perl.perl5.changes

[Perl/perl5] by Tony Cook via perl5-changes

Tue, 16 Dec 2025 22:56:06 +0000

Branch: refs/heads/23918-cmp-api
Home: https://github.com/Perl/perl5

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 16717d: add sv_numne() to the API by Tony Cook via perl5-changes

Tue, 16 Dec 2025 22:55:57 +0000

Branch: refs/heads/23918-cmp-api
Home: https://github.com/Perl/perl5
Commit: 16717d448459b076eee9955b9b9923de17cbf6b5
https://github.com/Perl/perl5/commit/16717d448459b076eee9955b9b9923de17cbf6b5
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M MANIFEST
M embed.fnc
M embed.h
M ext/XS-APItest/APItest.xs
M ext/XS-APItest/t/sv_numeq.t
A ext/XS-APItest/t/sv_numne.t
M proto.h
M sv.c
M sv.h

Log Message:
-----------
add sv_numne() to the API

some refactoring next, since sv_numeq_flags and sv_numne_flags are
similar.

Used a separate test file since putting every sv_num*() variant in the
one file would be ugly

Addresses GH #23918 but isn't a direct fix

Commit: d06055fc3362aa7ca9af4eb22fa414cfeb431dc7
https://github.com/Perl/perl5/commit/d06055fc3362aa7ca9af4eb22fa414cfeb431dc7
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M ext/XS-APItest/APItest.xs
M ext/XS-APItest/t/sv_numeq.t
M ext/XS-APItest/t/sv_numne.t
M proto.h
M sv.c

Log Message:
-----------
sv.c: extract the common parts of sv_numeq_flags and sv_numne_flags

Commit: bd4bba678c8e556fa082f87eeaa8eda489693b8f
https://github.com/Perl/perl5/commit/bd4bba678c8e556fa082f87eeaa8eda489693b8f
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M sv.c

Log Message:
-----------
sv_numeq/sv_numne: consolidate the similar documentation

If nothing else putting them together may avoid someone doing
`!sv_numeq(...)`

Commit: d41101a00dfbe22df4ee5c25c1ead01a6668c632
https://github.com/Perl/perl5/commit/d41101a00dfbe22df4ee5c25c1ead01a6668c632
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M MANIFEST
M embed.fnc
M embed.h
M ext/XS-APItest/APItest.xs
A ext/XS-APItest/t/sv_numcmp.t
M ext/XS-APItest/t/sv_numeq.t
M ext/XS-APItest/t/sv_numne.t
M proto.h
M sv.c
M sv.h

Log Message:
-----------
add sv_numcmp() to the API

Commit: ecff2db8c63f27bd4f055e13f2176b18910c9b28
https://github.com/Perl/perl5/commit/ecff2db8c63f27bd4f055e13f2176b18910c9b28
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M ext/XS-APItest/t/sv_numcmp.t
M ext/XS-APItest/t/sv_numeq.t
M ext/XS-APItest/t/sv_numne.t
M sv.c

Log Message:
-----------
sv_num*: correctly handle "0+" overloaded values

do_ncmp() expects simple SVs and for overloaded SVs will just compare
the SvNV() of each SV, mishandling the case where the 0+ overload
returns a large UV or IV that isn't exactly representable as an NV.

# Conflicts:
# ext/XS-APItest/t/sv_numeq.t
# ext/XS-APItest/t/sv_numne.t
# sv.c

Commit: 5c2ee02593cacca06d81769bc9486627d7c1a046
https://github.com/Perl/perl5/commit/5c2ee02593cacca06d81769bc9486627d7c1a046
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M MANIFEST
M embed.fnc
M embed.h
M ext/XS-APItest/APItest.xs
A ext/XS-APItest/t/sv_numlget.t
M proto.h
M sv.c
M sv.h

Log Message:
-----------
add sv_numle(), sv_numlt(), sv_numge(), sv_numgt() APIs

These are all needed because overloading may make them inconsistent
with <=> overloading.

Commit: 2dd07a2e7fca1f23aba7e89992259301d3673849
https://github.com/Perl/perl5/commit/2dd07a2e7fca1f23aba7e89992259301d3673849
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M sv.c

Log Message:
-----------
sv_numcmp_common: only call magic once if the SVs are the same

similar to try_amagic_bin()

Commit: 09fe3a12eea2360fa8d24811ed99727292d66681
https://github.com/Perl/perl5/commit/09fe3a12eea2360fa8d24811ed99727292d66681
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M ext/XS-APItest/APItest.xs
M ext/XS-APItest/t/sv_numne.t

Log Message:
-----------
add tests for void context calls to overloads

Discovered while working on another module, in many amagic_call() will
use the current context when calling the overload sub, but these APIs
might be called in XS code that simply needs a comparison, regardless
of the current OP context.

Commit: 0bb419d9ace4835b7d119b2d5b93f52dc4c6efb3
https://github.com/Perl/perl5/commit/0bb419d9ace4835b7d119b2d5b93f52dc4c6efb3
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M gv.c
M pp.h
M sv.c

Log Message:
-----------
amagic_call: accept a AMGf_force_scalar flag to force scalar context

and use it from the numeric comparison APIs.

Commit: 1ce4bc0d13d6268cce62bd1ac221c1f3b44069b8
https://github.com/Perl/perl5/commit/1ce4bc0d13d6268cce62bd1ac221c1f3b44069b8
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M gv.c
M pp_hot.c

Log Message:
-----------
pp_multiconcat: use the new AMGf_force_scalar

instead of a special case in amagic_call()

Commit: f4f735a8eb264c0c079b578c83dd1e790b436734
https://github.com/Perl/perl5/commit/f4f735a8eb264c0c079b578c83dd1e790b436734
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M ext/XS-APItest/t/sv_numeq.t
M proto.h
M sv.c
M sv.h

Log Message:
-----------
sv_numeq etc: don't do numify overloading with SV_SKIP_OVERLOAD

Commit: 2f8d46d1fef55f2cc85b77de1562e81b6eae4a9a
https://github.com/Perl/perl5/commit/2f8d46d1fef55f2cc85b77de1562e81b6eae4a9a
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M ext/XS-APItest/Makefile.PL
M ext/XS-APItest/t/sv_numne.t
M gv.c
M pp.h
M sv.c
M sv.h

Log Message:
-----------
add SV_FORCE_OVERLOAD to the sv_numcmp() APIs

and add AMGf_force_overload to amagic_call() which does the actual work.

Commit: 15e9ee5594598906c45e939b0d43a43d8eb2caa4
https://github.com/Perl/perl5/commit/15e9ee5594598906c45e939b0d43a43d8eb2caa4
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M ext/XS-APItest/APItest.xs
M ext/XS-APItest/t/sv_numeq.t

Log Message:
-----------
check the void context fix for sv_numeq too

Commit: 9db4e168080eaea3c6f9df8782220250b7171a2b
https://github.com/Perl/perl5/commit/9db4e168080eaea3c6f9df8782220250b7171a2b
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M pod/perldelta.pod
M sv.c

Log Message:
-----------
add perldelta for sv_numeq fix and other sv_num* additions

modified the sv.c documentation since the perldelta sv_numeq link had
multiple targets.

Commit: 178011c1679060e99d900a0e2e949be08faee2e3
https://github.com/Perl/perl5/commit/178011c1679060e99d900a0e2e949be08faee2e3
Author: Tony Cook <[email protected]>
Date: 2025-12-17 (Wed, 17 Dec 2025)

Changed paths:
M embed.fnc
M sv.c

Log Message:
-----------
Add some documentation for sv_numcmp_common()

Compare: https://github.com/Perl/perl5/compare/16717d448459%5E...178011c16790

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] d30679: Fatalize use of goto to jump into construct by James E Keenan via perl5-changes

Tue, 16 Dec 2025 15:14:12 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: d306795336cc8ce0c7c80e251adb1343497abbbd
https://github.com/Perl/perl5/commit/d306795336cc8ce0c7c80e251adb1343497abbbd
Author: James E Keenan <[email protected]>
Date: 2025-12-16 (Tue, 16 Dec 2025)

Changed paths:
M lib/warnings.pm
M pod/perldelta.pod
M pod/perldeprecation.pod
M pod/perldiag.pod
M pod/perlfunc.pod
M pp_ctl.c
M regen/warnings.pl
M t/comp/package_block.t
M t/op/goto.t
M t/porting/deprecation.t
M t/uni/labels.t
M warnings.h

Log Message:
-----------
Fatalize use of goto to jump into construct

* Adapt all tests.

* Revise discussion of 'goto' in perlfunc.

* Perl 5.44 fatalization: document goto_construct Also, note that no
deprecations or fatalizations occurred in perl-5.42.

* Remove 'goto_construct' from list of warnings. This entails:

** Entering 'deprecated::goto_construct" into %NO_BIT_FOR inside
regen/warnings.pl;

** Incrementing $VERSION in that file;

** Running `make regen` to propagate these changes to lib/warnings.pm and
warnings.h, in the process renumbering the warnings categories and
adding 'deprecated::goto_construct' to %NoOp inside lib/warnings.pm.

* Add entry to perldiag.pod

* Update perldelta: partial fatalization of goto construct.

* t/op/goto.t: Restore tests deleted during work on branch.

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 966e97: perlfunc: document return behavior of requireand ... by Grinnz via perl5-changes

Tue, 16 Dec 2025 12:08:36 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 966e97bd7e70f107042cfdeef1bc5bc7f69cb0c8
https://github.com/Perl/perl5/commit/966e97bd7e70f107042cfdeef1bc5bc7f69cb0c8
Author: Grinnz <[email protected]>
Date: 2025-12-16 (Tue, 16 Dec 2025)

Changed paths:
M .mailmap
M pod/perlfunc.pod

Log Message:
-----------
perlfunc: document return behavior of require and module_true feature

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 386a35: regex: eval/cut: fix premature local undo and$^R by David Mitchell via perl5-changes

Tue, 16 Dec 2025 11:19:26 +0000

Branch: refs/heads/davem/re_savestack_eval
Home: https://github.com/Perl/perl5
Commit: 386a359b4adc7a9b56b300cab367d14f7cddd068
https://github.com/Perl/perl5/commit/386a359b4adc7a9b56b300cab367d14f7cddd068
Author: David Mitchell <[email protected]>
Date: 2025-12-16 (Tue, 16 Dec 2025)

Changed paths:
M regexec.c
M regexp.h
M t/op/svleak.t
M t/re/pat_re_eval.t

Log Message:
-----------
regex: eval/cut: fix premature local undo and $^R

GH #16197

The main purpose of this commit is to stop premature scope unwinding
within eval code in regexes. Aside from backtracking on failure, the
scopes of every eval, such as (?{ local $x }), are supposed to accumulate,
and are only unwound en-masse at the very end of the pattern match.
However, a combination of a sub-pattern call, such as (??{...}) or
(?&...), combined with a cut, (?>...), can trigger earlier savestack
popping.

The direct fix for this, as explained below, is to remove this single
line from the EVAL_postponed_A/B: case:

REGCP_UNWIND(ST.cp); /* LEAVE in disguise */

However, that line is entwined with code which attempts to preserve the
final value of $^R during scope unwinding. Since that code was kind of
working around the misplaced REGCP_UNWIND(), it needs ripping out and
re-implementing. This has to be done at the same time, so the bulk of
this commit is actually concerned with $^R, even though it isn't the
subject of this ticket. So this commit doesn't change the behaviour of
$^R, but just changes its implementation somewhat.

The $^R issue is that every /(?{...})/ causes the equivalent of

local $^R = ...;

to be executed. During final exit, the savestack gets unwound, and all
those local's get undone, leaving $^R with the value it had before the
match started. But we promise that after the match, $^R will hold the
value of the most recent (?{...}). The code which this commit rips out
restored that value in one way; the new code in this commit does it a
different way. Basically, almost the last thing S_regmatch() does is a
LEAVE_SCOPE(orig_savestack_ix); This commit makes it so that the current
value of $^R is copied just before the LEAVE_SCOPE(), and that value is
copied back to $^R just after the LEAVE_SCOPE().

For efficiency, we only do the copy if we've actually set $^R. A mechanism
is also needed to ensure that the temporary copy doesn't leak if we die
during the savestack unwind. This is achieved by holding a pointer to
the copy in the aux_eval struct, which gets processed even if we die.

Now back to the main purpose of this commit, the premature stack unwind
in the presence of a cut with a sub-pattern. This bug has been there
since these features were added. It is instructive to look at a somewhat
idealised overview of the S_regmatch() function from around 5.6.0 (with
some bug fixes from later releases added). This was while the function
was still recursive. It looks approximately like:

pp_match(...) {
I32 ix = PL_savestack_ix;
... do a match ...
LEAVE_SCOPE(ix);
}

S_regmatch(...) {
while (scan) {
switch (OP(scan)) {
case FOO:
if (! there's a FOO)
return 0;
I32 ix = PL_savestack_ix;
if (regmatch(...)) /* recursively match rest of pattern */
return 1;
LEAVE_SCOPE(ix);
return 0;

case END:
if (doing a (??{...}) ) {
I32 ix = PL_savestack_ix;
if (regmatch(...) { /* recursively match rest of pattern */
LEAVE_SCOPE(ix);
return 1;
}
LEAVE_SCOPE(ix);
return 0;
}
return 1;

case EVAL:
... run the code, then, if its a (?{...}) ...
I32 ix = PL_savestack_ix;
if (regmatch(...) { /* recursively run subpattern */
LEAVE_SCOPE(ix)
return 1;
}
LEAVE_SCOPE(ix);
return 0;
}
}

Here, the FOO: case represents all the various ops which recurse. In
general, they match the next item and then recurse to match the rest of
the pattern.

Note that they all do a LEAVE_SCOPE() only in the *failure* branch.

At the end of a successful match, there is potentially much recursion,
and much stuff on the savestack. When the END op is reached, the series
of 'return 1's causes all the recursion to unwind, while leaving the
savestack untouched. Finally, the caller - such as pp_match() - clears
the savestack. In more recent perls the recursion has been removed and
the final LEAVE_SCOPE() is done within S_regmatch() itself, but the
principle remains the same: no stack freeing is done *during* matching,
and instead there's a single big clean up at the end.

Once (??{...}) enters the picture, that changes a bit. When the END op
associated with the '...' sub-pattern is reached, regmatch() is called
recursively to process any pattern after the (??{..}); then on success,
while working its way back through the nested regmatch() calls, both the
END and the EVAL code do a LEAVE_SCOPE() in the *success* branch.

This is anomalous, and those two LEAVE_SCOPE()'s are what this commit
removes (although in the current non-recursive regex engine, they are
shared by the same piece of code, so only one had to be removed). By
removing them, this regularises the behaviour of sub-patterns. I can't
think why those LEAVE_SCOPE()s were originally added, and assume it was a
thinko.

Normally it makes no difference whether the savestack is popped near the
end, interleaved with popping all recursive regmatch calls (or
equivalently on non-recursive engines popping the regmatch_state stack),
or whether the savestack is popped only after all the recursion is
exited from.

However, it makes a difference in the presence of a cut, (?>...). Here,
the final op of the sub-pattern '...' is SUCCEED, which rather than
recursing to match anything following the cut block, just returns. The
recursion pops back to the SUSPEND op which started the cut, which then
continues with the op loop as normal.

Thus when about to match the first op following a (?>...), the recursion
*within* the cut has been blown away as if it never happened, but the
accumulated savestack entries (e.g. from evals within the cut block) are
preserved and continue to accumulate.

Now, if there is a (??{...}) sub-pattern, or similarly a (?&FOO), within
the cut, then at the end of the cut, the recursion is unwound, which
includes the stacked EVAL and END recursions, which at this time call
LEAVE_SCOPE(), which frees part of the savestack, even though the
pattern match hasn't ended yet. That's the bug which this commit fixes.

The tests added to pat_re_eval.t check that the scope bug has been
fixed. The test added to svleak.t checks that the new $^R copying code
doesn't leak.

Commit: d377999ae732c860f975e756f66b36e7dafeb2f2
https://github.com/Perl/perl5/commit/d377999ae732c860f975e756f66b36e7dafeb2f2
Author: David Mitchell <[email protected]>
Date: 2025-12-16 (Tue, 16 Dec 2025)

Changed paths:
M regexec.c

Log Message:
-----------
regex: don't LEAVE_SCOPE() in S_regtry()

This commit should produce no practical change in functionality.

Currently, S_regtry() notes the position of PL_savestack_ix, calls
S_regmatch(), then pops the savestack back to that position. However,
S_regmatch() also does this just before returning, so it's redundant in
S_regtry(). (A temporary assert confirmed that lastcp == PL_savestack_ix
in S_regtry always while running the test suite).

So this commit removes the REGCP_UNWIND(lastcp) and associated machinery
from S_regtry().

It also regularises the "note current ix; pop back to old ix" code at
the start and end of S_regmatch() to use the standard REGCP_SET() and
REGCP_UNWIND() macros which do the same thing but also produce debugging
messages.

Commit: 8f9d341018ab6d57e646d218fc8e5ed5555963d8
https://github.com/Perl/perl5/commit/8f9d341018ab6d57e646d218fc8e5ed5555963d8
Author: David Mitchell <[email protected]>
Date: 2025-12-16 (Tue, 16 Dec 2025)

Changed paths:
M pod/perldelta.pod

Log Message:
-----------
perldelta for GH #16197 fix

Compare: https://github.com/Perl/perl5/compare/97c59b72e204...8f9d341018ab

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 8fd53b: RMG, remove trailing whitespace by Eric Herman via perl5-changes

Tue, 16 Dec 2025 09:10:51 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 8fd53b46c59e507154709ba96ed3456263999135
https://github.com/Perl/perl5/commit/8fd53b46c59e507154709ba96ed3456263999135
Author: Eric Herman <[email protected]>
Date: 2025-12-16 (Tue, 16 Dec 2025)

Changed paths:
M Porting/release_managers_guide.pod

Log Message:
-----------
RMG, remove trailing whitespace

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] acc8ad: embed.fnc: Rename EPTR qualifiers by Karl Williamson via perl5-changes

Mon, 15 Dec 2025 23:58:17 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: acc8ad3a2865555ac1ee10ef8e728321bb8e7076
https://github.com/Perl/perl5/commit/acc8ad3a2865555ac1ee10ef8e728321bb8e7076
Author: Karl Williamson <[email protected]>
Date: 2025-12-15 (Mon, 15 Dec 2025)

Changed paths:
M embed.fnc
M regen/embed.pl

Log Message:
-----------
embed.fnc: Rename EPTR qualifiers

People found the former names confusing. This changes to use EPTRgt and
EPTRge that explicitly give the relation that gets used.

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 7ddc00: regex: eval/cut: fix premature local undo and$^R by David Mitchell via perl5-changes

Sun, 14 Dec 2025 14:43:45 +0000

Branch: refs/heads/davem/re_savestack_eval
Home: https://github.com/Perl/perl5
Commit: 7ddc0085812ab89a573ada9988fdb81e91f1a041
https://github.com/Perl/perl5/commit/7ddc0085812ab89a573ada9988fdb81e91f1a041
Author: David Mitchell <[email protected]>
Date: 2025-12-14 (Sun, 14 Dec 2025)

Changed paths:
M regexec.c
M regexp.h
M t/op/svleak.t
M t/re/pat_re_eval.t

Log Message:
-----------
regex: eval/cut: fix premature local undo and $^R

GH #16197

The main purpose of this commit is to stop premature scope unwinding
within eval code in regexes. Aside from backtracking on failure, the
scopes of every eval, such as (?{ local $x }), are supposed to accumulate,
and are only unwound en-masse at the very end of the pattern match.
However, a combination of a sub-pattern call, such as (??{...}) or
(?&...), combined with a cut, (?>...), can trigger earlier savestack
popping.

The direct fix for this, as explained below, is to remove this single
line from the EVAL_postponed_A/B: case:

REGCP_UNWIND(ST.cp); /* LEAVE in disguise */

However, that line is entwined with code which attempts to preserve the
final value of $^R during scope unwinding. Since that code was kind of
working around the misplaced REGCP_UNWIND(), it needs ripping out and
re-implementing. This has to be done at the same time, so the bulk of
this commit is actually concerned with $^R, even though it isn't the
subject of this ticket. So this commit doesn't change the behaviour of
$^R, but just changes its implementation somewhat.

The $^R issue is that every /(?{...})/ causes the equivalent of

local $^R = ...;

to be executed. During final exit, the savestack gets unwound, and all
those local's get undone, leaving $^R with the value it had before the
match started. But we promise that after the match, $^R will hold the
value of the most recent (?{...}). The code which this commit rips out
restored that value in one way; the new code in this commit does it a
different way. Basically, almost the last thing S_regmatch() does is a
LEAVE_SCOPE(orig_savestack_ix); This commit makes it so that the current
value of $^R is copied just before the LEAVE_SCOPE(), and that value is
copied back to $^R just after the LEAVE_SCOPE().

For efficiency, we only do the copy if we've actually set $^R. A mechanism
is also needed to ensure that the temporary copy doesn't leak if we die
during the savestack unwind. This is achieved by holding a pointer to
the copy in the aux_eval struct, which gets processed even if we die.

Now back to the main purpose of this commit, the premature stack unwind
in the presence of a cut with a sub-pattern. This bug has been there
since these features were added. It is instructive to look at a somewhat
idealised overview of the S_regmatch() function from around 5.6.0 (with
some bug fixes from later releases added). This was while the function
was still recursive. It looks approximately like:

pp_match(...) {
I32 ix = PL_savestack_ix;
... do a match ...
LEAVE_SCOPE(ix);
}

S_regmatch(...) {
while (scan) {
switch (OP(scan)) {
case FOO:
if (! there's a FOO)
return 0;
I32 ix = PL_savestack_ix;
if (regmatch(...)) /* recursively match rest of pattern */
return 1;
LEAVE_SCOPE(ix);
return 0;

case END:
if (doing a (??{...}) ) {
I32 ix = PL_savestack_ix;
if (regmatch(...) { /* recursively match rest of pattern */
LEAVE_SCOPE(ix);
return 1;
}
LEAVE_SCOPE(ix);
return 0;
}
return 1;

case EVAL:
... run the code, then, if its a (?{...}) ...
I32 ix = PL_savestack_ix;
if (regmatch(...) { /* recursively run subpattern */
LEAVE_SCOPE(ix)
return 1;
}
LEAVE_SCOPE(ix);
return 0;
}
}

Here, the FOO: case represents all the various ops which recurse. In
general, they match the next item and then recurse to match the rest of
the pattern.

Note that they all do a LEAVE_SCOPE() only in the *failure* branch.

At the end of a successful match, there is potentially much recursion,
and much stuff on the savestack. When the END op is reached, the series
of 'return 1's causes all the recursion to unwind, while leaving the
savestack untouched. Finally, the caller - such as pp_match() - clears
the savestack. In more recent perls the recursion has been removed and
the final LEAVE_SCOPE() is done within S_regmatch() itself, but the
principle remains the same: no stack freeing is done *during* matching,
and instead there's a single big clean up at the end.

Once (??{...}) enters the picture, that changes a bit. When the END op
associated with the '...' sub-pattern is reached, regmatch() is called
recursively to process any pattern after the (??{..}); then on success,
while working its way back through the nested regmatch() calls, both the
END and the EVAL code do a LEAVE_SCOPE() in the *success* branch.

This is anomalous, and those two LEAVE_SCOPE()'s are what this commit
removes (although in the current non-recursive regex engine, they are
shared by the same piece of code, so only one had to be removed). By
removing them, this regularises the behaviour of sub-patterns. I can't
think why those LEAVE_SCOPE()s were originally added, and assume it was a
thinko.

Normally it makes no difference whether the savestack is popped near the
end, interleaved with popping all recursive regmatch calls (or
equivalently on non-recursive engines popping the regmatch_state stack),
or whether the savestack is popped only after all the recursion is
exited from.

However, it makes a difference in the presence of a cut, (?>...). Here,
the final op of the sub-pattern '...' is SUCCEED, which rather than
recursing to match anything following the cut block, just returns. The
recursion pops back to the SUSPEND op which started the cut, which then
continues with the op loop as normal.

Thus when about to match the first op following a (?>...), the recursion
*within* the cut has been blown away as if it never happened, but the
accumulated savestack entries (e.g. from evals within the cut block) are
preserved and continue to accumulate.

Now, if there is a (??{...}) sub-pattern, or similarly a (?&FOO), within
the cut, then at the end of the cut, the recursion is unwound, which
includes the stacked EVAL and END recursions, which at this time call
LEAVE_SCOPE(), which frees part of the savestack, even though the
pattern match hasn't ended yet. That's the bug which this commit fixes.

The tests added to pat_re_eval.t check that the scope bug has been
fixed. The test added to svleak.t checks that the new $^R copying code
doesn't leak.

Commit: eaa68d31d4451ce89c75e3c1ca265479259a6a8f
https://github.com/Perl/perl5/commit/eaa68d31d4451ce89c75e3c1ca265479259a6a8f
Author: David Mitchell <[email protected]>
Date: 2025-12-14 (Sun, 14 Dec 2025)

Changed paths:
M regexec.c

Log Message:
-----------
regex: don't LEAVE_SCOPE() in S_regtry()

This commit should produce no practical change in functionality.

Currently, S_regtry() notes the position of PL_savestack_ix, calls
S_regmatch(), then pops the savestack back to that position. However,
S_regmatch() also does this just before returning, so it's redundant in
S_regtry(). (A temporary assert confirmed that lastcp == PL_savestack_ix
in S_regtry always while running the test suite).

So this commit removes the REGCP_UNWIND(lastcp) and associated machinery
from S_regtry().

It also regularises the "note current ix; pop back to old ix" code at
the start and end of S_regmatch() to use the standard REGCP_SET() and
REGCP_UNWIND() macros which do the same thing but also produce debugging
messages.

Commit: 97c59b72e2049cf3c8787e81f291fbf7a3ba21de
https://github.com/Perl/perl5/commit/97c59b72e2049cf3c8787e81f291fbf7a3ba21de
Author: David Mitchell <[email protected]>
Date: 2025-12-14 (Sun, 14 Dec 2025)

Changed paths:
M pod/perldelta.pod

Log Message:
-----------
perldelta for GH #16197 fix

Compare: https://github.com/Perl/perl5/compare/7ddc0085812a%5E...97c59b72e204

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 55631d: Refactor Perl_av_fetch: remove goto, reduce thenu... by Bartosz Jarzyna via perl5-changes

Sun, 14 Dec 2025 10:39:28 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 55631dd1f86b43a00fcc9fb5a70ed8fc52eeaa93
https://github.com/Perl/perl5/commit/55631dd1f86b43a00fcc9fb5a70ed8fc52eeaa93
Author: bbrtj <[email protected]>
Date: 2025-12-14 (Sun, 14 Dec 2025)

Changed paths:
M av.c

Log Message:
-----------
Refactor Perl_av_fetch: remove goto, reduce the number of branches

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 10adbc: cpan/Test-Simple - Update to version 1.302219 by James E Keenan via perl5-changes

Sun, 14 Dec 2025 02:18:36 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 10adbc06411081cf1df37c2a3cd8546d565f1f1e
https://github.com/Perl/perl5/commit/10adbc06411081cf1df37c2a3cd8546d565f1f1e
Author: Chad Granum <[email protected]>
Date: 2025-12-13 (Sat, 13 Dec 2025)

Changed paths:
M MANIFEST
M Makefile.SH
M Porting/Maintainers.pl
M cpan/Test-Simple/lib/Test/Builder.pm
M cpan/Test-Simple/lib/Test/Builder/Formatter.pm
M cpan/Test-Simple/lib/Test/Builder/Module.pm
M cpan/Test-Simple/lib/Test/Builder/Tester.pm
M cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm
M cpan/Test-Simple/lib/Test/Builder/TodoDiag.pm
M cpan/Test-Simple/lib/Test/More.pm
M cpan/Test-Simple/lib/Test/Simple.pm
M cpan/Test-Simple/lib/Test/Tester.pm
M cpan/Test-Simple/lib/Test/Tester/Capture.pm
M cpan/Test-Simple/lib/Test/Tester/CaptureRunner.pm
M cpan/Test-Simple/lib/Test/Tester/Delegate.pm
M cpan/Test-Simple/lib/Test/use/ok.pm
M cpan/Test-Simple/lib/Test2.pm
M cpan/Test-Simple/lib/Test2/API.pm
M cpan/Test-Simple/lib/Test2/API/Breakage.pm
M cpan/Test-Simple/lib/Test2/API/Context.pm
M cpan/Test-Simple/lib/Test2/API/Instance.pm
M cpan/Test-Simple/lib/Test2/API/InterceptResult.pm
M cpan/Test-Simple/lib/Test2/API/InterceptResult/Event.pm
M cpan/Test-Simple/lib/Test2/API/InterceptResult/Facet.pm
M cpan/Test-Simple/lib/Test2/API/InterceptResult/Hub.pm
M cpan/Test-Simple/lib/Test2/API/InterceptResult/Squasher.pm
M cpan/Test-Simple/lib/Test2/API/Stack.pm
M cpan/Test-Simple/lib/Test2/AsyncSubtest.pm
M cpan/Test-Simple/lib/Test2/AsyncSubtest/Event/Attach.pm
M cpan/Test-Simple/lib/Test2/AsyncSubtest/Event/Detach.pm
M cpan/Test-Simple/lib/Test2/AsyncSubtest/Formatter.pm
M cpan/Test-Simple/lib/Test2/AsyncSubtest/Hub.pm
M cpan/Test-Simple/lib/Test2/Bundle.pm
M cpan/Test-Simple/lib/Test2/Bundle/Extended.pm
M cpan/Test-Simple/lib/Test2/Bundle/More.pm
M cpan/Test-Simple/lib/Test2/Bundle/Simple.pm
M cpan/Test-Simple/lib/Test2/Compare.pm
M cpan/Test-Simple/lib/Test2/Compare/Array.pm
M cpan/Test-Simple/lib/Test2/Compare/Bag.pm
M cpan/Test-Simple/lib/Test2/Compare/Base.pm
M cpan/Test-Simple/lib/Test2/Compare/Bool.pm
M cpan/Test-Simple/lib/Test2/Compare/Custom.pm
M cpan/Test-Simple/lib/Test2/Compare/DeepRef.pm
M cpan/Test-Simple/lib/Test2/Compare/Delta.pm
M cpan/Test-Simple/lib/Test2/Compare/Event.pm
M cpan/Test-Simple/lib/Test2/Compare/EventMeta.pm
M cpan/Test-Simple/lib/Test2/Compare/Float.pm
M cpan/Test-Simple/lib/Test2/Compare/Hash.pm
M cpan/Test-Simple/lib/Test2/Compare/Isa.pm
M cpan/Test-Simple/lib/Test2/Compare/Meta.pm
M cpan/Test-Simple/lib/Test2/Compare/Negatable.pm
M cpan/Test-Simple/lib/Test2/Compare/Number.pm
M cpan/Test-Simple/lib/Test2/Compare/Object.pm
M cpan/Test-Simple/lib/Test2/Compare/OrderedSubset.pm
M cpan/Test-Simple/lib/Test2/Compare/Pattern.pm
M cpan/Test-Simple/lib/Test2/Compare/Ref.pm
M cpan/Test-Simple/lib/Test2/Compare/Regex.pm
M cpan/Test-Simple/lib/Test2/Compare/Scalar.pm
M cpan/Test-Simple/lib/Test2/Compare/Set.pm
M cpan/Test-Simple/lib/Test2/Compare/String.pm
M cpan/Test-Simple/lib/Test2/Compare/Undef.pm
M cpan/Test-Simple/lib/Test2/Compare/Wildcard.pm
M cpan/Test-Simple/lib/Test2/Env.pm
M cpan/Test-Simple/lib/Test2/Event.pm
M cpan/Test-Simple/lib/Test2/Event/Bail.pm
M cpan/Test-Simple/lib/Test2/Event/Diag.pm
M cpan/Test-Simple/lib/Test2/Event/Encoding.pm
M cpan/Test-Simple/lib/Test2/Event/Exception.pm
M cpan/Test-Simple/lib/Test2/Event/Fail.pm
M cpan/Test-Simple/lib/Test2/Event/Generic.pm
M cpan/Test-Simple/lib/Test2/Event/Note.pm
M cpan/Test-Simple/lib/Test2/Event/Ok.pm
M cpan/Test-Simple/lib/Test2/Event/Pass.pm
M cpan/Test-Simple/lib/Test2/Event/Plan.pm
M cpan/Test-Simple/lib/Test2/Event/Skip.pm
M cpan/Test-Simple/lib/Test2/Event/Subtest.pm
M cpan/Test-Simple/lib/Test2/Event/TAP/Version.pm
M cpan/Test-Simple/lib/Test2/Event/V2.pm
M cpan/Test-Simple/lib/Test2/Event/Waiting.pm
M cpan/Test-Simple/lib/Test2/EventFacet.pm
M cpan/Test-Simple/lib/Test2/EventFacet/About.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Amnesty.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Assert.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Control.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Error.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Hub.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Info.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Info/Table.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Meta.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Parent.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Plan.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Render.pm
M cpan/Test-Simple/lib/Test2/EventFacet/Trace.pm
M cpan/Test-Simple/lib/Test2/Formatter.pm
M cpan/Test-Simple/lib/Test2/Formatter/TAP.pm
A cpan/Test-Simple/lib/Test2/Handle.pm
M cpan/Test-Simple/lib/Test2/Hub.pm
M cpan/Test-Simple/lib/Test2/Hub/Interceptor.pm
M cpan/Test-Simple/lib/Test2/Hub/Interceptor/Terminator.pm
M cpan/Test-Simple/lib/Test2/Hub/Subtest.pm
M cpan/Test-Simple/lib/Test2/IPC.pm
M cpan/Test-Simple/lib/Test2/IPC/Driver.pm
M cpan/Test-Simple/lib/Test2/IPC/Driver/Files.pm
M cpan/Test-Simple/lib/Test2/Manual.pm
M cpan/Test-Simple/lib/Test2/Manual/Anatomy.pm
M cpan/Test-Simple/lib/Test2/Manual/Anatomy/API.pm
M cpan/Test-Simple/lib/Test2/Manual/Anatomy/Context.pm
M cpan/Test-Simple/lib/Test2/Manual/Anatomy/EndToEnd.pm
M cpan/Test-Simple/lib/Test2/Manual/Anatomy/Event.pm
M cpan/Test-Simple/lib/Test2/Manual/Anatomy/Hubs.pm
M cpan/Test-Simple/lib/Test2/Manual/Anatomy/IPC.pm
M cpan/Test-Simple/lib/Test2/Manual/Anatomy/Utilities.pm
M cpan/Test-Simple/lib/Test2/Manual/Concurrency.pm
M cpan/Test-Simple/lib/Test2/Manual/Contributing.pm
M cpan/Test-Simple/lib/Test2/Manual/Testing.pm
M cpan/Test-Simple/lib/Test2/Manual/Testing/Introduction.pm
M cpan/Test-Simple/lib/Test2/Manual/Testing/Migrating.pm
M cpan/Test-Simple/lib/Test2/Manual/Testing/Planning.pm
M cpan/Test-Simple/lib/Test2/Manual/Testing/Todo.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/FirstTool.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/Formatter.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/Nesting.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/Plugin/TestExit.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/Plugin/TestingDone.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/Plugin/ToolCompletes.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/Plugin/ToolStarts.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/Subtest.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/TestBuilder.pm
M cpan/Test-Simple/lib/Test2/Manual/Tooling/Testing.pm
M cpan/Test-Simple/lib/Test2/Mock.pm
M cpan/Test-Simple/lib/Test2/Plugin.pm
M cpan/Test-Simple/lib/Test2/Plugin/BailOnFail.pm
M cpan/Test-Simple/lib/Test2/Plugin/DieOnFail.pm
M cpan/Test-Simple/lib/Test2/Plugin/ExitSummary.pm
M cpan/Test-Simple/lib/Test2/Plugin/SRand.pm
M cpan/Test-Simple/lib/Test2/Plugin/Times.pm
M cpan/Test-Simple/lib/Test2/Plugin/UTF8.pm
M cpan/Test-Simple/lib/Test2/Require.pm
M cpan/Test-Simple/lib/Test2/Require/AuthorTesting.pm
M cpan/Test-Simple/lib/Test2/Require/AutomatedTesting.pm
M cpan/Test-Simple/lib/Test2/Require/EnvVar.pm
M cpan/Test-Simple/lib/Test2/Require/ExtendedTesting.pm
M cpan/Test-Simple/lib/Test2/Require/Fork.pm
M cpan/Test-Simple/lib/Test2/Require/Module.pm
M cpan/Test-Simple/lib/Test2/Require/NonInteractiveTesting.pm
M cpan/Test-Simple/lib/Test2/Require/Perl.pm
M cpan/Test-Simple/lib/Test2/Require/RealFork.pm
M cpan/Test-Simple/lib/Test2/Require/ReleaseTesting.pm
M cpan/Test-Simple/lib/Test2/Require/Threads.pm
M cpan/Test-Simple/lib/Test2/Suite.pm
M cpan/Test-Simple/lib/Test2/Todo.pm
M cpan/Test-Simple/lib/Test2/Tools.pm
M cpan/Test-Simple/lib/Test2/Tools/AsyncSubtest.pm
M cpan/Test-Simple/lib/Test2/Tools/Basic.pm
M cpan/Test-Simple/lib/Test2/Tools/Class.pm
M cpan/Test-Simple/lib/Test2/Tools/ClassicCompare.pm
M cpan/Test-Simple/lib/Test2/Tools/Compare.pm
M cpan/Test-Simple/lib/Test2/Tools/Defer.pm
M cpan/Test-Simple/lib/Test2/Tools/Encoding.pm
M cpan/Test-Simple/lib/Test2/Tools/Event.pm
M cpan/Test-Simple/lib/Test2/Tools/Exception.pm
M cpan/Test-Simple/lib/Test2/Tools/Exports.pm
M cpan/Test-Simple/lib/Test2/Tools/GenTemp.pm
M cpan/Test-Simple/lib/Test2/Tools/Grab.pm
M cpan/Test-Simple/lib/Test2/Tools/Mock.pm
M cpan/Test-Simple/lib/Test2/Tools/Ref.pm
M cpan/Test-Simple/lib/Test2/Tools/Refcount.pm
M cpan/Test-Simple/lib/Test2/Tools/Spec.pm
M cpan/Test-Simple/lib/Test2/Tools/Subtest.pm
M cpan/Test-Simple/lib/Test2/Tools/Target.pm
M cpan/Test-Simple/lib/Test2/Tools/Tester.pm
M cpan/Test-Simple/lib/Test2/Tools/Tiny.pm
M cpan/Test-Simple/lib/Test2/Tools/Warnings.pm
M cpan/Test-Simple/lib/Test2/Util.pm
M cpan/Test-Simple/lib/Test2/Util/ExternalMeta.pm
M cpan/Test-Simple/lib/Test2/Util/Facets2Legacy.pm
M cpan/Test-Simple/lib/Test2/Util/Grabber.pm
M cpan/Test-Simple/lib/Test2/Util/Guard.pm
M cpan/Test-Simple/lib/Test2/Util/HashBase.pm
M cpan/Test-Simple/lib/Test2/Util/Importer.pm
M cpan/Test-Simple/lib/Test2/Util/Ref.pm
M cpan/Test-Simple/lib/Test2/Util/Sig.pm
M cpan/Test-Simple/lib/Test2/Util/Stash.pm
M cpan/Test-Simple/lib/Test2/Util/Sub.pm
M cpan/Test-Simple/lib/Test2/Util/Table.pm
M cpan/Test-Simple/lib/Test2/Util/Table/Cell.pm
M cpan/Test-Simple/lib/Test2/Util/Table/LineBreak.pm
M cpan/Test-Simple/lib/Test2/Util/Term.pm
M cpan/Test-Simple/lib/Test2/Util/Times.pm
M cpan/Test-Simple/lib/Test2/Util/Trace.pm
M cpan/Test-Simple/lib/Test2/V0.pm
A cpan/Test-Simple/lib/Test2/V1.pm
A cpan/Test-Simple/lib/Test2/V1/Base.pm
A cpan/Test-Simple/lib/Test2/V1/Handle.pm
M cpan/Test-Simple/lib/Test2/Workflow.pm
M cpan/Test-Simple/lib/Test2/Workflow/BlockBase.pm
M cpan/Test-Simple/lib/Test2/Workflow/Build.pm
M cpan/Test-Simple/lib/Test2/Workflow/Runner.pm
M cpan/Test-Simple/lib/Test2/Workflow/Task.pm
M cpan/Test-Simple/lib/Test2/Workflow/Task/Action.pm
M cpan/Test-Simple/lib/Test2/Workflow/Task/Group.pm
M cpan/Test-Simple/lib/ok.pm
A cpan/Test-Simple/t/modules/Handle.t
A cpan/Test-Simple/t/modules/V1.t
A cpan/Test-Simple/t/modules/V1/Base.t
A cpan/Test-Simple/t/modules/V1/Handle.t

Log Message:
-----------
cpan/Test-Simple - Update to version 1.302219

1.302219 2025-12-08 16:09:01-08:00 America/Los_Angeles

- Fix downstream test not to install Test2::Suite
- Fix documentation mistakes

Committer: Changes to Makefile.SH.

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 53d914: perlxs.pod: remove section on %v by David Mitchell via perl5-changes

Fri, 12 Dec 2025 14:57:54 +0000

Branch: refs/heads/davem/xspod
Home: https://github.com/Perl/perl5
Commit: 53d9146f7eaf0eccd08f2bc301ce2424170f9218
https://github.com/Perl/perl5/commit/53d9146f7eaf0eccd08f2bc301ce2424170f9218
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: remove section on %v

The XS parser supported an extremely obscure bit of functionality which
made use of the %v package variable to maintain state between different
bits of typemap processing. This was accidentally broken in 5.10.0:
refactoring removed the 'use vars "%v"' line, and no one seemed to
notice or care.

Also, the sole example of its use in the docs seemed to be obscure,
confusing and probably wrong.

There was a consensus in the discussion at

http://nntp.perl.org/group/perl.perl5.porters/267667

that we should stop documenting this feature rather than trying to fix
it.

Commit: 88fb9b221bcba5d6c75dcb76060fbf3f9245e45d
https://github.com/Perl/perl5/commit/88fb9b221bcba5d6c75dcb76060fbf3f9245e45d
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: reindent and reformat code examples

The various XS code examples had odd and inconsistent indentation (often
with 5 leading spaces) and inconsistent formatting, e.g. foo(a,b) vs
foo( a, b ) vs foo(a, b). Fix that, and also remove any tab chars.

Whitespace-only change.

Commit: 74260e4e198ac6c139ad7d8ed73c135538033195
https://github.com/Perl/perl5/commit/74260e4e198ac6c139ad7d8ed73c135538033195
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: delete most non-ref sections

This commit is a simple cut which deletes several '=head2' sections from
perlxs.pod. The next commit will tidy up and fix any broken links etc.

These sections are more tutorial-like, and aren't in line with the goal
of this branch that perlxs.pod becomes purely a reference manual for XS.
Any relevant information from these sections may be incorporated later
into new sections in perlxs.pod and/or be included in a future rewrite
of perlxstut.pod.

The sections deleted are:

=head2 Introduction
=head2 On The Road
=head2 The Anatomy of an XSUB
=head2 The Argument Stack
=head2 The RETVAL Variable
=head2 Returning SVs, AVs and HVs through RETVAL
=head2 Returning Undef And Empty Lists
=head2 Interface Strategy
=head2 Perl Objects And C Structures

Commit: a15b0e56b70a167456221e3ed10721c15fdefc13
https://github.com/Perl/perl5/commit/a15b0e56b70a167456221e3ed10721c15fdefc13
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M XSUB.h
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: fix up links after deleting sections

The previous commit deleted several sections from perlxs.pod. This
commit fixes things up; done as a separate commit so that the changes
aren't drowned out in the diff listing.

Commit: 1463566d83d7b94c5045970689ceb475767f8560
https://github.com/Perl/perl5/commit/1463566d83d7b94c5045970689ceb475767f8560
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: reorder sections

This big commit does a series of plain cut+pastes to reorder all the
=head2 sections within the file.

This changes the order from semi-random into roughly the order the
various XS keywords would appear within an XS file, and then within an
XSUB declaration/definition.

No changes have been made to the text: simply that all lines from a
particular '^=head2' up until the next head2 have been cut+paste as
a single unit.

No attempt has been made yet to make the text consistent with the new
ordering; that will be done by the subsequent commits of this branch.

The previous ordering in this file was:

=head1 NAME
=head1 DESCRIPTION
=head2 The MODULE Keyword
=head2 The PACKAGE Keyword
=head2 The PREFIX Keyword
=head2 The OUTPUT: Keyword
=head2 The NO_OUTPUT Keyword
=head2 The CODE: Keyword
=head2 The INIT: Keyword
=head2 The NO_INIT Keyword
=head2 The TYPEMAP: Keyword
=head2 Initializing Function Parameters
=head2 Default Parameter Values
=head2 The PREINIT: Keyword
=head2 The SCOPE: Keyword
=head2 The INPUT: Keyword
=head2 The IN/OUTLIST/IN_OUTLIST/OUT/IN_OUT Keywords
=head2 The C<length(NAME)> Keyword
=head2 Variable-length Parameter Lists
=head2 The C_ARGS: Keyword
=head2 The PPCODE: Keyword
=head2 The REQUIRE: Keyword
=head2 The CLEANUP: Keyword
=head2 The POSTCALL: Keyword
=head2 The BOOT: Keyword
=head2 The VERSIONCHECK: Keyword
=head2 The PROTOTYPES: Keyword
=head2 The PROTOTYPE: Keyword
=head2 The ALIAS: Keyword
=head2 The OVERLOAD: Keyword
=head2 The FALLBACK: Keyword
=head2 The INTERFACE: Keyword
=head2 The INTERFACE_MACRO: Keyword
=head2 The INCLUDE: Keyword
=head2 The INCLUDE_COMMAND: Keyword
=head2 The CASE: Keyword
=head2 The EXPORT_XSUB_SYMBOLS: Keyword
=head2 The & Unary Operator
=head2 Inserting POD, Comments and C Preprocessor Directives
=head2 Using XS With C++
=head2 Safely Storing Static Data in XS
=head3 MY_CXT REFERENCE
=head1 EXAMPLES
=head1 CAVEATS
=head2 Use of standard C library functions
=head2 Event loops and control flow
=head1 XS VERSION
=head1 AUTHOR DIAGNOSTICS
=head1 AUTHOR

and is now:

=head1 NAME
=head1 DESCRIPTION
=head2 The MODULE Keyword
=head2 The PACKAGE Keyword
=head2 The PREFIX Keyword
=head2 Inserting POD, Comments and C Preprocessor Directives
=head2 The REQUIRE: Keyword
=head2 The VERSIONCHECK: Keyword
=head2 The PROTOTYPES: Keyword
=head2 The EXPORT_XSUB_SYMBOLS: Keyword
=head2 The INCLUDE: Keyword
=head2 The INCLUDE_COMMAND: Keyword
=head2 The TYPEMAP: Keyword
=head2 The BOOT: Keyword
=head2 The FALLBACK: Keyword
=head2 The NO_OUTPUT Keyword
=head2 The IN/OUTLIST/IN_OUTLIST/OUT/IN_OUT Keywords
=head2 Default Parameter Values
=head2 The C<length(NAME)> Keyword
=head2 Variable-length Parameter Lists
=head2 The PREINIT: Keyword
=head2 The INPUT: Keyword
=head2 The NO_INIT Keyword
=head2 Initializing Function Parameters
=head2 The & Unary Operator
=head2 The SCOPE: Keyword
=head2 The INIT: Keyword
=head2 The C_ARGS: Keyword
=head2 The CODE: Keyword
=head2 The PPCODE: Keyword
=head2 The POSTCALL: Keyword
=head2 The OUTPUT: Keyword
=head2 The CLEANUP: Keyword
=head2 The PROTOTYPE: Keyword
=head2 The OVERLOAD: Keyword
=head2 The ALIAS: Keyword
=head2 The INTERFACE: Keyword
=head2 The INTERFACE_MACRO: Keyword
=head2 The CASE: Keyword
=head2 Using XS With C++
=head2 Safely Storing Static Data in XS
=head3 MY_CXT REFERENCE
=head1 EXAMPLES
=head1 CAVEATS
=head2 Use of standard C library functions
=head2 Event loops and control flow
=head1 XS VERSION
=head1 AUTHOR DIAGNOSTICS
=head1 AUTHOR

Commit: 855df4221c8d8d5403cc74e1f213850723a6a0fa
https://github.com/Perl/perl5/commit/855df4221c8d8d5403cc74e1f213850723a6a0fa
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add group headers

Following the previous commit's reordering of the all the =head2
sections, demote most of the =head2 headers to =head3, and add some new
=head2 headers which group together related headers.

Also add some =head3's for a few missing keywords.

Subsequent commits will flesh out the new sections.

Commit: c0cd43bca2c31a8e93c5e6ea9af7861864ed00f9
https://github.com/Perl/perl5/commit/c0cd43bca2c31a8e93c5e6ea9af7861864ed00f9
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add a new introductory part

Four commits ago, I removed most of the general text sections in
perlxs (i.e. the ones not specifically about a particular keyword).

Now this commit adds a completely new introductory part to perlxs, about
1200 lines long. It represents an attempt to write a background to what
XS and XSUBs, SVs, typemaps etc are, in a complete and modern way.
The existing reference section for each keyword follows it.

I tried to avoid getting too tutorial-like (that's what perlxstut is
for), but I may have crossed the line in various places. In particular
it has a new section which could have been titled "all the bits of
perlguts you need to know in order to write non-trivial XSUBs without
having to actually read perlguts".

Commit: 7088754f0657847522ab4e483736b667d4576e06
https://github.com/Perl/perl5/commit/7088754f0657847522ab4e483736b667d4576e06
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add BNF definition section

Add a section which semi-formally tries to define the syntax and
structue of an XS file, using a BNF-like format.

See http://nntp.perl.org/group/perl.perl5.porters/268701 for the
discussion of this part.

Commit: d07e065215e984be915d4b6dfaa4f0b20e66d76d
https://github.com/Perl/perl5/commit/d07e065215e984be915d4b6dfaa4f0b20e66d76d
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update MODULE/PACKAGE/PREFIX

Rewrite the POD for these three keywords, and in particular, treat
them as one declaration, rather than three unrelated keywords.

Commit: b934cb14a43ea759b4129784c7a110af89ab423c
https://github.com/Perl/perl5/commit/b934cb14a43ea759b4129784c7a110af89ab423c
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update file-scoped directive text

Populate the new

=head2 File-scoped XS Keywords and Directives

section, partially by cannibalising (and then deleting) the old

=head3 Inserting POD, Comments and C Preprocessor Directives

subsection. This commit only adds text about directives; subsequent
commits will update the various file-scoped keywords.

Commit: 68b7aca119ecd2cd566f25d640f239fda59a338b
https://github.com/Perl/perl5/commit/68b7aca119ecd2cd566f25d640f239fda59a338b
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update REQUIRE, VERSIONCHECK keywords

Commit: 261412c92ea524c2a5326166006790cd292b5c39
https://github.com/Perl/perl5/commit/261412c92ea524c2a5326166006790cd292b5c39
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update PROTOTYPES: keyword

Commit: 5bfc7766c1a430033c8be34bbd412279050c8100
https://github.com/Perl/perl5/commit/5bfc7766c1a430033c8be34bbd412279050c8100
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update EXPORT_XSUB_SYMBOLS, INCLUDE(_COMMAND)

Commit: 2d3d34823f2ce229b7f6bd99051a5fd4ef00ed3f
https://github.com/Perl/perl5/commit/2d3d34823f2ce229b7f6bd99051a5fd4ef00ed3f
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update TYPEMAP: keyword

Commit: b518bc737a5e3290f32489d7dd8767aee1cdb11c
https://github.com/Perl/perl5/commit/b518bc737a5e3290f32489d7dd8767aee1cdb11c
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update BOOT: keyword

Commit: 73049754fa3516b64ec1ab4831c735020bcc679f
https://github.com/Perl/perl5/commit/73049754fa3516b64ec1ab4831c735020bcc679f
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update FALLBACK: keyword

Commit: 64cb00d0abfce627d2134e318e5c23dcf2a740bf
https://github.com/Perl/perl5/commit/64cb00d0abfce627d2134e318e5c23dcf2a740bf
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update XSUB Structure + Declaration

Populate the new

=head2 The Structure of an XSUB
=head2 An XSUB Declaration

sections

Commit: e3ea242f73af0be2c0f9579f85f79a63fb42c1c5
https://github.com/Perl/perl5/commit/e3ea242f73af0be2c0f9579f85f79a63fb42c1c5
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update section 'An XSUB Parameter'

Add some initial text for this new section, and also add a new
subsection "XSUB Parameter Placeholders".

Commit: 66c57eb0a573e753b9352948902ebe72cc4bc66b
https://github.com/Perl/perl5/commit/66c57eb0a573e753b9352948902ebe72cc4bc66b
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update IN_OUT etc section

Commit: 286e90fc6e3a071353940dc7596c46fbd51273c9
https://github.com/Perl/perl5/commit/286e90fc6e3a071353940dc7596c46fbd51273c9
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M XSUB.h
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update default, length, ellipis params

Rewrite (and retitle) these three subsections:

=head3 Default Parameter Values
=head3 The C<length(NAME)> Keyword
=head3 Variable-length Parameter Lists

Commit: 86f59bdda7fb7381d783f50aa522402626996be5
https://github.com/Perl/perl5/commit/86f59bdda7fb7381d783f50aa522402626996be5
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: Input part, PREINIT sections

Add text for the new '=head2 The XSUB Input Part' section, and rewrite
the existing entry for the PREINIT keyword.

Commit: 2ab94532396ddf8a5ebd7798d0f5297f6554c2af
https://github.com/Perl/perl5/commit/2ab94532396ddf8a5ebd7798d0f5297f6554c2af
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update 'The INPUT: Keyword' section

This commit completely rewrites this section and subsections:

=head3 The INPUT: Keyword
=head4 The NO_INIT Keyword
=head4 Initializing Function Parameters
=head4 The & Unary Operator

It de-emphasises the INPUT keyword and suggests using ANSI XS signatures
etc instead.

Commit: ef9e62342c6961bf2f39a4614f30d5c620243a0c
https://github.com/Perl/perl5/commit/ef9e62342c6961bf2f39a4614f30d5c620243a0c
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update 'SCOPE: Keyword' section

Commit: 342d0e87c4337f236399c662689809aa5e9a15da
https://github.com/Perl/perl5/commit/342d0e87c4337f236399c662689809aa5e9a15da
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: init part, INIT sections

Add text for the new '=head2 The XSUB Init Part' section, and rewrite
the existing entry for the INIT keyword.

Commit: bd046ce1ad0c43731de527df86a55600910ef353
https://github.com/Perl/perl5/commit/bd046ce1ad0c43731de527df86a55600910ef353
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: code part, autocall, C_ARGS

Add text to the new

=head2 The XSUB Code Part
=head3 Auto-calling a C function

sections, and rewrite the existing

=head4 The C_ARGS: Keyword

section

Commit: 257e27e13ed86d3912c73da850abc323c72c07a2
https://github.com/Perl/perl5/commit/257e27e13ed86d3912c73da850abc323c72c07a2
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: CODE, PPCODE

Rewrite these sections:

=head3 The CODE: Keyword
=head3 The PPCODE: Keyword

Commit: 44b0c500e99fae923935d3d82478e27c1518931c
https://github.com/Perl/perl5/commit/44b0c500e99fae923935d3d82478e27c1518931c
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update NOT_IMPLEMENTED_YET: keyword

This keyword formerly wasn't documented. The docs now say "this is what
it is, but don't use it".

Commit: c0065254f7b0aef8a149d8bf98b3f5210a0098cc
https://github.com/Perl/perl5/commit/c0065254f7b0aef8a149d8bf98b3f5210a0098cc
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: output part

Add text to the new

=head2 The XSUB Output Part

section, and rewrite the text in these existing sections:

=head3 The POSTCALL: Keyword
=head3 The OUTPUT: Keyword

Commit: 027c0006e5a25cfefa3127879a42927b1ed81e16
https://github.com/Perl/perl5/commit/027c0006e5a25cfefa3127879a42927b1ed81e16
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: cleanup part

Add text to the new

=head2 The XSUB Cleanup Part

section, and rewrite the text in this existing section:

=head3 The CLEANUP: Keyword

Commit: c9ef9317973afb3bdbae9dad3bbad17bda1b9f83
https://github.com/Perl/perl5/commit/c9ef9317973afb3bdbae9dad3bbad17bda1b9f83
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update generic intro, PROTOTYPE

Add text to the new

=head2 XSUB Generic Keywords

section, and rewrite the text in this existing section:

=head3 The PROTOTYPE: Keyword

Commit: 815fd708c551410ac81766e80b27940ebb799221
https://github.com/Perl/perl5/commit/815fd708c551410ac81766e80b27940ebb799221
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: mention package name types

Explain that a 'C' parameter type in an XSUB declaration can actually
be a Perl package name or similar, e.g.

Foo::Bar
f(Foo::Bar obj, char *s)

Commit: c9498f597725a2820cb7f32113a0b44a7ae7c780
https://github.com/Perl/perl5/commit/c9498f597725a2820cb7f32113a0b44a7ae7c780
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update OVERLOAD, add T_PTROBJ

First, add a new subsection

=head3 T_PTROBJ and opaque handles

to the TYPEMAPs section explaining how this typemap can be used to
map between Perl objects and C library handles. It provides a
fully-worked example of wrapping a simple arithmetic library.

Then completely rewrite the

=head3 The OVERLOAD: Keyword

section. In particular, it now refers to the new T_PTROBJ example and
shows how it can be extended to use overloading.

Commit: f81709256c762f91c27298e62c90dd996b904bd5
https://github.com/Perl/perl5/commit/f81709256c762f91c27298e62c90dd996b904bd5
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: document ATTRS

This keyword was undocumented, even though it had been added 25 years
ago.

Commit: d716b2e269eed58285f4f64efcb21375aeaf6825
https://github.com/Perl/perl5/commit/d716b2e269eed58285f4f64efcb21375aeaf6825
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add "Sharing XSUB bodies" section

Populate the introduction to this new section.

Commit: 19d1401d728d1124af8b21612a840fa397846401
https://github.com/Perl/perl5/commit/19d1401d728d1124af8b21612a840fa397846401
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: ALIAS

Rewrite this section:

=head3 The ALIAS: Keyword

Commit: 7f027d039364e50dc3dc7451d4a3980bcf69571d
https://github.com/Perl/perl5/commit/7f027d039364e50dc3dc7451d4a3980bcf69571d
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update INTERFACE, INTERFACE_MACRO

Rewrite these sections:

=head3 The INTERFACE: Keyword
=head3 The INTERFACE_MACRO: Keyword

also demote the second to be a head4 child of the first. Then expand
the T_PTROBJ example to use INTERFACE as an alternative to ALIAS.

Commit: df225cdecba56e8ae15a2b49fecf663c5a51cfbe
https://github.com/Perl/perl5/commit/df225cdecba56e8ae15a2b49fecf663c5a51cfbe
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: CASE

Rewrite this section:

=head3 The CASE: Keyword

Commit: 494b7e45afd7639656e77483dcb327c74a295ff7
https://github.com/Perl/perl5/commit/494b7e45afd7639656e77483dcb327c74a295ff7
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add "Using Typemaps" section

Populate this new section (except for the T_PTROBJ subsection, which had
already been added by an earlier commit within this branch).

Note that the "Common typemaps" subsection could probably benefit
from some further expansion by someone familiar with which built-in
T_FOO entries are useful.

Commit: b0ed1804cf9d343c5ae6c478226b13d17ceab248
https://github.com/Perl/perl5/commit/b0ed1804cf9d343c5ae6c478226b13d17ceab248
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod
M t/porting/known_pod_issues.dat

Log Message:
-----------
perlxs.pod: update "Using XS With C++" section

Rewrite this section:

=head2 Using XS With C++

Disclaimer: I've never written a proper C++ program. I had to
(literally) dust off my 34-year old copy of Stroustrup(*) and also do
some Googling. Hopefully what I've written is sane.

(*) This was bought back in the days when people used to to learn things
by buying books, and when I thought that I ought to know something about
this newfangled C++ thing. I never got round to reading all of it: I
discovered Perl around the same time, which looked to be a lot more fun.

Commit: 9a6dfffd6036adbf71be567d4acb8d11a006dec6
https://github.com/Perl/perl5/commit/9a6dfffd6036adbf71be567d4acb8d11a006dec6
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update MY_CXT section

Revise the text in this section:

=head2 Safely Storing Static Data in XS

Commit: dcc8456229ac9489f905153d983a13712b37b442
https://github.com/Perl/perl5/commit/dcc8456229ac9489f905153d983a13712b37b442
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update EXAMPLES section

Rewrite this section:

=head1 EXAMPLES

Basically, delete the one big example in this section and instead
provide links to various other examples already present in this document
instead.

Commit: dedc44bbe5632ff056dcbef73dcab00dcd3a3dca
https://github.com/Perl/perl5/commit/dedc44bbe5632ff056dcbef73dcab00dcd3a3dca
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update CAVEATS, AUTHOR, A DIAGNOSTICS

Tweak the final few sections of perlxs.pod.

Commit: 63ac12a28e5522c694d52a22d2a7b0fe5c8df591
https://github.com/Perl/perl5/commit/63ac12a28e5522c694d52a22d2a7b0fe5c8df591
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
perlxs.pod: standardise version numbers in the doc

After the big rewrite, various bits of text which describe when a
particular feature was introduced or changed have ended up using a
random mixture of Perl, ParseXS and xsubpp version numbers.

This commit standardises on xsubpp version numbers. These are mostly
the same as ParseXS, but this handles the cases before ParseXS was
split off from xsubpp. Perl versions suffer from not exactly matching
when an xsubpp version number was incremented, and not matching what is
used by the "REQUIRE:" keyword.

This commit also adds a short new section which tries to explain how the
three sets of version numbers are related.

Commit: 6abfedb04644697b18271a2c702dde305749f57a
https://github.com/Perl/perl5/commit/6abfedb04644697b18271a2c702dde305749f57a
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs: tweaks following code review

The many commits in this branch have completely rewritten perlxs.pod.
This commit applies all the minor tweaks suggested by reviewers.

(It was far too much like hard work to try and update each individual
commit with the various changes.)

Commit: 45673ef5d75d1e58f576cb8206b96f8e98000710
https://github.com/Perl/perl5/commit/45673ef5d75d1e58f576cb8206b96f8e98000710
Author: David Mitchell <[email protected]>
Date: 2025-12-12 (Fri, 12 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs: more tweaks following code review

The many commits in this branch have completely rewritten perlxs.pod.
This commit applies a second set of minor tweaks suggested by reviewers.

Compare: https://github.com/Perl/perl5/compare/4f3cd1361655...45673ef5d75d

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 55c5b5: RMG - No more update of perlweb for BLEAD-POINT by Thibault Duponchelle via perl5-changes

Thu, 11 Dec 2025 09:57:42 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 55c5b505cb6ecaf868f1326432801de3a67c9943
https://github.com/Perl/perl5/commit/55c5b505cb6ecaf868f1326432801de3a67c9943
Author: Thibault DUPONCHELLE <[email protected]>
Date: 2025-12-11 (Thu, 11 Dec 2025)

Changed paths:
M Porting/release_managers_guide.pod

Log Message:
-----------
RMG - No more update of perlweb for BLEAD-POINT

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 34bbff: RMG - Link reference commit in section Bump theve... by Thibault Duponchelle via perl5-changes

Wed, 10 Dec 2025 16:49:14 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 34bbffc475fff2abb4bf857b8975281ae6dfe5c1
https://github.com/Perl/perl5/commit/34bbffc475fff2abb4bf857b8975281ae6dfe5c1
Author: Thibault DUPONCHELLE <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M Porting/release_managers_guide.pod

Log Message:
-----------
RMG - Link reference commit in section Bump the version number

Co-authored-by: Eric Herman <[email protected]>

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 509e33: mathoms.c: Add a few coments by Karl Williamson via perl5-changes

Wed, 10 Dec 2025 15:50:45 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 509e33b8716f2eef4f4b4d8334de4317531c0bdf
https://github.com/Perl/perl5/commit/509e33b8716f2eef4f4b4d8334de4317531c0bdf
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M mathoms.c

Log Message:
-----------
mathoms.c: Add a few coments

Commit: 9d8dbcc0beda9176883aa1bd6b751def7f7d17e3
https://github.com/Perl/perl5/commit/9d8dbcc0beda9176883aa1bd6b751def7f7d17e3
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M proto.h
M regen/embed.pl

Log Message:
-----------
regen/embed.pl: Don't create prototype for private macros

Prior to this commit, a commented-out prototype was created for all
macros that we have argument and return type information for. This
might be useful information for a reader of proto.h.

This commits stops doing this for private macros. It makes a future
commit slightly easier, and I'm unsure of the usefulness of this anyway.
But I could be persuaded otherwise.

Commit: eac6222943a92d0fe32cd359bc6eb12cba140784
https://github.com/Perl/perl5/commit/eac6222943a92d0fe32cd359bc6eb12cba140784
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M proto.h
M regen/embed.pl

Log Message:
-----------
regen/embed.pl: Don't create prototype for weird macros

Where 'weird' is defined as meaning something where the normal rules
don't apply, so something we generate is unlikely to be correct.

This only affects one element, and it uses aTHX in a way that is
incompatible with it being automated.

Commit: 2444829c88e91c6e4d1f5b1283233a026dea347e
https://github.com/Perl/perl5/commit/2444829c88e91c6e4d1f5b1283233a026dea347e
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.fnc
M proto.h
M regen/embed.pl

Log Message:
-----------
embed.pl: enum and union are like struct in parameters

These keywords all need another word to indicate the parameter type.
Previously only 'struct' was considered to have.

This changed showed an error in one entry embed.fnc, which is also
corrected in this commit.

Commit: 32aaa22eec0a2a4e403b010518b39801679776d4
https://github.com/Perl/perl5/commit/32aaa22eec0a2a4e403b010518b39801679776d4
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.fnc
M embed.h

Log Message:
-----------
embed.fnc: Drop Perl_ on do_aexec my_stat my_lstat

These macros are not for external use, so don't need a Perl_ prefix

Commit: 838b774823624d57f1b2a1369aa0d423e440eb57
https://github.com/Perl/perl5/commit/838b774823624d57f1b2a1369aa0d423e440eb57
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M hv.h
M proto.h

Log Message:
-----------
Move hv_stores() declaration from embed.fnc to hv.h

This is required for the next few commits that start automatically
creating long Perl_name functions for the elements in embed.fnc that are
macros and don't already have them in the source.

Only macros can take a parameter that has to be a literal string, so
don't fit with the next few commits. This is the only case in embed.fnc
like that, so I'm deferring dealing with it for now.

Commit: 3abbf2e1b897a9339dffed87699a0752ef9b629d
https://github.com/Perl/perl5/commit/3abbf2e1b897a9339dffed87699a0752ef9b629d
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.fnc

Log Message:
-----------
embed.fnc: free_c_backtrace is only Perl_free_c_backtrace

Commit: 9df1c067a997d2510c7e898ff968d2c55f939caa
https://github.com/Perl/perl5/commit/9df1c067a997d2510c7e898ff968d2c55f939caa
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.fnc
M proto.h

Log Message:
-----------
Remove const char * const return value

A function can't return something of that type, but this has always been
a macro, so this hasn't been caught.

Commit: 1c0e0dceeb1cc405db3e1d50925e413c27771ac3
https://github.com/Perl/perl5/commit/1c0e0dceeb1cc405db3e1d50925e413c27771ac3
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.fnc
M handy.h
M proto.h

Log Message:
-----------
Three macros are used in perl extensions

They were marked as core-only.

Commit: 2635eab6d1ef7dcd3d7e4ce408f8a30c108f71b4
https://github.com/Perl/perl5/commit/2635eab6d1ef7dcd3d7e4ce408f8a30c108f71b4
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M regen/embed.pl

Log Message:
-----------
regen/embed.pl: Move declaration

Future commits make the new location a better choice

Commit: 3783dac29fd28b8fb3695bbb3b2483833e211737
https://github.com/Perl/perl5/commit/3783dac29fd28b8fb3695bbb3b2483833e211737
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M regen/embed.pl

Log Message:
-----------
regen/embed.pl: Move declaration to first use

Commit: 92dcf59a90bfbb545599098d7043c96abb783ee5
https://github.com/Perl/perl5/commit/92dcf59a90bfbb545599098d7043c96abb783ee5
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.h
M perl.h
M regen/embed.pl

Log Message:
-----------
Gain control of macro namespace visibility

This commit adds the capability to undefine macros that are visible to
XS code but shouldn't be. This can be used to stop macro namespace
pollution by perl.

It works by changing embed.h to have two modes, controlled by a #ifdef
that is set by perl.h. perl.h now #includes embed.h twice. The first
time works as it always has. The second sets the #ifdef, and causes
embed.h to #undef the macros that shouldn't be visible. This call is
just before perl.h returns to its includer, so that these macros have
come and gone before the file that #included perl.h is affected by them.
It comes after the inline headers get included, so they have access to
all the symbols that are defined.

The list of macros is determined by the visibility given by the apidoc
lines documenting them, plus several exception lists that allow a symbol
to be visible even though it is not documented as such.

In this commit, the main exception list contains everything that is
currently visible outside the Perl core, so this should not break any
code. But it means that the visibility control is established for
future changes to our code base. New macros will not be visible except
when documented as needing to be such. We can no longer inadvertently
add new names to pollute the user's.

I expect that over time, the exception list will become smaller, as we
go through it and remove the items that really shouldn't be visible. We
can then see via smoking if someone is actually using them, and either
decide that these should be visible, or work with the module author for
another way to accomplish their needs. (I would hope this would lead to
proper documentation of the ones that need to be visible.)

There are currently four lists of symbols.

One list is for symbols that are used by libc functions, and that Perl
may redefine (usually so that code doesn't have to know if it is running
on a platform that is lacking the given feature.) The algorithm added
here catches most of these and keeps them visible, but there are a few
items that currently must be manually listed.

A second list is of symbols that the re extension to Perl requires, but
no one else needs to. This list is currently empty, as everything
initially is in the main exception list.

A third list is for items that other Perl extensions require, but no one
else needs to. This list is currently empty, as everything initially is
in the main exception list.

The final list is for items that currently are visible to the whole
world. It contains thousands of items. This list should be examined
for:

1) Names that shouldn't be so visible; and
2) Names that need to remain visible but should be changed so they
are less likely to clash with anything the user might come up
with.

I have wanted this ability to happen for a long time; and now things
have come together to enable it.

This allows us to have a clear-cut boundary with CPAN.

It means you can add macros that have internal-only use without having
to worry about making them likely not to clash with user names.

It shows precisely in one place what our names are that are visible to
CPAN.

Commit: b847495def21301a78c7aad050b8d47d51482258
https://github.com/Perl/perl5/commit/b847495def21301a78c7aad050b8d47d51482258
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M regen/HeaderParser.pm
M regen/embed.pl

Log Message:
-----------
Macros guarded by some #ifdef's aren't globally visible

The previous commit undefines macros that aren't supposed to be visible
to XS code. But, to avoid any possible breakage, it creates an
exception list of symbols that may have been visible, and leaves them
so. The goal is to stop the list from growing as new code is developed,
and to shorten the list by various means.

This is the first commit to do that, by looking to see if any symbols
aren't actually externally visible because they are guarded by #ifdef's
that evaluate to false. For example a symbol that is #defined only if
PERL_CORE is defined won't be visible, and need not be on the exception
list.

This cuts almost 30% off the initial list.

Commit: ebbe6ac0f7628af109f6d3fdcc0bc859c8236ce8
https://github.com/Perl/perl5/commit/ebbe6ac0f7628af109f6d3fdcc0bc859c8236ce8
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.h
M regen/embed.pl

Log Message:
-----------
Remove a few more macros from being visible to XS code

These are a few macros dealing with inversion lists that were never
intended to be visible to general XS code, and they actually can't be in
use in cpan because the mechanisms to create inversion lists are private
to perl.

Commit: 7e1ae0c850febc62d152ad00d44c48954a6b66cc
https://github.com/Perl/perl5/commit/7e1ae0c850febc62d152ad00d44c48954a6b66cc
Author: Karl Williamson <[email protected]>
Date: 2025-12-10 (Wed, 10 Dec 2025)

Changed paths:
M embed.h
M regen/embed.pl

Log Message:
-----------
Remove SBOX case statements from external visibility

I'm pretty sure there is no use case for these, and very unlikely to
have any actual uses.

Compare: https://github.com/Perl/perl5/compare/c0b8ab1ec691...7e1ae0c850fe

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 6ae61b: 4 by Karl Williamson via perl5-changes

Mon, 08 Dec 2025 22:20:49 +0000

Branch: refs/heads/smoke-me/khw-scan_num
Home: https://github.com/Perl/perl5
Commit: 6ae61b4213a3f2c5ceeb72178414ab55f60e124c
https://github.com/Perl/perl5/commit/6ae61b4213a3f2c5ceeb72178414ab55f60e124c
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M numeric.c
M proto.h

Log Message:
-----------
4

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 4e0f1e: embed.fnc: Use correct name for formalparameter by Karl Williamson via perl5-changes

Mon, 08 Dec 2025 20:25:20 +0000

Branch: refs/heads/smoke-me/khw-scan_num
Home: https://github.com/Perl/perl5
Commit: 4e0f1ed4b7fb85e604e01e54e0694b17df1fec14
https://github.com/Perl/perl5/commit/4e0f1ed4b7fb85e604e01e54e0694b17df1fec14
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M proto.h

Log Message:
-----------
embed.fnc: Use correct name for formal parameter

The prototype here did not match the actual function

Commit: 02eea7664a2b28c93c1240e1892bf5ce4b41038e
https://github.com/Perl/perl5/commit/02eea7664a2b28c93c1240e1892bf5ce4b41038e
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M perl.h
M proto.h

Log Message:
-----------
Inline Perl_grok_(bin|oct|hex)

Commit: 149403bb583e627a351b3dadb1e8fe117434d546
https://github.com/Perl/perl5/commit/149403bb583e627a351b3dadb1e8fe117434d546
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
XXX Add grok_bin_hex()

grok_bin() and grok_hex() differ from grok_oct() in that they optionally
accept a leading 0x or 0b, based on a flag. Then they all share common
code. This commit splits out the check for the leading 0[bx] into a
separate function.

Commit: 7b31255c4a6d06e287fa34a643f3f76b9c532615
https://github.com/Perl/perl5/commit/7b31255c4a6d06e287fa34a643f3f76b9c532615
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Add two UNLIKELY()s

Underscores in numbers are much less common than digits, and its
unlikely that a number will begin with all zero's and underscores

Commit: db452f0554f4d1f308770ddf2973eaf05b3d22e2
https://github.com/Perl/perl5/commit/db452f0554f4d1f308770ddf2973eaf05b3d22e2
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Add/revise some comments

And reflow. This is in preparation for the next commit.

Commit: 09ceb17bedc6b84877190689a74ef904f3b110b6
https://github.com/Perl/perl5/commit/09ceb17bedc6b84877190689a74ef904f3b110b6
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Speed up overflow detection

We can compute outside the loop the exact value at which the next
iteration wil overflow, saving some operations

Commit: fa84f11a0b0ce9fb6c9627fc25e2b2e651ba2434
https://github.com/Perl/perl5/commit/fa84f11a0b0ce9fb6c9627fc25e2b2e651ba2434
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Improve speed, precision

This replaces a floating multiply each loop iteration with an integer
increment, plus after the loop completes, a call to ldexp() or pow().

Most of the floating multiplies are done on integral values, so
not much precision is lost, but this gets it down to just one
precision-losing operation.

Commit: 7def20c12216eb778aa589f03035db3811ef8568
https://github.com/Perl/perl5/commit/7def20c12216eb778aa589f03035db3811ef8568
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Macroize duplicated code

This is now complicated enough to benefit being placed in a macro
instead of duplicating it.

Commit: 611a8db0aa65cbd916250666d5b6cb61cb6ca30c
https://github.com/Perl/perl5/commit/611a8db0aa65cbd916250666d5b6cb61cb6ca30c
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M perl.h
M pp.c

Log Message:
-----------
Add definitions for [IU}V_BITS to perl.h

replacing the pp.c definition of IV_BITS

Commit: 7fef94ae1af51155c7f40b9d211b2a0fc6adf7ca
https://github.com/Perl/perl5/commit/7fef94ae1af51155c7f40b9d211b2a0fc6adf7ca
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Eke more precision when overflows

The algorithm for dealing with input that overflows a UV is to parse in
batches that don't overflow. This means when we would otherwise
overflow the UV, we store the unoverflowed value in an NV, and start
over with the next batch, combining that with the first batch at the end
or when it too overflows. This minimizes the number of floating point
operations.

But on platforms where a UV contains more bits than an NV significand,
that first batch when stored loses precision. This is typical on a
64-bit machine using IEEE 754 double precision NVs, which have 53 bits
of precision.

This precision loss can be avoided by stopping accumulating the first
batch just before it would overflow the significand. That effectively
is what this commit does.

Actually it is a bit more complicated. Instead of stopping there, it
adds a checkpoint and goes on. If the final result doesn't overflow the
UV, the checkpoint is simply ignored. But if it does overflow, the code
returns to the checkpoint and creates a first batch from the data as of
there, and reparses those typically few bytes of input (11 bits is 2-3
hex digits) as the first bytes of the second batch.

Commit: 3db8d12edb73e5286b43b4ea89b2d5dbfab2754a
https://github.com/Perl/perl5/commit/3db8d12edb73e5286b43b4ea89b2d5dbfab2754a
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Move error handling to outside loop

This makes the loop occupy less memory, but it's mainly for the next
commit which will remove a conditional each time through the loop, and
the tests want the error messages displayed in a particular order.

Commit: dd5e2a76b1b61c1370a7b5c739ddad62269a1ee6
https://github.com/Perl/perl5/commit/dd5e2a76b1b61c1370a7b5c739ddad62269a1ee6
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
debug

Commit: a27b879dc1697f9fdd0a89f6dde858a973856a8c
https://github.com/Perl/perl5/commit/a27b879dc1697f9fdd0a89f6dde858a973856a8c
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
rm debug

Commit: fdc0e2641ed70e6a018eb35bb4c635f1e4361fb4
https://github.com/Perl/perl5/commit/fdc0e2641ed70e6a018eb35bb4c635f1e4361fb4
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Move overflow notification outside loop

There are several code section that deal with overflow; this allows some
of them to be combined, in future commits

Commit: be2c7f40d2c55f0a2b678c11fe2da997d7a5d932
https://github.com/Perl/perl5/commit/be2c7f40d2c55f0a2b678c11fe2da997d7a5d932
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Combine two overflow handling sections

Commit: 576ce40db2462def08e82070b612d2d5b2f5105f
https://github.com/Perl/perl5/commit/576ce40db2462def08e82070b612d2d5b2f5105f
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Cleanup the function end code

This rationalizes a bit more the code at the end of the function to make
it a bit more readable.

Commit: 5a9d382210fab2ce5a5668647865d5a085e8a80a
https://github.com/Perl/perl5/commit/5a9d382210fab2ce5a5668647865d5a085e8a80a
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Some more function termination cleanup

This moves some code into an else, and replaces it with an UNLIKELY.

That could be eliminated if we didn't care about the order the warning
messages are raised, but the test suite does test for a particular
order, so I'm leaving it as-is.

Commit: f944ab905fc119159af965002522dc3e6b04cfce
https://github.com/Perl/perl5/commit/f944ab905fc119159af965002522dc3e6b04cfce
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M inline.h
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Add ability to discard not overflow

This will be used in a future commit

Commit: 264257aa033ebfba1c3a8028d1fe41d1718ec506
https://github.com/Perl/perl5/commit/264257aa033ebfba1c3a8028d1fe41d1718ec506
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Macroize common expression

This is in preparation for a future commit where the expression will
change, and this way, we can explain it in one place

Commit: 2aa3b02ba49745eb63531557bdd15d2a61ea7b2f
https://github.com/Perl/perl5/commit/2aa3b02ba49745eb63531557bdd15d2a61ea7b2f
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_oct_bin_hex: Change OR to addition

Currently an OR is all that is needed as a shift is used to make space
for the next digit. But a future commit will change that so that we
will need to add the new digit, not just OR it

Commit: f69a4a158a4eadf6b9e7fa24e87fcd7b2c5a36a1
https://github.com/Perl/perl5/commit/f69a4a158a4eadf6b9e7fa24e87fcd7b2c5a36a1
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
shift1,2

Commit: d4a1145ba31b99d59c99b2354cb29fcdb091981b
https://github.com/Perl/perl5/commit/d4a1145ba31b99d59c99b2354cb29fcdb091981b
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Move declaration

A future commit will want it earlier

Commit: f1c312542abede9beb028d6c8c438758bc558045
https://github.com/Perl/perl5/commit/f1c312542abede9beb028d6c8c438758bc558045
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Save expression in a variable

The next commit will reference this in multiple places

Commit: b107a3fe8360ecf5c2380af8dc03abd9ed7151d6
https://github.com/Perl/perl5/commit/b107a3fe8360ecf5c2380af8dc03abd9ed7151d6
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Also accept decimal

Commit: 2a78847ca112ddf63bda18aaa0615ca4ef4300c1
https://github.com/Perl/perl5/commit/2a78847ca112ddf63bda18aaa0615ca4ef4300c1
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
use-decimal

Commit: a9728e86876d13e889e0a21e9a530f74053c46cb
https://github.com/Perl/perl5/commit/a9728e86876d13e889e0a21e9a530f74053c46cb
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
Rename grok_bin_oct_hex()

It now also accepts decimal, so that name no longer is appropriate.
grok_uint_by_base() is what I chose.

Commit: ea27df71a180fab982de6d472418ab5da834ef50
https://github.com/Perl/perl5/commit/ea27df71a180fab982de6d472418ab5da834ef50
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
XXX fnc changed earlier grok_uint_by_base: Change name of formal parameter

The new name is clearer that this is passed to a function to look
something up.

Commit: 17f8a602d0c057992d95d7ef3463beb9438153f4
https://github.com/Perl/perl5/commit/17f8a602d0c057992d95d7ef3463beb9438153f4
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Combine an if with its else

By changing the conditional, the two branches can be combined to avoid
some duplication.

Commit: 0996159d1e6a4827d3911db45b75170618729008
https://github.com/Perl/perl5/commit/0996159d1e6a4827d3911db45b75170618729008
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Clean up loop some

As the comments say, this loop is "wonky". But it doesn't have to be as
wonky as it is. This commit reverses the meaning of some conditionals
so that when true they exit the loop. And this allows the removal of a
'break' just at the end of the loop scope, making the loop much more
conventional.

In reversing the meaning, I split some off into more conditionals. This
makes it easier to understand, and prepares for a future commit.

Commit: 25be079205ecdf63b662a8fd0f9fd327f85f64ab
https://github.com/Perl/perl5/commit/25be079205ecdf63b662a8fd0f9fd327f85f64ab
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Move declaration until needed

This code may not need to get executed; so move it until it is needed.

Commit: c182b3d6260c9a009330dbeba0a7100f0d5eb3c1
https://github.com/Perl/perl5/commit/c182b3d6260c9a009330dbeba0a7100f0d5eb3c1
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Optimize for more likely case

This changes the code to prefer the case where the input is a string
without underscores that contains 8 or fewer digits in the given base.
Doing so eliminates two conditionals from what I think is what the vast
majority of calls to this function look like.

This is accomplished in this commit by making the default: case for at
least 9 digits and adds a case: for 8. There is no need for anything
then but to return for any string 8 or fewer digit characters long.

The default does try to handle leading zeros efficiently by having a
tight loop to strip them off.

Commit: 3263805341152b3464e8b354e19f247a41699f7c
https://github.com/Perl/perl5/commit/3263805341152b3464e8b354e19f247a41699f7c
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Round discarded digits

When remaining digits are discarded instead of overflowing, this commit
causes the returned non-overflowed value to be rounded (towards even)
instead of truncated.

This will be useful when future commits cause this function to be called
to convert a string of numbers following a decimal point. Instead of
truncating the final significant value, it will be rounded.

Commit: b1180f1606d35ebd487ea5e88dd90f47975a6e4e
https://github.com/Perl/perl5/commit/b1180f1606d35ebd487ea5e88dd90f47975a6e4e
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
XXX grok_uint_by_base: Outdent lines

Commit: 518058522419c6a1f0949f017135d51dea56d893
https://github.com/Perl/perl5/commit/518058522419c6a1f0949f017135d51dea56d893
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M inline.h
M numeric.c

Log Message:
-----------
grok_uint_by_base: Allow sequential underscores

For one future use of this function, multiple consecutive underscores
are tolerated. This prepares this function for this use.

Commit: 78c9bbb5ffc5137b917f768fd9bccbaa63cf1f59
https://github.com/Perl/perl5/commit/78c9bbb5ffc5137b917f768fd9bccbaa63cf1f59
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
grok_uint_by_base: Allow caller to specify overflow bits

Usually this will be UV_BITS to allow something up to UV_MAX, but a
future use will want this to be able to be smaller

Commit: eee67839e43327c6d8c69124d1960f35bbe7e9bb
https://github.com/Perl/perl5/commit/eee67839e43327c6d8c69124d1960f35bbe7e9bb
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
pod

Commit: ca3cfc43023bf20c2f68cbd66a1e0215beec0919
https://github.com/Perl/perl5/commit/ca3cfc43023bf20c2f68cbd66a1e0215beec0919
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
digit-count

Commit: 977690d67679b0d4ac1bf867c14982766ed9024c
https://github.com/Perl/perl5/commit/977690d67679b0d4ac1bf867c14982766ed9024c
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
XXX unsure LIKELY

Commit: b6f2c7f86e75837b38c7bfb89e6def5e94577697
https://github.com/Perl/perl5/commit/b6f2c7f86e75837b38c7bfb89e6def5e94577697
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c
M perl.h

Log Message:
-----------
notes

Commit: bb19d7e9b0729700eca39745c3ba30fc8504482a
https://github.com/Perl/perl5/commit/bb19d7e9b0729700eca39745c3ba30fc8504482a
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M numeric.c
M proto.h

Log Message:
-----------
uintmax_t

Commit: 134774c9b5c017f70580f3da31a01530acc85169
https://github.com/Perl/perl5/commit/134774c9b5c017f70580f3da31a01530acc85169
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
toke.c: scan_num: %c format to print a non digit byte

This panic croak was wrong, should it ever get executed. Any byte that
reaches it isn't a digit, so %d would give wrong results.

Commit: 97e08e073eedd4b5effbb1bc6d381ced8e5cbc60
https://github.com/Perl/perl5/commit/97e08e073eedd4b5effbb1bc6d381ced8e5cbc60
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Reformat/fix comments

These comments that gave complicated regex patterns that show the syntax
of the various types of numerical constants recognized by this function
had some oversights in them, and were hard to read.

Commit: f6c0b5e190ed4b8a9a8e0d09aa6dc1ff0563acd9
https://github.com/Perl/perl5/commit/f6c0b5e190ed4b8a9a8e0d09aa6dc1ff0563acd9
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Reorder case: statements in switch()

Move the default to the end, and the shortest to the beginning. This is
in preparation for future commits.

Commit: e98ba497ea1df71ba605c09a133aa28c19b7a519
https://github.com/Perl/perl5/commit/e98ba497ea1df71ba605c09a133aa28c19b7a519
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Convert switch() to elsif series

By using isDIGIT(), the cases for individual digits 1-9 collapse into a
single one, leaving just three possibilities, which are more clearly
handled by an if and two 'else if's

Commit: 2ccb12760275c990b4a5bb638598d42e078dbcfa
https://github.com/Perl/perl5/commit/2ccb12760275c990b4a5bb638598d42e078dbcfa
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: indentation, reorder-comments only

Some blocks have been removed, so can outdent; others will be added in
future commits, so indent.

Commit: 9a1a19b66b5dafd9f1c6d1af108148376b6b51ee
https://github.com/Perl/perl5/commit/9a1a19b66b5dafd9f1c6d1af108148376b6b51ee
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
toke.c: Convert do/while to modern STMT_START/END

Commit: 7a3e8493bc776081a0c9f74c2555a35c01ce0acd
https://github.com/Perl/perl5/commit/7a3e8493bc776081a0c9f74c2555a35c01ce0acd
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M numeric.c
M proto.h

Log Message:
-----------
numeric.c: Make output_nonportable() callable from core

Instead of being internal to this file.

Commit: ac08fae4726ac8dbc6468e3dbd98ad1e7f5e8a8c
https://github.com/Perl/perl5/commit/ac08fae4726ac8dbc6468e3dbd98ad1e7f5e8a8c
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Replace code by equivalent function call

The previous commit has made this function, long in numeric.c,
available to the rest of core. The code here removed duplicated what it
does. Two variables are now unused, and are removed.

Commit: 1554f421529bf41feca8d71ffaac613854ec7742
https://github.com/Perl/perl5/commit/1554f421529bf41feca8d71ffaac613854ec7742
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Macroize common code

This avoids repeating code snippets

Commit: c2678c87006881b63600e9a9e628e6caf95786b4
https://github.com/Perl/perl5/commit/c2678c87006881b63600e9a9e628e6caf95786b4
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Remove now irrelevant code

The previous commit created macros to handle underscores in a numeric
constant. They changed things so adjacent underscores are all absorbed
at once (and warned about). That means we no longer have to keep track
of if the previous character was an underscore.

Commit: 32ca2f8ae8631fdd0f127ef9fb426d3021cb168a
https://github.com/Perl/perl5/commit/32ca2f8ae8631fdd0f127ef9fb426d3021cb168a
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Remove duplicated logic

The previous commit enhanced grok_bin_oct_hex to handle multiple
consecutive underscores, which is all that was missing for this code to
be able to call it instead of doing the same task itself.

This saves quite a bit of code

Commit: 2a7485c510947bbcae6e3ef43c9bfcb1ef7837e7
https://github.com/Perl/perl5/commit/2a7485c510947bbcae6e3ef43c9bfcb1ef7837e7
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Extract common code to a single place

Commit: d6fba6b39a57203870bab8028f4865ceb359cc2e
https://github.com/Perl/perl5/commit/d6fba6b39a57203870bab8028f4865ceb359cc2e
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
DEBUG

Commit: b9de8bd8ed25c4d3c9978c90624e534b0865bad0
https://github.com/Perl/perl5/commit/b9de8bd8ed25c4d3c9978c90624e534b0865bad0
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
XXX Make nonportable a switch

Commit: edda673029c2801e8ad62bc452ff6017ebe0a5ef
https://github.com/Perl/perl5/commit/edda673029c2801e8ad62bc452ff6017ebe0a5ef
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
white after grok_num

Commit: d679e668240305a6014f4e98593531690c2099a6
https://github.com/Perl/perl5/commit/d679e668240305a6014f4e98593531690c2099a6
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
frac

Commit: 36ab005f1a7ca51c484aa9a6db1239c71cbefa9b
https://github.com/Perl/perl5/commit/36ab005f1a7ca51c484aa9a6db1239c71cbefa9b
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M embed.fnc
M handy.h
M inline.h
M numeric.c
M pod/perldiag.pod
M proto.h
M t/op/hexfp.t
M t/op/oct.t
M toke.c

Log Message:
-----------
trial

Commit: 46d9961515cabd29b3accf9a6b3a6ab01ca4c5f9
https://github.com/Perl/perl5/commit/46d9961515cabd29b3accf9a6b3a6ab01ca4c5f9
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M pod/perldata.pod

Log Message:
-----------
perldata: Add details to hex, etc floats

Commit: 63ce45dce9ea747235bb355387f02c39ef96f71e
https://github.com/Perl/perl5/commit/63ce45dce9ea747235bb355387f02c39ef96f71e
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
f

Commit: 5bd2008b9a7effcc5247f52577e3565c1192f11f
https://github.com/Perl/perl5/commit/5bd2008b9a7effcc5247f52577e3565c1192f11f
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M Configure
M config_h.SH
M metaconfig.h

Log Message:
-----------
Configure

Commit: 76b758e070597e12ed20112459276e19c1b7099e
https://github.com/Perl/perl5/commit/76b758e070597e12ed20112459276e19c1b7099e
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M handy.h
M perl.h
M toke.c

Log Message:
-----------
2

Commit: 784cf63b12f35f3118f3beac8d695b5b64eece6b
https://github.com/Perl/perl5/commit/784cf63b12f35f3118f3beac8d695b5b64eece6b
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M Configure

Log Message:
-----------
f

Commit: aba63fe5ffa1721f4c85648fa044b84289f957b2
https://github.com/Perl/perl5/commit/aba63fe5ffa1721f4c85648fa044b84289f957b2
Author: Karl Williamson <[email protected]>
Date: 2025-12-08 (Mon, 08 Dec 2025)

Changed paths:
M Configure
M Cross/config.sh-arm-linux
M Cross/config.sh-arm-linux-n770
M Porting/config.sh
M configure.com
M plan9/config_sh.sample
M win32/config.gc
M win32/config.vc

Log Message:
-----------
3

Compare: https://github.com/Perl/perl5/compare/c9a5117dbd72...aba63fe5ffa1

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] c0b8ab: Grammatical corrections by James E Keenan via perl5-changes

Mon, 08 Dec 2025 02:30:05 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: c0b8ab1ec6910f423d6cf550207f108b32af8beb
https://github.com/Perl/perl5/commit/c0b8ab1ec6910f423d6cf550207f108b32af8beb
Author: James E Keenan <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M README.openbsd

Log Message:
-----------
Grammatical corrections

As reported by hmartink.

Fixes: GH #23933

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 72f140: embed.fnc: Use correct name for formalparameter by Karl Williamson via perl5-changes

Sun, 07 Dec 2025 19:19:50 +0000

Branch: refs/heads/smoke-me/khw-scan_num
Home: https://github.com/Perl/perl5
Commit: 72f140eed6b84de675a3e221cb055bfaabe07e27
https://github.com/Perl/perl5/commit/72f140eed6b84de675a3e221cb055bfaabe07e27
Author: Karl Williamson <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M embed.fnc
M proto.h

Log Message:
-----------
embed.fnc: Use correct name for formal parameter

The prototype here did not match the actual function

Commit: 5a881ccb91c3cefa724301ce0607b9c8c4889f43
https://github.com/Perl/perl5/commit/5a881ccb91c3cefa724301ce0607b9c8c4889f43
Author: Karl Williamson <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M perl.h
M proto.h

Log Message:
-----------
Inline Perl_grok_(bin|oct|hex)

Commit: 56255e4219d9f3083b4f708b9230572e6e0ff9a9
https://github.com/Perl/perl5/commit/56255e4219d9f3083b4f708b9230572e6e0ff9a9
Author: Karl Williamson <[email protected]>
Date: 2025-12-06 (Sat, 06 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
XXX Add grok_bin_hex()

grok_bin() and grok_hex() differ from grok_oct() in that they optionally
accept a leading 0x or 0b, based on a flag. Then they all share common
code. This commit splits out the check for the leading 0[bx] into a
separate function.

Commit: 305de213cb4e88ca282c6e224091f0b63673d4d3
https://github.com/Perl/perl5/commit/305de213cb4e88ca282c6e224091f0b63673d4d3
Author: Karl Williamson <[email protected]>
Date: 2025-12-06 (Sat, 06 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Add two UNLIKELY()s

Underscores in numbers are much less common than digits, and its
unlikely that a number will begin with all zero's and underscores

Commit: 05a85b46974ded881cedaa86b1272ccc1d3c14e9
https://github.com/Perl/perl5/commit/05a85b46974ded881cedaa86b1272ccc1d3c14e9
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Add/revise some comments

And reflow. This is in preparation for the next commit.

Commit: 77cb443a431b909a186ead557672cb98b8974601
https://github.com/Perl/perl5/commit/77cb443a431b909a186ead557672cb98b8974601
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Speed up overflow detection

We can compute outside the loop the exact value at which the next
iteration wil overflow, saving some operations

Commit: 4814771d47ef99e54b5fd1480fd06307e7a6757b
https://github.com/Perl/perl5/commit/4814771d47ef99e54b5fd1480fd06307e7a6757b
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Improve speed, precision

This replaces a floating multiply each loop iteration with an integer
increment, plus after the loop completes, a call to ldexp() or pow().

Most of the floating multiplies are done on integral values, so
not much precision is lost, but this gets it down to just one
precision-losing operation.

Commit: 3c5324b605b12798c870cb1a587ca470a37c0780
https://github.com/Perl/perl5/commit/3c5324b605b12798c870cb1a587ca470a37c0780
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Macroize duplicated code

This is now complicated enough to benefit being placed in a macro
instead of duplicating it.

Commit: 32be410fb67785f7e54043c02f41271c3f37f826
https://github.com/Perl/perl5/commit/32be410fb67785f7e54043c02f41271c3f37f826
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M perl.h
M pp.c

Log Message:
-----------
Add definitions for [IU}V_BITS to perl.h

replacing the pp.c definition of IV_BITS

Commit: 3d56e5df5d195773a22394ddf1e122b6ec14a461
https://github.com/Perl/perl5/commit/3d56e5df5d195773a22394ddf1e122b6ec14a461
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Eke more precision when overflows

The algorithm for dealing with input that overflows a UV is to parse in
batches that don't overflow. This means when we would otherwise
overflow the UV, we store the unoverflowed value in an NV, and start
over with the next batch, combining that with the first batch at the end
or when it too overflows. This minimizes the number of floating point
operations.

But on platforms where a UV contains more bits than an NV significand,
that first batch when stored loses precision. This is typical on a
64-bit machine using IEEE 754 double precision NVs, which have 53 bits
of precision.

This precision loss can be avoided by stopping accumulating the first
batch just before it would overflow the significand. That effectively
is what this commit does.

Actually it is a bit more complicated. Instead of stopping there, it
adds a checkpoint and goes on. If the final result doesn't overflow the
UV, the checkpoint is simply ignored. But if it does overflow, the code
returns to the checkpoint and creates a first batch from the data as of
there, and reparses those typically few bytes of input (11 bits is 2-3
hex digits) as the first bytes of the second batch.

Commit: 709f38456972b69506ec179c29c8a841b22fdcdb
https://github.com/Perl/perl5/commit/709f38456972b69506ec179c29c8a841b22fdcdb
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Move error handling to outside loop

This makes the loop occupy less memory, but it's mainly for the next
commit which will remove a conditional each time through the loop, and
the tests want the error messages displayed in a particular order.

Commit: 30cf2e57b835e8a903c9f9cd59d0c6049b3cf9cd
https://github.com/Perl/perl5/commit/30cf2e57b835e8a903c9f9cd59d0c6049b3cf9cd
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
debug

Commit: 61bbd5b0399b6a5e9fc9d430bef7315f45b9c663
https://github.com/Perl/perl5/commit/61bbd5b0399b6a5e9fc9d430bef7315f45b9c663
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
rm debug

Commit: f593c80ab7e5f70a233401c3b96cf03edfcfcd4f
https://github.com/Perl/perl5/commit/f593c80ab7e5f70a233401c3b96cf03edfcfcd4f
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Move overflow notification outside loop

There are several code section that deal with overflow; this allows some
of them to be combined, in future commits

Commit: c43ff3cbd4bacb39d92c078b0fab1f7be9403461
https://github.com/Perl/perl5/commit/c43ff3cbd4bacb39d92c078b0fab1f7be9403461
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Combine two overflow handling sections

Commit: ac20a27bc48083be60349c44a67d0ea9f4933613
https://github.com/Perl/perl5/commit/ac20a27bc48083be60349c44a67d0ea9f4933613
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Cleanup the function end code

This rationalizes a bit more the code at the end of the function to make
it a bit more readable.

Commit: 13263e18956650b9e6d0b8d9d9c89b5eb707e303
https://github.com/Perl/perl5/commit/13263e18956650b9e6d0b8d9d9c89b5eb707e303
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Some more function termination cleanup

This moves some code into an else, and replaces it with an UNLIKELY.

That could be eliminated if we didn't care about the order the warning
messages are raised, but the test suite does test for a particular
order, so I'm leaving it as-is.

Commit: 9ad2906170fdfa53de4240f006023188841d45cd
https://github.com/Perl/perl5/commit/9ad2906170fdfa53de4240f006023188841d45cd
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M inline.h
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Add ability to discard not overflow

This will be used in a future commit

Commit: 63f6281f137042ea04dfa37e65f619785fdb46a1
https://github.com/Perl/perl5/commit/63f6281f137042ea04dfa37e65f619785fdb46a1
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Macroize common expression

This is in preparation for a future commit where the expression will
change, and this way, we can explain it in one place

Commit: a1d81ab2f64cd94c0b3217397cffb86d60833c89
https://github.com/Perl/perl5/commit/a1d81ab2f64cd94c0b3217397cffb86d60833c89
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_oct_bin_hex: Change OR to addition

Currently an OR is all that is needed as a shift is used to make space
for the next digit. But a future commit will change that so that we
will need to add the new digit, not just OR it

Commit: c625d4402d93dc883784d8369c291d1a47d678a8
https://github.com/Perl/perl5/commit/c625d4402d93dc883784d8369c291d1a47d678a8
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
shift1,2

Commit: 7f93a9a38109ee2cdf42ca9229960fc0b832f985
https://github.com/Perl/perl5/commit/7f93a9a38109ee2cdf42ca9229960fc0b832f985
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Move declaration

A future commit will want it earlier

Commit: 26a69b88c21e9437d51f34af0ea3b558eac0dc3a
https://github.com/Perl/perl5/commit/26a69b88c21e9437d51f34af0ea3b558eac0dc3a
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Save expression in a variable

The next commit will reference this in multiple places

Commit: 6a177879f1e181ce3c93c55fdcc588194e515c83
https://github.com/Perl/perl5/commit/6a177879f1e181ce3c93c55fdcc588194e515c83
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_bin_oct_hex: Also accept decimal

Commit: ac3c225c53dc76122e24a43821ed503809ff931c
https://github.com/Perl/perl5/commit/ac3c225c53dc76122e24a43821ed503809ff931c
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
use-decimal

Commit: d76a59e1ef7199d8c2c87a92597b6d2dcd4bdf04
https://github.com/Perl/perl5/commit/d76a59e1ef7199d8c2c87a92597b6d2dcd4bdf04
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
Rename grok_bin_oct_hex()

It now also accepts decimal, so that name no longer is appropriate.
grok_uint_by_base() is what I chose.

Commit: d6d07bc0c3495fd6d27afffddc4606bcc02d4d6c
https://github.com/Perl/perl5/commit/d6d07bc0c3495fd6d27afffddc4606bcc02d4d6c
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
XXX fnc changed earlier grok_uint_by_base: Change name of formal parameter

The new name is clearer that this is passed to a function to look
something up.

Commit: 794d17d4da1adc34f54ee3e6e0c49234dbc6909d
https://github.com/Perl/perl5/commit/794d17d4da1adc34f54ee3e6e0c49234dbc6909d
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Combine an if with its else

By changing the conditional, the two branches can be combined to avoid
some duplication.

Commit: fd0431973f389d56a59bf2657dd3aded79d0da96
https://github.com/Perl/perl5/commit/fd0431973f389d56a59bf2657dd3aded79d0da96
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Clean up loop some

As the comments say, this loop is "wonky". But it doesn't have to be as
wonky as it is. This commit reverses the meaning of some conditionals
so that when true they exit the loop. And this allows the removal of a
'break' just at the end of the loop scope, making the loop much more
conventional.

In reversing the meaning, I split some off into more conditionals. This
makes it easier to understand, and prepares for a future commit.

Commit: 2c298d9bbd4eff08a761f0c9d744b5691df082f5
https://github.com/Perl/perl5/commit/2c298d9bbd4eff08a761f0c9d744b5691df082f5
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Move declaration until needed

This code may not need to get executed; so move it until it is needed.

Commit: 8600e9cf5c33ce038b8721ea3ef82981842407a2
https://github.com/Perl/perl5/commit/8600e9cf5c33ce038b8721ea3ef82981842407a2
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Optimize for more likely case

This changes the code to prefer the case where the input is a string
without underscores that contains 8 or fewer digits in the given base.
Doing so eliminates two conditionals from what I think is what the vast
majority of calls to this function look like.

This is accomplished in this commit by making the default: case for at
least 9 digits and adds a case: for 8. There is no need for anything
then but to return for any string 8 or fewer digit characters long.

The default does try to handle leading zeros efficiently by having a
tight loop to strip them off.

Commit: a4c63ff56030a5d97865efd014e61e216ae281aa
https://github.com/Perl/perl5/commit/a4c63ff56030a5d97865efd014e61e216ae281aa
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
grok_uint_by_base: Round discarded digits

When remaining digits are discarded instead of overflowing, this commit
causes the returned non-overflowed value to be rounded (towards even)
instead of truncated.

This will be useful when future commits cause this function to be called
to convert a string of numbers following a decimal point. Instead of
truncating the final significant value, it will be rounded.

Commit: 0870a9f212fbbcda56ddcb173ccce671ba030ce5
https://github.com/Perl/perl5/commit/0870a9f212fbbcda56ddcb173ccce671ba030ce5
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
XXX grok_uint_by_base: Outdent lines

Commit: 97cf71e197ccc027d4add3d749a10ce2218e5478
https://github.com/Perl/perl5/commit/97cf71e197ccc027d4add3d749a10ce2218e5478
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M inline.h
M numeric.c

Log Message:
-----------
grok_uint_by_base: Allow sequential underscores

For one future use of this function, multiple consecutive underscores
are tolerated. This prepares this function for this use.

Commit: 9c37b7083b6a0694f4e72e28173113489824e2a3
https://github.com/Perl/perl5/commit/9c37b7083b6a0694f4e72e28173113489824e2a3
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
grok_uint_by_base: Allow caller to specify overflow bits

Usually this will be UV_BITS to allow something up to UV_MAX, but a
future use will want this to be able to be smaller

Commit: 2f0a21dcb50ed3a708fbbcf3ae4e67dfa84dc68f
https://github.com/Perl/perl5/commit/2f0a21dcb50ed3a708fbbcf3ae4e67dfa84dc68f
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
pod

Commit: ee788aa702dec627374585fe693264f49ac711d2
https://github.com/Perl/perl5/commit/ee788aa702dec627374585fe693264f49ac711d2
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M inline.h
M numeric.c
M proto.h

Log Message:
-----------
digit-count

Commit: 3c25553c071e52236752c215129f0906b3658504
https://github.com/Perl/perl5/commit/3c25553c071e52236752c215129f0906b3658504
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
XXX unsure LIKELY

Commit: 4cf2f814af461900f37a801a53a66c1a77ac36b2
https://github.com/Perl/perl5/commit/4cf2f814af461900f37a801a53a66c1a77ac36b2
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c
M perl.h

Log Message:
-----------
notes

Commit: e10c82ae9208780331f7c0c00485fd7f54e8dd4a
https://github.com/Perl/perl5/commit/e10c82ae9208780331f7c0c00485fd7f54e8dd4a
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M embed.fnc
M numeric.c
M proto.h

Log Message:
-----------
uintmax_t

Commit: aee7b559bc5cb0749b7a157a8091bc0c2a7d7ef1
https://github.com/Perl/perl5/commit/aee7b559bc5cb0749b7a157a8091bc0c2a7d7ef1
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
toke.c: scan_num: %c format to print a non digit byte

This panic croak was wrong, should it ever get executed. Any byte that
reaches it isn't a digit, so %d would give wrong results.

Commit: 6c7767e26faa9e89a49049775b20ff117be2db95
https://github.com/Perl/perl5/commit/6c7767e26faa9e89a49049775b20ff117be2db95
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Reformat/fix comments

These comments that gave complicated regex patterns that show the syntax
of the various types of numerical constants recognized by this function
had some oversights in them, and were hard to read.

Commit: 0e377d98b2c03a48dbd7ba60d5ba4ce68b63ec6b
https://github.com/Perl/perl5/commit/0e377d98b2c03a48dbd7ba60d5ba4ce68b63ec6b
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Reorder case: statements in switch()

Move the default to the end, and the shortest to the beginning. This is
in preparation for future commits.

Commit: 95668c87e9c5bdebd9ecd11362fced4439e2df12
https://github.com/Perl/perl5/commit/95668c87e9c5bdebd9ecd11362fced4439e2df12
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Convert switch() to elsif series

By using isDIGIT(), the cases for individual digits 1-9 collapse into a
single one, leaving just three possibilities, which are more clearly
handled by an if and two 'else if's

Commit: 02f446f25ef4c08852020268439d951b5690103a
https://github.com/Perl/perl5/commit/02f446f25ef4c08852020268439d951b5690103a
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: indentation, reorder-comments only

Some blocks have been removed, so can outdent; others will be added in
future commits, so indent.

Commit: d376d2e4e87f57b05de5573613f9a9cb037e2346
https://github.com/Perl/perl5/commit/d376d2e4e87f57b05de5573613f9a9cb037e2346
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
toke.c: Convert do/while to modern STMT_START/END

Commit: e3bff16824dd5dbcc1bd464dd859d76300142705
https://github.com/Perl/perl5/commit/e3bff16824dd5dbcc1bd464dd859d76300142705
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M embed.fnc
M embed.h
M numeric.c
M proto.h

Log Message:
-----------
numeric.c: Make output_nonportable() callable from core

Instead of being internal to this file.

Commit: c79459e9eb5c72b2ed17c8f9be461bcc4a917212
https://github.com/Perl/perl5/commit/c79459e9eb5c72b2ed17c8f9be461bcc4a917212
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Replace code by equivalent function call

The previous commit has made this function, long in numeric.c,
available to the rest of core. The code here removed duplicated what it
does. Two variables are now unused, and are removed.

Commit: eb8e56b9d855060356963e82089844afb5e65b87
https://github.com/Perl/perl5/commit/eb8e56b9d855060356963e82089844afb5e65b87
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Macroize common code

This avoids repeating code snippets

Commit: 0b547afb5e35c1880abc1b3cea4aff1516fd741d
https://github.com/Perl/perl5/commit/0b547afb5e35c1880abc1b3cea4aff1516fd741d
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Remove now irrelevant code

The previous commit created macros to handle underscores in a numeric
constant. They changed things so adjacent underscores are all absorbed
at once (and warned about). That means we no longer have to keep track
of if the previous character was an underscore.

Commit: 8a1d85eab2e555e4ed5699b0e177a67523f58620
https://github.com/Perl/perl5/commit/8a1d85eab2e555e4ed5699b0e177a67523f58620
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Remove duplicated logic

The previous commit enhanced grok_bin_oct_hex to handle multiple
consecutive underscores, which is all that was missing for this code to
be able to call it instead of doing the same task itself.

This saves quite a bit of code

Commit: 50836729e4e6d070ba0729d16d4cb48d41d1ad87
https://github.com/Perl/perl5/commit/50836729e4e6d070ba0729d16d4cb48d41d1ad87
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
scan_num: Extract common code to a single place

Commit: 2fe9323b032fbc28e3f03ee6dc4c926dd6fc99ca
https://github.com/Perl/perl5/commit/2fe9323b032fbc28e3f03ee6dc4c926dd6fc99ca
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
DEBUG

Commit: 0a4fc0b11e375565fae8d07605b5be5124dc8958
https://github.com/Perl/perl5/commit/0a4fc0b11e375565fae8d07605b5be5124dc8958
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
XXX Make nonportable a switch

Commit: 7b4f2c8c0daab6ada1dca83f1c6db3a8b9be794d
https://github.com/Perl/perl5/commit/7b4f2c8c0daab6ada1dca83f1c6db3a8b9be794d
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
white after grok_num

Commit: 492fc7dfcd28d74f7a878ace83ef3c89479b332d
https://github.com/Perl/perl5/commit/492fc7dfcd28d74f7a878ace83ef3c89479b332d
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M toke.c

Log Message:
-----------
frac

Commit: c9a5117dbd7206555005f7a9907a0e9341409c3f
https://github.com/Perl/perl5/commit/c9a5117dbd7206555005f7a9907a0e9341409c3f
Author: Karl Williamson <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M embed.fnc
M handy.h
M inline.h
M numeric.c
M pod/perldiag.pod
M proto.h
M t/op/hexfp.t
M t/op/oct.t
M toke.c

Log Message:
-----------
trial

Compare: https://github.com/Perl/perl5/compare/72f140eed6b8%5E...c9a5117dbd72

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] d5c986: Correct misspelling in inline comment by James E Keenan via perl5-changes

Sun, 07 Dec 2025 12:03:41 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: d5c986535bafb40a8586761bd554986e1be88b1c
https://github.com/Perl/perl5/commit/d5c986535bafb40a8586761bd554986e1be88b1c
Author: James E Keenan <[email protected]>
Date: 2025-12-07 (Sun, 07 Dec 2025)

Changed paths:
M regen/warnings.pl

Log Message:
-----------
Correct misspelling in inline comment

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] by James E Keenan via perl5-changes

Sat, 06 Dec 2025 23:25:07 +0000

Branch: refs/heads/goto-label-fatal
Home: https://github.com/Perl/perl5

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] by James E Keenan via perl5-changes

Sat, 06 Dec 2025 23:17:52 +0000

Branch: refs/heads/goto-label-jump-fatalization-20250823
Home: https://github.com/Perl/perl5

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 5863d2: Revert "handy.h: Add void * casts to memEQ,memNE" by Karl Williamson via perl5-changes

Thu, 04 Dec 2025 20:22:49 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 5863d273902a6f74becb7a8fac8e3c8874bbcef9
https://github.com/Perl/perl5/commit/5863d273902a6f74becb7a8fac8e3c8874bbcef9
Author: Karl Williamson <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M handy.h

Log Message:
-----------
Revert "handy.h: Add void * casts to memEQ, memNE"

This reverts commit 9d3980bc229750e6c07726fe529f02bf4dc6a5a5.

It can create problems to add casts to macros; potentially hiding real
issues. These casts caused Tony Cook some lost time recently; whatever
the cause for this commit isn't showing up on my box; so I'll try
reverting it to see if something shows up on some other platform.

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] by David Mitchell via perl5-changes

Thu, 04 Dec 2025 18:43:00 +0000

Branch: refs/heads/davem/gh23903
Home: https://github.com/Perl/perl5

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 343b97: make pregexec() handle zero-length stringsagain by David Mitchell via perl5-changes

Thu, 04 Dec 2025 18:41:16 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 343b977ef3c2b03da4546280d57e3ce3d40a40d4
https://github.com/Perl/perl5/commit/343b977ef3c2b03da4546280d57e3ce3d40a40d4
Author: David Mitchell <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M embed.fnc
M ext/XS-APItest/APItest.pm
M ext/XS-APItest/APItest.xs
M ext/XS-APItest/t/callregexec.t
M proto.h

Log Message:
-----------
make pregexec() handle zero-length strings again

GH #23903

In embed.fnc, commit v5.43.3-167-g45ea12db26 added SPTR, EPTR parameter
modifiers to (amongst other API functions), Perl_pregexec().

These cause assert constraints to be added to the effect that SPTR <
EPTR (since the latter is supposed to be a pointer to the byte after the
last character in the string).

This falls down for an empty string since in this case pregexec() is
called with strbeg == strend.

This was causing an assert failure in the test suite for
Package-Stash-XS.

The reason it wasn't noticed before is because:

1) pregexec() is a thin wrapper over regexec_flags();

2) The perl core (e.g. pp_match()) calls regexec_flags() rather than
pregexec();

3) Package::Stash::XS has XS code which calls pregexec() directly rather
than using CALLREGEXEC() (which would call regexec_flags());

4) In embed.fnc, regexec_flags()'s strend parameter is declared as
NN rather than EPTR, so it doesn't get the assert added.

So very little code was actually using pregexec().

This commit, for now, changes pregexec()'s strend parameter from EPTR to
EPTRQ, which has the net effect of allowing zero-length strings to be
passed, and thus fixes the CPAN issue.

But longer term, we need to decide: is the general logic for EPTR wrong?
Should the assert be SPTR <= EPTR? And should EPTR be applied to
regexec_flags()'s strend parameter too?

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 8e698e: perlapi: Combine grok_(bin|hex|oct) into singleentry by Karl Williamson via perl5-changes

Thu, 04 Dec 2025 17:08:45 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 8e698ee4907b52797e37e7cf9ab70591d9af90c1
https://github.com/Perl/perl5/commit/8e698ee4907b52797e37e7cf9ab70591d9af90c1
Author: Karl Williamson <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M numeric.c

Log Message:
-----------
perlapi: Combine grok_(bin|hex|oct) into single entry

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 209bfe: handy.h: Avoid UB in nBIT_MASK() by Karl Williamson via perl5-changes

Thu, 04 Dec 2025 17:08:13 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 209bfe369d94aa6e6d0ab65c27920b649c61cdd6
https://github.com/Perl/perl5/commit/209bfe369d94aa6e6d0ab65c27920b649c61cdd6
Author: Karl Williamson <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M handy.h

Log Message:
-----------
handy.h: Avoid UB in nBIT_MASK()

I discovered the hard way that this is undefined behavior when operating
on the widest unsigned integer type available on the platform.

I couldn't think of a way to write this without a branch that worked
both for that condition and a zero length mask

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 8ac8d0: Reformat XDIGIT_VALUE() by Karl Williamson via perl5-changes

Thu, 04 Dec 2025 17:07:43 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 8ac8d0564ad9f91017d8d2bf93db71daa718817b
https://github.com/Perl/perl5/commit/8ac8d0564ad9f91017d8d2bf93db71daa718817b
Author: Karl Williamson <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M handy.h

Log Message:
-----------
Reformat XDIGIT_VALUE()

This makes it clearer what is going on, and may encourage compiler
optimization

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 980fad: Rename THING token when used for attributes by Paul Evans via perl5-changes

Thu, 04 Dec 2025 10:23:20 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 980fade0ba45c1fcccc26a93f4aee086b80b357a
https://github.com/Perl/perl5/commit/980fade0ba45c1fcccc26a93f4aee086b80b357a
Author: Paul "LeoNerd" Evans <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M perly.act
M perly.h
M perly.tab
M perly.y
M toke.c

Log Message:
-----------
Rename THING token when used for attributes

Prior to this, toke.c and perly.y used the token name `THING` for a
number of apparently totally unrelated reasons. While I haven't
attempted to fix most of them, I did find that the one used for
attribute lists is unrelated to the others, so I have renamed that to
something more sensible and meaningful.

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] e5694b: embed.fnc/group_end: correct the patend pointerde... by Tony Cook via perl5-changes

Thu, 04 Dec 2025 04:48:39 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: e5694b9e05a3e5427ca966f0a7d0dc60b5519e62
https://github.com/Perl/perl5/commit/e5694b9e05a3e5427ca966f0a7d0dc60b5519e62
Author: Tony Cook <[email protected]>
Date: 2025-12-04 (Thu, 04 Dec 2025)

Changed paths:
M embed.fnc
M proto.h
M t/op/pack.t

Log Message:
-----------
embed.fnc/group_end: correct the patend pointer decorator

This was incorrectly asserting the supplied string had at least one
character, which could produce an assertion instead of a useful error
message for the user of pack().

Related to #23980

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] d79d16: pp_ctl.c: Comment on 2 exceptions in pp_goto() by James E Keenan via perl5-changes

Wed, 03 Dec 2025 18:23:58 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: d79d16960362c0639aa2ab6586aa35176975d9af
https://github.com/Perl/perl5/commit/d79d16960362c0639aa2ab6586aa35176975d9af
Author: James E Keenan <[email protected]>
Date: 2025-12-03 (Wed, 03 Dec 2025)

Changed paths:
M pp_ctl.c

Log Message:
-----------
pp_ctl.c: Comment on 2 exceptions in pp_goto()

Per discussion in GH #23969, neither of these two exceptions can be
reached from our test suite. The exceptions address the possibility of a
subroutine without a name, which is unlikely except possibly from XS
code.

The exceptions immediately preceding each of the two upon which we are
commenting have messages composed to include a subroutine's name. These
exceptions can be reached from our test suite and are covered as of the
time of this commit.

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 6ff96b: Remove period from end of abstract (subject) by James E Keenan via perl5-changes

Wed, 03 Dec 2025 18:23:30 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 6ff96b930bf8b81b1c8c8d3434f52814e307c8ff
https://github.com/Perl/perl5/commit/6ff96b930bf8b81b1c8c8d3434f52814e307c8ff
Author: James E Keenan <[email protected]>
Date: 2025-12-03 (Wed, 03 Dec 2025)

Changed paths:
M utils/pl2pm.PL

Log Message:
-----------
Remove period from end of abstract (subject)

This should satisfy concerns raised in GH #23939.

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 8993af: make pregexec() handle zero-length stringsagain by David Mitchell via perl5-changes

Wed, 03 Dec 2025 12:01:53 +0000

Branch: refs/heads/davem/gh23903
Home: https://github.com/Perl/perl5
Commit: 8993af9d1f44c04b8feaea56f46fbb33703e67a5
https://github.com/Perl/perl5/commit/8993af9d1f44c04b8feaea56f46fbb33703e67a5
Author: David Mitchell <[email protected]>
Date: 2025-12-03 (Wed, 03 Dec 2025)

Changed paths:
M embed.fnc
M ext/XS-APItest/APItest.pm
M ext/XS-APItest/APItest.xs
M ext/XS-APItest/t/callregexec.t
M proto.h

Log Message:
-----------
make pregexec() handle zero-length strings again

GH #23903

In embed.fnc, commit v5.43.3-167-g45ea12db26 added SPTR, EPTR parameter
modifiers to (amongst other API functions), Perl_pregexec().

These cause assert constraints to be added to the effect that SPTR <
EPTR (since the latter is supposed to be a pointer to the byte after the
last character in the string).

This falls down for an empty string since in this case pregexec() is
called with strbeg == strend.

This was causing an assert failure in the test suite for
Package-Stash-XS.

The reason it wasn't noticed before is because:

1) pregexec() is a thin wrapper over regexec_flags();

2) The perl core (e.g. pp_match()) calls regexec_flags() rather than
pregexec();

3) Package::Stash::XS has XS code which calls pregexec() directly rather
than using CALLREGEXEC() (which would call regexec_flags());

4) In embed.fnc, regexec_flags()'s strend parameter is declared as
NN rather than EPTR, so it doesn't get the assert added.

So very little code was actually using pregexec().

This commit, for now, changes pregexec()'s strend parameter from EPTR to
EPTRQ, which has the net effect of allowing zero-length strings to be
passed, and thus fixes the CPAN issue.

But longer term, we need to decide: is the general logic for EPTR wrong?
Should the assert be SPTR <= EPTR? And should EPTR be applied to
regexec_flags()'s strend parameter too?

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] e3b0f7: Perl_sv_backoff - only copy the buffer contentsif... by Richard Leach via perl5-changes

Wed, 03 Dec 2025 00:28:49 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: e3b0f74b8368e302993dcab22ba6d42613dd8ef5
https://github.com/Perl/perl5/commit/e3b0f74b8368e302993dcab22ba6d42613dd8ef5
Author: Richard Leach <[email protected]>
Date: 2025-12-03 (Wed, 03 Dec 2025)

Changed paths:
M sv.c

Log Message:
-----------
Perl_sv_backoff - only copy the buffer contents if SvOK(sv)

When undoing an OOK, `Perl_sv_backoff` will always copy the shifted buffer
contents down to the start of the buffer. That's required when the contents
are live, unfortunate when the contents are defunct but tiny, and can have a
noticeably bad effect on performance when the contents are defunct but large.

See https://github.com/Perl/perl5/issues/23967 for an example of the latter.

With this commit the copy will only occur if `SvOK(sv)`. This test is expected
to indicate whether the buffer contents are active or defunct.

Commit: 4342210875c1fea7efb91105b1d600cbea2cc012
https://github.com/Perl/perl5/commit/4342210875c1fea7efb91105b1d600cbea2cc012
Author: Richard Leach <[email protected]>
Date: 2025-12-03 (Wed, 03 Dec 2025)

Changed paths:
M scope.c

Log Message:
-----------
Perl_leave_scope - sv_backoff shouldn't do an unnecessay string copy

When a `my` SV goes out of scope, any OOK hack on its string buffer is
undone by `Perl_sv_backoff`. If the SV is `SvOK`, a copy of the buffer
contents will occur, but since the contents are defunct at this point,
the copy is unnecessary.

See https://github.com/Perl/perl5/issues/23967 as an example of where
this unnecessary copy had a noticeable effect on performance.

This commit essentially inlines the necessary parts of `sv_backoff` to
avoid the copy, without excessive messing around with `sv`'s flags at
the call site in `Perl_leave_scope`.

Commit: c8c2426c21c92620ec98c0e3309b732188960e5a
https://github.com/Perl/perl5/commit/c8c2426c21c92620ec98c0e3309b732188960e5a
Author: Richard Leach <[email protected]>
Date: 2025-12-03 (Wed, 03 Dec 2025)

Changed paths:
M pod/perldelta.pod

Log Message:
-----------
Perldelta for GH#23968: OOK buffer copy only when SvOK(sv)

Compare: https://github.com/Perl/perl5/compare/dd8309d23d71...c8c2426c21c9

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 4e9d64: pp_reverse - chunk-at-a-time string reversal by Richard Leach via perl5-changes

Tue, 02 Dec 2025 22:44:27 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 4e9d64bf17d377e8e2f1412ef936b804fb0ee215
https://github.com/Perl/perl5/commit/4e9d64bf17d377e8e2f1412ef936b804fb0ee215
Author: Richard Leach <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M pp.c

Log Message:
-----------
pp_reverse - chunk-at-a-time string reversal

The performance characteristics of string reversal in blead is very
variable depending upon the capabilities of the C compiler. Some
compilers are able to vectorize some cases for better performance.

This commit introduces explicit reversal and swapping of whole
registers at a time, which all builds seem to be able to
benefit from.

The `_swab_xx_` macros for doing this already exist in perl.h,
using them for this purpose was inspired by
https://dev.to/wunk/fast-array-reversal-with-simd-j3p
The bit shifting done by these macros should be portable and reasonably
performant if not optimised further, but it is likely that they will
be optimised to bswap, rev, movbe instructions.

Some performance comparisons:

1. Large string reversal, with different source & destination buffers
my $x = "X"x(1024*1000*10); my $y; for (0..1_000) { $y = reverse $x }

gcc blead:
2,388.30 msec task-clock # 0.993 CPUs utilized
10,574,195,388 cycles # 4.427 GHz
61,520,672,268 instructions # 5.82 insn per cycle
10,255,049,869 branches # 4.294 G/sec

clang blead:
688.37 msec task-clock # 0.946 CPUs utilized
3,161,754,439 cycles # 4.593 GHz
8,986,420,860 instructions # 2.84 insn per cycle
324,734,391 branches # 471.745 M/sec

gcc patched:
408.39 msec task-clock # 0.936 CPUs utilized
1,617,273,653 cycles # 3.960 GHz
6,422,991,675 instructions # 3.97 insn per cycle
644,856,283 branches # 1.579 G/sec

clang patched:
397.61 msec task-clock # 0.924 CPUs utilized
1,655,838,316 cycles # 4.165 GHz
5,782,487,237 instructions # 3.49 insn per cycle
324,586,437 branches # 816.350 M/sec

2. Large string reversal, but reversing the buffer in-place
my $x = "X"x(1024*1000*10); my $y; for (0..1_000) { $y = reverse "foo",$x }

gcc blead:
6,038.06 msec task-clock # 0.996 CPUs utilized
27,109,273,840 cycles # 4.490 GHz
41,987,097,139 instructions # 1.55 insn per cycle
5,211,350,347 branches # 863.083 M/sec

clang blead:
5,815.86 msec task-clock # 0.995 CPUs utilized
26,962,768,616 cycles # 4.636 GHz
47,111,208,664 instructions # 1.75 insn per cycle
5,211,117,921 branches # 896.018 M/sec

gcc patched:
1,003.49 msec task-clock # 0.999 CPUs utilized
4,298,242,624 cycles # 4.283 GHz
7,387,822,303 instructions # 1.72 insn per cycle
725,892,855 branches # 723.367 M/sec

clang patched:
970.78 msec task-clock # 0.973 CPUs utilized
4,436,489,695 cycles # 4.570 GHz
8,028,374,567 instructions # 1.81 insn per cycle
725,867,979 branches # 747.713 M/sec

3. Short string reversal, different source & destination (checking performance on
smaller string reversals - note: this one's vary variable due to noise)
my $x = "1234567"; my $y; for (0..10_000_000) { $y = reverse $x }

gcc blead:
401.20 msec task-clock # 0.916 CPUs utilized
1,672,263,966 cycles # 4.168 GHz
5,564,078,603 instructions # 3.33 insn per cycle
1,250,983,219 branches # 3.118 G/sec

clang blead:
380.58 msec task-clock # 0.998 CPUs utilized
1,615,634,265 cycles # 4.245 GHz
5,583,854,366 instructions # 3.46 insn per cycle
1,300,935,443 branches # 3.418 G/sec

gcc patched:
381.62 msec task-clock # 0.999 CPUs utilized
1,566,807,988 cycles # 4.106 GHz
5,474,069,670 instructions # 3.49 insn per cycle
1,240,983,221 branches # 3.252 G/sec

clang patched:
346.21 msec task-clock # 0.999 CPUs utilized
1,600,780,787 cycles # 4.624 GHz
5,493,773,623 instructions # 3.43 insn per cycle
1,270,915,076 branches # 3.671 G/sec

Commit: dd8309d23d7193f614a9c35def5da5bc2fed9300
https://github.com/Perl/perl5/commit/dd8309d23d7193f614a9c35def5da5bc2fed9300
Author: Richard Leach <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M pp.c

Log Message:
-----------
Add MSVC support for chunk-at-a-time string reversal

The heavy lifting for this commit was done by bulk88 in GH#23374.

Any deficiencies in transcription are down to richardleach. ;)

Compare: https://github.com/Perl/perl5/compare/407267be97c8...dd8309d23d71

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] by David Mitchell via perl5-changes

Tue, 02 Dec 2025 21:02:23 +0000

Branch: refs/heads/davem/re_debug
Home: https://github.com/Perl/perl5

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 407267: regex engine: split EVAL_postponed_AB state by David Mitchell via perl5-changes

Tue, 02 Dec 2025 21:00:50 +0000

Branch: refs/heads/davem/re_debug
Home: https://github.com/Perl/perl5
Commit: 407267be97c83fd72b334a0953ab6142a27de01f
https://github.com/Perl/perl5/commit/407267be97c83fd72b334a0953ab6142a27de01f
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M regcomp.sym
M regexec.c
M regnodes.h

Log Message:
-----------
regex engine: split EVAL_postponed_AB state

(This commit makes no practical changes in behaviour except for
debugging output.)

Currently, one of the regex engine stack states is EVAL_postponed_AB.

When executing something like /(??{'A'})B/ where A and B represent
general subpatterns, the engine executes the eval code, which returns
the string 'A', which is compiled into a subpattern. Then the engine
pushes an EVAL_postponed_AB state and runs the subpattern until it
reaches an END op. Then it pushes *another* EVAL_postponed_AB state and
runs the B part of the pattern until the final END. Then before
returning success, it pops EVAL_postponed_AB off the stack (twice),
executing any cleanup required. Similarly during failure, the
EVAL_postponed_AB_fail action will be executed once or twice (depending
on whether it failed during A or B).

This commit splits that state into two,

EVAL_postponed_A
EVAL_postponed_B

The first is pushed before running A, the second before running B.
The actions currently remain the same and share the same code; i.e. this
commit just does the equivalent of:

- case EVAL_postponed_AB:
+ case EVAL_postponed_A:
+ case EVAL_postponed_B:
... cleanup code ....

But it makes the code easier to understand, makes debugging output
clearer, and will allow in future for the cleanup behaviours to differ
between A and B.

This commit also fixes up a few debugging messages and code comments
which were still referring to 'EVAL_AB', which was renamed to
EVAL_postponed_AB some years ago.

Commit: 5e0ff3da62d49aca7f4ad15653d602630d049c87
https://github.com/Perl/perl5/commit/5e0ff3da62d49aca7f4ad15653d602630d049c87
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:

Log Message:
-----------
Merge branch 'blead' into davem/re_debug

Compare: https://github.com/Perl/perl5/compare/6dee5e6b86cb...5e0ff3da62d4

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 676faf: regmatch(): improve savestack ix debuggingoutput by David Mitchell via perl5-changes

Tue, 02 Dec 2025 20:58:51 +0000

Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: 676faf69f7f048f041c0569cb843f62e6e605762
https://github.com/Perl/perl5/commit/676faf69f7f048f041c0569cb843f62e6e605762
Author: David Mitchell <[email protected]>
Date: 2025-12-01 (Mon, 01 Dec 2025)

Changed paths:
M regexec.c

Log Message:
-----------
regmatch(): improve savestack ix debugging output

Under

use re Debug => 'STATE';

S_regmatch() produces various bits of debugging output regarding
savestack manipulation; e.g. when the current index is saved; or when
the savestack is popped back to that index on success or failure etc.

Unfortunately that output is patchy and inconsistent. For example it
often talks confusingly about "Setting an EVAL scope" even when an EVAL
hasn't been executed. And it doesn't log all savestack manipulation.

After this commit there is a standard set of debugging messages, such
as:

savestack: snapping ix=14
savestack: regcppush filled in range [14..18)
savestack: regcppop freed in range [14..18)
savestack: freeing in range [10..14)

which are consistently applied.

There should be no change in functionality unless that specific
debugging is enabled.

Commit: 407267be97c83fd72b334a0953ab6142a27de01f
https://github.com/Perl/perl5/commit/407267be97c83fd72b334a0953ab6142a27de01f
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M regcomp.sym
M regexec.c
M regnodes.h

Log Message:
-----------
regex engine: split EVAL_postponed_AB state

(This commit makes no practical changes in behaviour except for
debugging output.)

Currently, one of the regex engine stack states is EVAL_postponed_AB.

When executing something like /(??{'A'})B/ where A and B represent
general subpatterns, the engine executes the eval code, which returns
the string 'A', which is compiled into a subpattern. Then the engine
pushes an EVAL_postponed_AB state and runs the subpattern until it
reaches an END op. Then it pushes *another* EVAL_postponed_AB state and
runs the B part of the pattern until the final END. Then before
returning success, it pops EVAL_postponed_AB off the stack (twice),
executing any cleanup required. Similarly during failure, the
EVAL_postponed_AB_fail action will be executed once or twice (depending
on whether it failed during A or B).

This commit splits that state into two,

EVAL_postponed_A
EVAL_postponed_B

The first is pushed before running A, the second before running B.
The actions currently remain the same and share the same code; i.e. this
commit just does the equivalent of:

- case EVAL_postponed_AB:
+ case EVAL_postponed_A:
+ case EVAL_postponed_B:
... cleanup code ....

But it makes the code easier to understand, makes debugging output
clearer, and will allow in future for the cleanup behaviours to differ
between A and B.

This commit also fixes up a few debugging messages and code comments
which were still referring to 'EVAL_AB', which was renamed to
EVAL_postponed_AB some years ago.

Compare: https://github.com/Perl/perl5/compare/0e943cea3fce...407267be97c8

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 694c5f: Bump ExtUtils::ParseXS to 3.61 by David Mitchell via perl5-changes

Tue, 02 Dec 2025 13:13:00 +0000

Branch: refs/heads/davem/xs_ast
Home: https://github.com/Perl/perl5
Commit: 694c5fe954b335a2977fbe94b7bb2d695db2ed68
https://github.com/Perl/perl5/commit/694c5fe954b335a2977fbe94b7bb2d695db2ed68
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Constants.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/CountLines.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Eval.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Utilities.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/Typemaps.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/Typemaps/Cmd.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/Typemaps/InputMap.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/Typemaps/OutputMap.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/Typemaps/Type.pm
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
Bump ExtUtils::ParseXS to 3.61

Commit: d268acde34767d535cb67120e98ba670332fdd28
https://github.com/Perl/perl5/commit/d268acde34767d535cb67120e98ba670332fdd28
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add stub top-level node

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add ExtUtils::ParseXS::Node::XS_file class.
It doesn't do a lot yet.

Commit: ae631d9a69f7cb20a7032be23c0d1bd296eee619
https://github.com/Perl/perl5/commit/ae631d9a69f7cb20a7032be23c0d1bd296eee619
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::preamble

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add a stub Node subclass which is responsible for emitting the preamble
at the start of the generated C file.

Commit: addb8b2b48dad2bf772e0213b04249cd62863d2d
https://github.com/Perl/perl5/commit/addb8b2b48dad2bf772e0213b04249cd62863d2d
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::C_part

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add these new AST node types:

ExtUtils::ParseXS::Node::C_part;
ExtUtils::ParseXS::Node::C_part_POD;
ExtUtils::ParseXS::Node::C_part_code;

and add this method:

ExtUtils::ParseXS::Node::is_xs_module_line()

These are collectively used to parse and hold the "C" part of the XS
file - i.e. everything that comes before the first MODULE line.

A C_part node has children consisting of C_part_POD and C_part_code
nodes.

Commit: ec74c45de0ca26f0f9df64595ba9e384287727f2
https://github.com/Perl/perl5/commit/ec74c45de0ca26f0f9df64595ba9e384287727f2
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::postamble

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add a stub Node subclass which is responsible for emitting the postamble
to the C file following any C code which has been copied as-is from the
C part of the XS file

Commit: 25044e131a1d15d1da8883447c74d2a8cefd0472
https://github.com/Perl/perl5/commit/25044e131a1d15d1da8883447c74d2a8cefd0472
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::global_cpp_line

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

This node stores a single "#foo BAR" C preprocessor line which appears
within the XS part of an XS file, and which is in global (file) scope
(as opposed to being within code in an XSUB or BOOT).

Commit: 47dc181c2f0d8de8eb3bf837ffe3a8e27d995127
https://github.com/Perl/perl5/commit/47dc181c2f0d8de8eb3bf837ffe3a8e27d995127
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: simplify double-negative regex

This condition:

match a line which doesn't start with anything other than '#'

can be more simply expressed as:

match a line which starts with '#'

They're equivalent. Or at least I think they are. This was doing my head
in.

Commit: ee9516a65953c93e7d595b7ab4e933ab7f45a2ca
https://github.com/Perl/perl5/commit/ee9516a65953c93e7d595b7ab4e933ab7f45a2ca
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::BOOT

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Remove the old BOOT_handler() and replace it with a new Node type.

Commit: 5fde9ee944d23cf436741be148f1f636a7f1bfea
https://github.com/Perl/perl5/commit/5fde9ee944d23cf436741be148f1f636a7f1bfea
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: regularise BOOT processing

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

For some reason most of the file-scoped keywords (PROTOTYPES etc) are
handled together in a single 'while /REQUIRE|PROTOTYPES|.../' loop, but
then BOOT is checked for separately after that loop has completed.

This commit just moves checking for and handling BOOT into that same
loop. The only significant difference between BOOT and the other
keywords is that BOOT consumes all line until the end of the paragraph,
rather than stopping at the next keyword. But this shouldn't make any
difference to the top-level parsing loop.

Commit: 4c5d768c20c71fc51abd8635f0cd034c990522ec
https://github.com/Perl/perl5/commit/4c5d768c20c71fc51abd8635f0cd034c990522ec
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::FALLBACK

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Remove the old FALLBACK_handler() and replace it with a new Node type.
Also, change $pxs->{map_package_to_fallback_string} so that it
now holds logical values ('TRUE' etc) rather than the actual strings
to later emit into the code ('&PL_sv_yes' etc). Then do the translation
at the point the code is emitted.

Commit: 54e27538b1f745b03f59bf1d5976be9de4f40665
https://github.com/Perl/perl5/commit/54e27538b1f745b03f59bf1d5976be9de4f40665
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::REQUIRE

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Remove the old REQUIRE_handler() and replace it with a new Node type.

Commit: 2790c5773be9de9fc7d1a28c1accdae0d7abf0f9
https://github.com/Perl/perl5/commit/2790c5773be9de9fc7d1a28c1accdae0d7abf0f9
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Eval.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: refactor: store current pkg in Node::xsub

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

In preparation for code generation to be done only after the whole file
has been parsed (rather than after each XSUB is parsed, as at present),
store the current package name in each Node::xsub object.

Then use that value, rather than $pxs->{PACKAGE_name}, to specify the
package name in various bits of boot code and in the $Package variable
in typemaps. This is because by the end of parsing, $pxs->{PACKAGE_name}
will have the value which was extracted from the *last* seen MODULE line
in the file, and not the value at the point where an XSUB was parsed.

Add several tests to confirm things work, and will continue to work in a
few commits' time, when we switch to code generation coming after *all*
parsing.

Commit: 321ee8f8a7ebe9ca35d69f72d317c0f30fde8416
https://github.com/Perl/perl5/commit/321ee8f8a7ebe9ca35d69f72d317c0f30fde8416
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::include etc

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

This commit adds these new Node types:

ExtUtils::ParseXS::Node::include
ExtUtils::ParseXS::Node::INCLUDE
ExtUtils::ParseXS::Node::INCLUDE_COMMAND

and removes these two (now obsolete) methods:

INCLUDE_handler
INCLUDE_COMMAND_handler

INCLUDE_COMMAND_handler() looks to have been a cut+paste of
INCLUDE_handler() with the relevant bits changed to handle a piped
command rather than a filename. I've combined their functionality back
into a single parse() method within the Node::include base class, then
made Node::INCLUDE and Node::INCLUDE_COMMAND trivial subclasses which
just set (or not) $self->{is_cmd} and then let the base class do all the
heavy lifting.

Commit: 9c64c8e15e0215709f092e357a34ee67923c4f4e
https://github.com/Perl/perl5/commit/9c64c8e15e0215709f092e357a34ee67923c4f4e
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: remove residual FOO_handler

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Remove a branch which says "if present, call new-style
Node::FOO::parse(), else call old-style call FOO_handler()", since there
are now AST nodes for all keywords.

Also update some of the code comments to remove mention of FOO_handler
and that way of working.

Commit: 620fa16460d9b359ed97455ee057895624e2d3d1
https://github.com/Perl/perl5/commit/620fa16460d9b359ed97455ee057895624e2d3d1
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: remove check_keyword() method

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

This method used to be used for scanning for keywords and then invoking
old-style FOO_handler()s. Replace its last usage with the new
parse_keywords() method (which looks for keywords and creates
Node::FOO's instead), and remove the method itself.

For now, create a temporary node to accumulate all the file-scoped
keywords in a continuous run, and then call as_code() once at the end.

I.e for the input:

PROTOTYPES: X
REQUIRE: Y
..

formerly a temporary Node::PROTOTYPES node was created, parsed, and then
as_code() immediately called on it, then freed. Then similarly for
REQUIRE etc.

After this commit, a temporary parent node is created, child nodes are
added to it for each of PROTOTYPES, REQUIRE, etc, and finally as_code()
is called once which calls as_code() for each child.

This is a small step in the process of moving towards deferring calling
as_code() until the whole file has been read in and parsed.

The next commit will make the temporary parent node into something
better.

Commit: 0081492bcb4b756aa0ed8b18dbb7c49fa85b322b
https://github.com/Perl/perl5/commit/0081492bcb4b756aa0ed8b18dbb7c49fa85b322b
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: add Node::as_concise method

For debugging: displays the node and its children in a compact style
similar to perl -MO=Concise

Commit: a16d2b475a63bd7f5a14c327f3f537c962b9aab5
https://github.com/Perl/perl5/commit/a16d2b475a63bd7f5a14c327f3f537c962b9aab5
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::cpp_scope

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add a new node type,

ExtUtils::ParseXS::Node::cpp_scope.

This node takes on responsibility for the handling the main file-scoped
parsing loop, which looks for file-scoped keywords, CPP directives,
the start of XSUB declarations etc. It is the main child of the top-most
Node::XS_file node type.

The bulk of this commit consists of cutting and pasting the main parsing
loop (including fetch_para()) out from ExtUtils::ParseXS::process_file()
and into ExtUtils::ParseXS::Node::cpp_scope::parse().

For now this as a bit hacky - all tests pass, but mainly by ensuring
(temporarily) that various bits of code are emitted during the parsing
phase, rather than via a call to as_code(), in order to preserve the
correct ordering. The next few commits will gradually clean things up.

This commit also converts the lexical variable $cpp_next_tmp_define
into a field of the ExtUtils::ParseXS object, as it now needs accessing
outside of process_file().

The intent for this node type is that it represents anywhere within a
single C preprocessor #if / #else / #endif branch / scope. For this
commit, there's only a single cpp_scope node, representing the whole XS
part of the main file. Over the next few commits, further nodes will be
added: during INCLUDE, which is a whole separate scope, and between CPP
conditional directives. When all is complete, it will allow for simpler
handling of handling "duplicate" XSUB declarations in different #if
branches, and the XS_parse_stack mechanism can be eliminated.

Commit: 9016336891184076fd8d591a6a0ecd57f85ccfb3
https://github.com/Perl/perl5/commit/9016336891184076fd8d591a6a0ecd57f85ccfb3
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: use Node::cpp_scope for INCLUDE

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

This commit makes the INCLUDE keyword's parse() method use a new
Node::cpp_scope node as a child to hold the nodes generated by parsing
the included file. So whereas before those nodes were flattened into the
parent list of nodes, they now form a distinct subtree.

This involves recursively calling Node::cpp_scope::parse() and
fetch_para(), rather than the included file being processed in the main
loop.

For example, given a main file which contains two declared XSUBs, main1
and main2, and an included file with XSUBs inc1 and inc2; then
previously, the parse tree would have had this nesting:

XS_file
cpp_scope[type=main]
xsub main1
include
xsub inc1
xsub inc2
xsub main2

whereas it now looks like:

XS_file
cpp_scope[type=main]
xsub main1
include
cpp_scope[type=include]
xsub inc1
xsub inc2
xsub main2

This commit introduces a slight bug in terms of the order in which BOOT
blocks in main and included files get emitted; this will be fixed in a
few commits' time.

Commit: 9e686aa48a8ea3b5978f5fb797be96a14a6d8147
https://github.com/Perl/perl5/commit/9e686aa48a8ea3b5978f5fb797be96a14a6d8147
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: rm push_parse_stack, PopFile

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Delete the ExtUtils::ParseXS methods push_parse_stack() and PopFile(),
and inline their single use each in
ExtUtils::ParseXS::Node::include::parse().

This is commit is just a crude cut+paste and s/$self/$pxs/g.
Subsequent commits will clean the code up.

Commit: b641429f637978db5aec52c0646d32719ae8627e
https://github.com/Perl/perl5/commit/b641429f637978db5aec52c0646d32719ae8627e
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: INCLUDE: save state locally

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Previously whenever INCLUDE or INCLUDE_COMMAND parsed a new file, the
old state (filename, old lines etc) were saved in a new entry pushed
onto the XS_parse_stack.

The previous commit inlined this push and pull into
Node::include::parse(), which now looked a bit like:

sub parse() {
...
push @{$pxs->{XS_parse_stack}}, {
type =>' file',
Filename => $pxs->{in_filename},
...
}

... open new file and parse ...

my $data = pop @{$pxs->{XS_parse_stack}};
$pxs->{in_pathname} = $data->{Filepathname};
...
}

The intention over the next few commits is to eliminate XS_parse_stack
completely: the functionality of the stack will be replaced by a nested
series of Node::cpp_scope nodes within the AST. For now, this commit
still pushes/pops an entry (so that #if / #endif scope analysing
continues to work), but saves all the values in a local var. So after
this commit, Node::include::parse() looks like:

sub parse() {
...
push @{$pxs->{XS_parse_stack}}, { type =>' file' };
my @saved = ($pxs->{in_filename}, ...);

... open new file and parse ...

pop @{$pxs->{XS_parse_stack}};
($pxs->{in_filename}, ...) = @saved;
}

The rest of this commit is just general tidying up and reordering of
some of the code within parse() without change in functionality. Except
I also added

$pxs->{line_no} = [];

in addition to the already present

$pxs->{line} = [];

Commit: 19c1a90a59891f072054f89a0cdc82247348ea2a
https://github.com/Perl/perl5/commit/19c1a90a59891f072054f89a0cdc82247348ea2a
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: global_cpp_line: parse CPP line

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Make the recently-added Node::global_cpp_line::parse() method
split a C preprocessor line into its components and set some flags
indicating whether its part of a conditional and if so, an if or variant.

Save these values as new fields in the node. They will come in useful
shortly.

Commit: a0b5f3109c531173bd7eaa12cc6822d420d43d16
https://github.com/Perl/perl5/commit/a0b5f3109c531173bd7eaa12cc6822d420d43d16
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: fix #elif processing

The XS parser is supposed to be able to do basic processing of
file-scoped C preprocessor directives: in particular, to handle
alternative XSUB declarations; e.g.:

#if X
void foo(short i)
#elif Y
void foo(long i)
#else
void foo(int i)
#endif

But due to a bug present since this feature was added in perl5.003
(and which no one has complained about for 29 years), the XS parser
was incorrectly looking for the string '#elsif' rather than '#elif'.

So the above would fail, but something using only 'else' would compile
okay:

#if X
void foo(short i)
#else
void foo(int i)
#endif

This commit fixes the bug and adds some tests.

Commit: ae5354281892081f70da55c64b277d3f5847528e
https://github.com/Perl/perl5/commit/ae5354281892081f70da55c64b277d3f5847528e
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: add more #if/#endif tests

Add more tests for how the XS parser processes conditional C
preprocessor directives in file scope.

Note that some of these tests will start to fail after the next commit,
which re-implements the processing. The fixups to these tests will
indicate what has changed.

Commit: 89452f0a04208bd9d90c346b2ee8f97c3a50ac9d
https://github.com/Perl/perl5/commit/89452f0a04208bd9d90c346b2ee8f97c3a50ac9d
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: refactor: reimplement #if processing

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

This commit includes a few significant changes to ParseXS which are not
separable into individual commits. Some of these also involve minor
changes in behaviour. Hence a rather long commit message, using
pseudo-POD headers to break things up a bit and give a bit of structure.

=head1 Background

The XS parser does some basic analysis of file-scoped C preprocessor
conditional directives such as #if, #elif, #endif. This is mainly for
the purpose of handling otherwise duplicate XSUBs, e.g.

#ifdef USE_2ARG

int foo(int i, int j)

#else

int foo(int i)

#endif

For the above, the parser suppresses any 'duplicate' warning, and adds a
'guard' define at the start of each branch, such as the following in the
emitted C code:

#ifdef USE_2ARG
#define XSubPPtmpAAAA 1

XS_EUPXS(XS_Foo__Bar_foo) { ... }

#else
#define XSubPPtmpAAAB 1

XS_EUPXS(XS_Foo__Bar_foo) { ... }

#endif

It then uses those guards in the boot code to ensure that only the
correct XSUB(s) (and aliases etc) are registered:

#if XSubPPtmpAAAA
newXS_deffile("Foo::Bar::foo", XS_Foo__Bar_foo);
#endif

#if XSubPPtmpAAAB
newXS_deffile("Foo::Bar::foo", XS_Foo__Bar_foo);
#endif

=head1 The changes in this commit

=head2 Add a Node::cpp_scope node between branches

The XS code above was, before this commit, compiled into an AST
which looked something like:

XS_file
preamble
C_part
postamble
cpp_scope[type=main]
global_cpp_line [if]
xsub
global_cpp_line [else]
xsub
global_cpp_line [endif]

After this commit, all nodes between each conditional CPP line are
now hung off a new cpp_scope node:

XS_file
preamble
C_part
postamble
cpp_scope[type=main]
global_cpp_line [if]
cpp_scope[type=if]
xsub
global_cpp_line [else]
cpp_scope[type=if]
xsub
global_cpp_line [endif]

=head2 The logic to do the #if/#endif processing is completely
reimplemented from scratch.

Formerly, @{pxs->{XS_parse_stack}} was a stack that was pushed or popped
for every INCLUDE and #if/#endif, and the state about conditionals was
contained in that stack.

After this commit, the state is instead implied by the structure of the
AST; in particular, in the currently active cpp_scope node.
The cpp_scope::parse() method has the fetch_para() loop, and each time
it encounters an #if etc it, it creates a child cpp_scope node and
recursively calls its parse() method. That method returns when it sees
the end of a scope, such as encountering an #elif, #else or #endif or
EOF.

Now, virtually all of the logic involved in maintaining '#if' state is
in cpp_scope::parse(), and all the logic to emit '#define XSubPPtmpAAAB
1' etc code is in cpp_scope::as_code(). This is in contrast to
previously, where the logic and code generation were liberally spread
around in ParseXS.pm and Node.pm

To facilitate this change, two fields have been added to the cpp_scope
node type:

guard_name - the 'XSubPPtmpAAAB' or whatever to use for this branch;
seen_xsubs - hash of XSUB names seen in this branch.

After this commit, the data in @{pxs->{XS_parse_stack}} isn't used;
the next commit will remove it completely.

=head2 Almost all code is now generated after parsing is finished.

At the start of this git branch, code was emitted after each XSUB was
parsed. Now (with a few residual exceptions to be addressed shortly) the
whole file is first parsed into a single AST, and *then* the top-level
as_code() method is called to generate the entire C file from the AST.

Some of the changes in this commit represent the removal of temporary
shims that preserved the old parse/emit ordering, and which are now no
longer required.

=head1 Visible changes in behaviour

Although this commit is mostly intended to be just a refactoring, some
changes in behaviour (in terms of what C code is generated and what
warnings are issued) has occurred.

These are made visible in the changes this commit makes to
t/001-basic.t, which deliberately contains tests for the behaviours
which were about to change.

The changes are:

=head2 fewer warnings

The "Warning: duplicate function definition" warning is now emitted
less often, to avoid false positives. In particular, this fixes GH
19661. That ticket concerned code like

#if C1
void foo()
#endif

#if C2
void foo()
#endif

The logic was such that only one of C1 and C2 could ever be true
(basically DBD:MariaDB was adding a missing DBI function if the
installed DBI was an old one). But the XS parser can't know that.
So it used to warn.

The big change in this commit is to now warn only for two identically
named XSUBs which appear in the *same* cpp_scope. This removes false
positives. It may add false negatives, but that is fairly harmless,
as it just means that bad C code will be generated, and the C compiler
will eventually complain anyway.

The guard definitions are still generated as before.

This change in behaviour was intentional, but was also the easiest: it
meant just checking the seen_xsubs hash for the current cpp_scope node.
For more complex behaviour (if you are considering implementing it), it
may require having the concept of a 'current' cpp_scope, e.g.
$pxs->{cur_cpp_scope}, plus a chain of pointers from the current scope
back to outer ones. (This is slightly problematic, as weak refs aren't
available, as Scalar::Util may not be built yet - so it might require
higher level node(s) to have a destructor which searches for those links
and deletes them first) Once the chain is in place, then it's possible
for a newly-added XSUB to update the parent seen_xsubs hashes too, etc.

=head2 The sequencing of XSubPPtmpAAA, XSubPPtmpAAB etc may change

Previously, the sequence of the guard defines was strictly in the order
they appear in the C file. Now, ones in a nested scope come before outer
ones, e.g.

#ifdef C1
#define XSubPPtmpAAB 1
# ifdef C2
# define XSubPPtmpAAA 1
...

This should be harmless.

=head2 XSubPPtmpAAA etc may be placed slightly differently

Previously in something like

#ifdef C1
#define X 1
#define Y 2

the guard would be added after any other CPP lines; the above would
cause this C code to be emitted:

#ifdef C1
#define X 1
#define Y 2
#define XSubPPtmpAAA 1

but is now:

#ifdef C1
#define XSubPPtmpAAA 1
#define X 1
#define Y 2

i.e.the guard now always comes *immediately* after the conditional at
the start of the branch it will guard.

This should be harmless.

Commit: 1ab2dcc845131f648cbacc80d9f745b7ea55a7ef
https://github.com/Perl/perl5/commit/1ab2dcc845131f648cbacc80d9f745b7ea55a7ef
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M MANIFEST
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Utilities.pm
R dist/ExtUtils-ParseXS/t/111-analyze_preprocessor_statements.t
M dist/ExtUtils-ParseXS/t/113-check_cond_preproc_statements.t

Log Message:
-----------
ParseXS: refactor: remove parser stack

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

The previous commits have changed how the parser tracks file-scoped
#if / #endif sections - this is now simply part of the structure
of the AST.

So this commit removes the old parser stack, @{$pxs->{XS_parse_stack}},
which used to be used to track this info. The stack was still present,
but was no longer used to deduce #if state. It also removes the
analyze_preprocessor_statements() method and the test file which used to
test it, t/111-analyze_preprocessor_statements.t.

One minor regression in this commit is that a warning about about
unbalanced #if/#endif within an XSUB or BOOT block used to sometimes
emit an extra hint about what the problem might be:

print STDERR " (precede it with a blank line if the matching #if is outside the function)\n"
if $self->{XS_parse_stack}->[-1]{type} eq 'if';

Since it's no longer trivial to check whether we're currently within a
global #if/endif section, I've kept the 'unbalanced' warning, but
deleted the additional hint.

If anybody wants to add it back, you'll need to add a field to the
EU::PXS class such as

$pxs->{current_cpp_scope}

which keeps track of which is the current (innermost) Node::cpp_scope
node being parsed. Then check_conditional_preprocessor_statements() can
use it in place of the "$self->{XS_parse_stack}->[-1]{type} eq 'if';"
test.

Commit: 165e25d1d53a9fd1426987e9fbd5249c9df1db9a
https://github.com/Perl/perl5/commit/165e25d1d53a9fd1426987e9fbd5249c9df1db9a
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: move main parsing into XS_file

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Move the parsing of the XS part of the main file from process_file()
into cpp_scope::parse().

No functional change.

Commit: 457af2777d90f9ea3260a1a432af73f02feba05e
https://github.com/Perl/perl5/commit/457af2777d90f9ea3260a1a432af73f02feba05e
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::pre_boot

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add a stub Node subclass which is responsible for emitting any C code
which comes after the XS file has been parsed (and any user XSUBs
emitted), but before the boot XSUB is emitted.

Currently this code is just concerned with adding overload methods.

The code has just been moved from process_file() to the node's
as_code method() and is otherwise unchanged.

Commit: 504148c22659a3320c7fc5e272d3478729e5d760
https://github.com/Perl/perl5/commit/504148c22659a3320c7fc5e272d3478729e5d760
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::boot_xsub

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add a stub Node subclass which is responsible for emitting the C code
for the boot XSUB.

The code has just been moved from process_file() to the node's
as_code method() and is otherwise unchanged.

Commit: 864b568d2620b5d6a3f4039933b7c36c8557d2d7
https://github.com/Perl/perl5/commit/864b568d2620b5d6a3f4039933b7c36c8557d2d7
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: tidy Node::boot_xsub::parse()

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

The previous commit moved code from process_file() into it's own
method, with minimal changes. Now clean the code up a bit:

- use $open_brace, $close_brace instead of the magic [[ ]];
- clean up comments, white space etc;
- use Q to make some code strings more readable.

Commit: e28a97b96fe1a6bfbf36a46f7fc54ea9a1779fef
https://github.com/Perl/perl5/commit/e28a97b96fe1a6bfbf36a46f7fc54ea9a1779fef
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: move some code

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Move some code from process_file() to XS_file::parse()

No change in functionality.

Commit: 06775815876913f1e32a41fe479f01f32385cba0
https://github.com/Perl/perl5/commit/06775815876913f1e32a41fe479f01f32385cba0
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: emit package overload registrations later

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

In the boot XSUB, a bunch of newXS() calls are added, to register all
the XSUBs in the perl namespace, plus all aliases, overload methods etc.

In addition, for every package which contains at least one overload
method, a special stub sub called Foo::Bar::() is registered. This is
used at run-time to determine whether overloading is present for a
particular class.

This commit changes the order, by changing an unshift to a push.
Previously the stub package subs were registered first, then the XSUBs;
now its the other way round. e.g.

/* mark these two packages as having overload methods */
newXS_deffile("X::Y::()", XS_X__Y_nil);
newXS_deffile("P::Q::()", XS_P__Q_nil);

/* foo XSUB used as an overload method for <=> etc */
newXS_deffile("P::Q::foo", XS_P__Q_foo);
newXS_deffile("P::Q::(<=>", XS_P__Q_foo);
...
/* bar XSUB used as an overload method for <=> etc */
newXS_deffile("X::Y::bar", XS_X__Y_bar);
newXS_deffile("X::Y::(<=>", XS_X__Y_bar);
...

becomes

/* foo XSUB used as an overload method for <=> etc */
newXS_deffile("P::Q::foo", XS_P__Q_foo);
newXS_deffile("P::Q::(<=>", XS_P__Q_foo);
...
/* bar XSUB used as an overload method for <=> etc */
newXS_deffile("X::Y::bar", XS_X__Y_bar);
newXS_deffile("X::Y::(<=>", XS_X__Y_bar);
...

/* mark these two packages as having overload methods */
newXS_deffile("P::Q::()", XS_P__Q_nil);
newXS_deffile("X::Y::()", XS_X__Y_nil);

This should make no difference to the functionality, but will make
a further refactoring of boot code generation easier.

AFAIKT, using unshift rather than push in the original 2002 commit
didn't have any particular intent.

Commit: e8fdab34e5780c03f4e82c53f1fa8a6f5883f11a
https://github.com/Perl/perl5/commit/e8fdab34e5780c03f4e82c53f1fa8a6f5883f11a
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add as_boot_code() methods

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Before this commit, as_code() methods were responsible for generating
both the general code that goes in the C file, and also code that
specifically goes in the boot SUB. They achieved this by pushing
lines into the arrays $pxs->{bootcode_early} and $pxs->{bootcode_later}.

This commit adds as_boot_code() methods to the relevant node types and
removes the bootcode_early and bootcode_later arrays.

Now, the C code file is generated by walking the AST, calling as_code()
methods. When it gets to the boot_xsub::as_code() method, which is
responsible for emitting the XS(boot_Foo__Bar) {} boot XSUB, that sub
itself then does a second treewalk, calling as_boot_code() methods, which
accumulates lines to be added early and later on in the boot SUB.

Commit: 0e56af3707404162e8391e55b6c1f4d629353e05
https://github.com/Perl/perl5/commit/0e56af3707404162e8391e55b6c1f4d629353e05
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M MANIFEST
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Utilities.pm
R dist/ExtUtils-ParseXS/t/109-standard_XS_defs.t

Log Message:
-----------
ParseXS: refactor: inline and rm standard_XS_defs

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

The ExtUtils::ParseXS::Utilities::standard_XS_defs() function
just returns a big string containing all the standard boilerplate code
that gets added to the C file. Delete this function, and instead
include the text directly within C_part_postamble::as_code().

Commit: 0bed57b7c7a26a0be0bdf5979967a39894826df3
https://github.com/Perl/perl5/commit/0bed57b7c7a26a0be0bdf5979967a39894826df3
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: don't use $_ for C part of file

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

When parsing the C part of the XS file (mainly looking for POD and C
code), don't use $_ and $. to hold the state of the next line. This was
ok when it was a simple loop, but now that the logic is spread over
multiple parse() methods, store the next line to process in
$pxs->{lastline} and $pxs->{lastline_no}.

(The XS half of the file already uses those variables, along with
@{$pxs->{line}} etc.)

Commit: 9e4ed57fe55c2c3090df9c1b92a90055b59a5053
https://github.com/Perl/perl5/commit/9e4ed57fe55c2c3090df9c1b92a90055b59a5053
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: remove $ExtUtils::ParseXS::END

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

The parser, when parsing an XSUB (but not any other parts of an XS file),
pushes a special token onto the end of the list of lines in
@{$pxs->{line}} which make up the current paragraph. This token,

$ExtUtils::ParseXS::END = "!End!\n\n";

(along with a trailing ':') is designed to look like an impossible
keyword which can't actually appear in the source code (due to the
multiple newlines).

It looks like it was originally added in perl5.002 to make the parsing
code easier, but I don't really understand why. It just makes the
parser harder to understand.

So this commit removes it, and just relies on @{$pxs->{line}} being zero
to detect the end of the paragraph. This change doesn't alter the C code
generated from any of the XS files bundled with perl.

Commit: 0e526ab02ed31f1a2769789598b261e697f01972
https://github.com/Perl/perl5/commit/0e526ab02ed31f1a2769789598b261e697f01972
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: remove stray $_

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Remove this line:

$_ = '';

which is a leftover from when $_ used to maintain the current line
between calls to various parsing subs. It's redundant now.

Commit: 01a94f36db995af0aa4d305a0bbeae6610a3e900
https://github.com/Perl/perl5/commit/01a94f36db995af0aa4d305a0bbeae6610a3e900
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Utilities.pm
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: spot MODULE line syntax errors

Previously, a line which was not a completely syntactically correct
MODULE line was not treated as a module line; instead it got passed on
uninterpreted until likely causing an error further along in parsing.

This commit changes things so that anything that looks like the *start*
of a module line is treated as a module line, and is *only then*
examined for full syntactic correctness, giving an error if not ok.

For example: previously, this line:

MODULE = Foo XXXPACKAGE = Foo::Bar

gave the weird error message:

Error: Function definition too short 'MODULE = Foo XXXPACKAGE ...

but now gives the error message:

Error: unparseable MODULE line: 'MODULE = Foo XXXPACKAGE ...

In particular, any line starting with

/^MODULE\s*[=:]/

is now treated as an attempt to declare a module, including the
syntactically incorrect 'MODULE:' form.

This is in the same spirit that other keywords are already processed;
for example

PROTOTYPES: XXXENABLE

is treated as as a badly-formed PROTOTYPES line rather than an otherwise
unrecognised and unprocessed line.

Commit: 969eb6c21c29267a137c4808a7fdd99a52ffd6fd
https://github.com/Perl/perl5/commit/969eb6c21c29267a137c4808a7fdd99a52ffd6fd
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::MODULE

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add a new node type,

ExtUtils::ParseXS::Node::MODULE.

this (mostly) regularises the treatment of a MODULE line, now handled in
the usual way by parse_keywords(), rather than being a special magical
snowflake that got its own handling in fetch_para().

This commit also changes one parameter of parse_keywords() from being a
boolean to being a bit flag, now that there are now *two* slightly
special cases to flag up: MODULE, in addition to NOT_IMPLEMENTED_YET.

This commit is supposed to have no changes in behaviour, but there
*might* be some edge cases that I haven't thought of.

Commit: 264963f5e34f971e1af8231e8b5a721e177e5055
https://github.com/Perl/perl5/commit/264963f5e34f971e1af8231e8b5a721e177e5055
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: spot TYPEMAP: line syntax errors

Previously, a line which was not a completely syntactically correct
TYPEMAP: line was not treated as a TYPEMAP line; instead it got passed on
uninterpreted until likely causing an error further along in parsing.

This commit changes things so that anything starting with
/^TYPEMAP\s*:/ is treated as a TYPEMAP line, and is *only then*
examined for full syntactic correctness, giving an error if not ok.

This is similar to two commits ago which did the same for the MODULE
keyword.

Commit: d86a388f680484feed13be22bc3ff7649f13bf31
https://github.com/Perl/perl5/commit/d86a388f680484feed13be22bc3ff7649f13bf31
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::TYPEMAP

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Add a new node type,

ExtUtils::ParseXS::Node::TYPEMAP.

this (mostly) regularises the treatment of a TYPEMAP line, now handled in
the usual way by parse_keywords(), rather than being processed solely
within fetch_para(). fetch_para() does still need to do *some*
processing: it has to read the TYPEMAP line, extract out the <<EOF bit,
and use that to read lines up until the matchhing 'EOF'. It then returns
these lines (including the initial TYPEMAP line) as a paragraph which is
then handled as a normal file-scoped keyword.

With this commit, fetch_para() becomes closer to its official job of
just returning the next useful chunk of lines, rather than having hidden
side effects too.

This commit is supposed to have no changes in behaviour, but there
*might* be some edge cases that I haven't thought of.

Commit: 8ace24a19195a83d1c07025aaaa8a71c9e755f8d
https://github.com/Perl/perl5/commit/8ace24a19195a83d1c07025aaaa8a71c9e755f8d
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: inline _maybe_skip_pod()

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Inline the one use of _maybe_skip_pod() into fetch_para().

Commit: a83466f52d576694188e5ad1708a7de93b76a911
https://github.com/Perl/perl5/commit/a83466f52d576694188e5ad1708a7de93b76a911
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: move a chomp()

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

Change where $pxs->{lastline} gets chomped at the start of the XS half
of the file, where the fetch_para() loop is first entered - do it just
the once when transitioning rather than polluting the main fetch_para()
loop with it.

No functional changes.

Commit: fc392dbc1128de22b7b90c84473b12a79ffdaae1
https://github.com/Perl/perl5/commit/fc392dbc1128de22b7b90c84473b12a79ffdaae1
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: chomp TYPEMAP lines

(This commit is part of a series which will extend the AST parse tree
from just representing individual XSUBs to representing the whole XS
file.)

For consistency with everything else in fetch_para(), chomp any lines
extracted from a TYPEMAP section before stuffing them in
@{$pxs->{line}}. Later, make Node::TYPEMAP::parse() stick a newline
back on each line before passing them to the typemap parser.

Commit: 7c7e63d517e6646b3ae5d6bddc18a79e2237a7a7
https://github.com/Perl/perl5/commit/7c7e63d517e6646b3ae5d6bddc18a79e2237a7a7
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: fix POD on last line

If POD continued to the last line of the XS file, a bunch of
"Use of uninitialized value" warnings would be generated.

The fix is trivial, but it also required a slight tweak to the test
framework in t/001-basic.t, which was unconditionally adding a \n to the
end of each test string, thus making the '=cut' appear not to be at EOF.

Commit: 21e6471be3c9ecf6050995b09ecfe2b70994c2f6
https://github.com/Perl/perl5/commit/21e6471be3c9ecf6050995b09ecfe2b70994c2f6
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: fetch_para(): remove POD loop

Remove the loop which says 'foreach contiguous section of pod, delete
it'. Replace it with with something which removes just *one* section of
POD, then relies on the main loop to process the next line, which may
well be the start of another POD section.

Fewer nested loops is less cognitive load.

Should be no functional changes.

Commit: caa8c16f744eac5bc8987dc66976fa1c3b8282fb
https://github.com/Perl/perl5/commit/caa8c16f744eac5bc8987dc66976fa1c3b8282fb
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: fetch_para(): invert if

change

if (cond) {
... lots lines ...
}

to

if (!cond) {
goto read_next_line;
}
... lots lines ...

read_next_line:

which might be considered a bit regressive in a 'goto considered
harmful' sort of way, but will will help with further refactoring,
and emphasises that the current input line is discarded if the condition
isn't met (i.e. it strips out code comments, which wasn't obvious).

Commit: 35b69a9252e55177c62490999f9bcc112e6414c4
https://github.com/Perl/perl5/commit/35b69a9252e55177c62490999f9bcc112e6414c4
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: fetch_para(): reindent

Reindent a block of code after the previous commit removed a scope.
Whitespace-only.

Commit: 27eb48219099f2a56b966686d93f70cecc5a2d92
https://github.com/Perl/perl5/commit/27eb48219099f2a56b966686d93f70cecc5a2d92
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: process line continuations after POD

Within fetch_para(), the at-end-of-POD code and the main loop both had
separate (but similar) chunks of code to read in the next line, with the
latter one being more complete. In particular, the latter had logic to
process '\' continuation lines, while the POD one didn't. This meant
that this gave a parse error:

=pod
=cut
void foo(int i, \
int j)

The fix is to make them both use the same code to read the next line.

Commit: fbfd5fbdcb3403019e4d2cb1dcaab66a1e490e11
https://github.com/Perl/perl5/commit/fbfd5fbdcb3403019e4d2cb1dcaab66a1e490e11
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: process line continuations after TYPEMAP

(This is similar to the previous commit, which fixed the same bug for
POD)

Within fetch_para(), the at-end-of-TYPEMAP code and the main loop both
had separate (but similar) chunks of code to read in the next line, with
the latter one being more complete. In particular, the latter had logic
to process '\' continuation lines, while the TYPEMAP one didn't. This
meant that this gave a parse error:

TYPEMAP: <<EOF
...
EOF
void foo(int i, \
int j)

The fix is to make them both use the same code to read the next line.
This bug was probably actually introduced by me during the refactoring
in this branch.

By using the same read-next-line code, it also means that an immediately
following blank line is now stripped of whitespace. I'm not sure whether
that actually affects anything in the real word. Also, any trailing
lines in typemap are now stripped. Again, this shouldn't affect
anything.

Commit: 0028ecc6f630e6e90ffe6e00493378953c45d568
https://github.com/Perl/perl5/commit/0028ecc6f630e6e90ffe6e00493378953c45d568
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: fetch_para(): add more comments

Add more code comments to fetch_para(), especially explaining how
the "XSUB cuddled by #if/#endif without blank lines" mechanism works

Commit: f38a029e6612ebcb66ffecab371f74aa3f9ca7a3
https://github.com/Perl/perl5/commit/f38a029e6612ebcb66ffecab371f74aa3f9ca7a3
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: fetch_para(): simplify logic

Simplify the logic of the code which handles C-preprocessor directives
where there's not the (usually required) blank line directly before and
after an XSUB; e.g.

#if X
int foo(...)
...
#endif

This commit splits the logic into two halves: the first is responsible
for handling conditionals directly *before* the XSUB, and the other half
for ones during and directly after the XSUB. Then the second half is
further simplified.

I think this makes the code easier to understand.

I *think* the new code is logically equivalent.

Commit: a21bd3ae4e21cf2e1ec2e9f492963c9b563c9ce7
https://github.com/Perl/perl5/commit/a21bd3ae4e21cf2e1ec2e9f492963c9b563c9ce7
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm

Log Message:
-----------
ParseXS: refactor: fetch_para: update description

Update the code comments above this function.

This is both to reflect the changes made in this branch (e.g.
no longer processing MODULE lines in this function), and to reflect
my better understanding of this function since I wrote the original
text.

Commit: c7ee219e7c9be497a8532043699e65fd94e54cf1
https://github.com/Perl/perl5/commit/c7ee219e7c9be497a8532043699e65fd94e54cf1
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: fix comments re die if indented

Rewrite some code comments explaining why the XS parser sometimes dies
if the next line is indented. It's a bit complicated and the initial
comments I added a year ago weren't entirely accurate.

Commit: 5a14bdfd780cdf7850b68db06c9f8ade8dfe5f60
https://github.com/Perl/perl5/commit/5a14bdfd780cdf7850b68db06c9f8ade8dfe5f60
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: simplify main loop

Simplify the code in Node::cpp_scope::parse(), which contains the
main file-scoped parsing loop, and update / add code comments.

The code in this loop has gone through some rough times with all the
refactoring that has taken place in the last year, chiefly converting
the parser into being AST-based. This commit is kind of a final tidy up
to restore sanity. Apart from fixing lots of comments, the main thing
this commit does is flatten nested while loops. Before this commit, the
overall structure of the parse() function was like;

PARAGRAPH:
while (lines to process) {
while (blank line) { skip }
while (C-preprocessor line) { process }
... and similar loops ...
}

and is now:

while (lines to process) {
if (blank line) { skip; next }
if (C-preprocessor line) { process; next }

... and similar loops ...
}

i.e. rather than there being nested loops, just have one main loop
and go back to the start of that to do the next thing. This should
be functionally equivalent, but less complicated.

This new paradigm can't be applied yet to the last few items in the main
loop due to a backwards compatibility issue with the SCOPE keyword.
I'll fix that at some future date.

Commit: 98301ea422f5644f81d6bac24b85601d2183dbaa
https://github.com/Perl/perl5/commit/98301ea422f5644f81d6bac24b85601d2183dbaa
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: rationalise parse-fail action

Now that the XS parser builds a parse tree (AST), it spends most of its
time doing recursive-descent $some_node_object->parse() calls.

There is a common idiom within the current code along the lines of

$some_node_object->parse()
or return;

This makes a parse failure in a node propagate its way back up the call
chain. Note that many parse failures don't immediately croak; they
call blurt() which only warns, but also increments the error count;
parsing can continue.

During my big refactoring I haven't been entirely consistent here; this
commit adds a bunch of 'or return's that I missed. It also does a few
'or next's at higher levels, to get behaviour along the lines of:
"this XSUB had a syntax error, but continue trying to parse other XSUBs,
and only error out at the end."

This is all a bit going on gut feelings; in the final analysis it
doesn't matter too much how much more parsing gets done before xsubpp
finally croaks out.

Commit: e098f86a6623e49b5a27c4c5a1016fa136b860a2
https://github.com/Perl/perl5/commit/e098f86a6623e49b5a27c4c5a1016fa136b860a2
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS.pm
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: update POD and comments

Update the POD at the top of Node.pm and the code comments at the top
of ParseXS.pm to reflect the changes in this branch which have extended
the AST from representing just an XSUB to representing the whole XS
file.

Commit: 7c119be465d5c250a2545d6f663bddfc9b8a4ed6
https://github.com/Perl/perl5/commit/7c119be465d5c250a2545d6f663bddfc9b8a4ed6
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/ExtUtils/ParseXS/Node.pm

Log Message:
-----------
ParseXS: refactor: add Node::Q method

Add a method which is a simple wrapper around ExtUtils::ParseXS::Q.
This means throughout Node.pm, you can write

$self->Q(<<"EOF")

rather than the more long-winded

ExtUtils::ParseXS::Q(<<"EOF");

Commit: 7888dacc7017f682ca811d4fd7fd377cc4ed5d93
https://github.com/Perl/perl5/commit/7888dacc7017f682ca811d4fd7fd377cc4ed5d93
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
ParseXS: refactor: re-add tests for C preamble

This earlier commit in the branch:

ParseXS: refactor: inline and rm standard_XS_defs

removed a test file which checked that the standard C preamble text
was present. This commit re-adds those tests in a different way.

Compare: https://github.com/Perl/perl5/compare/d88ca93a8b61...7888dacc7017

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] d04442: perlxs.pod: remove section on %v by David Mitchell via perl5-changes

Tue, 02 Dec 2025 12:06:52 +0000

Branch: refs/heads/davem/xspod
Home: https://github.com/Perl/perl5
Commit: d044425cd0e30d2ae29d54a485d545df1b7716eb
https://github.com/Perl/perl5/commit/d044425cd0e30d2ae29d54a485d545df1b7716eb
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: remove section on %v

The XS parser supported an extremely obscure bit of functionality which
made use of the %v package variable to maintain state between different
bits of typemap processing. This was accidentally broken in 5.10.0:
refactoring removed the 'use vars "%v"' line, and no one seemed to
notice or care.

Also, the sole example of its use in the docs seemed to be obscure,
confusing and probably wrong.

There was a consensus in the discussion at

http://nntp.perl.org/group/perl.perl5.porters/267667

that we should stop documenting this feature rather than trying to fix
it.

Commit: c2d26f6785e44f1a4925f0de56ce947613844a8e
https://github.com/Perl/perl5/commit/c2d26f6785e44f1a4925f0de56ce947613844a8e
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: reindent and reformat code examples

The various XS code examples had odd and inconsistent indentation (often
with 5 leading spaces) and inconsistent formatting, e.g. foo(a,b) vs
foo( a, b ) vs foo(a, b). Fix that, and also remove any tab chars.

Whitespace-only change.

Commit: bafe93f7df38cbc46918a9507ddf53c55d8361d1
https://github.com/Perl/perl5/commit/bafe93f7df38cbc46918a9507ddf53c55d8361d1
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: delete most non-ref sections

This commit is a simple cut which deletes several '=head2' sections from
perlxs.pod. The next commit will tidy up and fix any broken links etc.

These sections are more tutorial-like, and aren't in line with the goal
of this branch that perlxs.pod becomes purely a reference manual for XS.
Any relevant information from these sections may be incorporated later
into new sections in perlxs.pod and/or be included in a future rewrite
of perlxstut.pod.

The sections deleted are:

=head2 Introduction
=head2 On The Road
=head2 The Anatomy of an XSUB
=head2 The Argument Stack
=head2 The RETVAL Variable
=head2 Returning SVs, AVs and HVs through RETVAL
=head2 Returning Undef And Empty Lists
=head2 Interface Strategy
=head2 Perl Objects And C Structures

Commit: 942ab5a5afb123ac8ffb7afaad86719882f2e7d3
https://github.com/Perl/perl5/commit/942ab5a5afb123ac8ffb7afaad86719882f2e7d3
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M XSUB.h
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: fix up links after deleting sections

The previous commit deleted several sections from perlxs.pod. This
commit fixes things up; done as a separate commit so that the changes
aren't drowned out in the diff listing.

Commit: e7ddc3c9b6802f96d3791773e22f2e7c7823e5b6
https://github.com/Perl/perl5/commit/e7ddc3c9b6802f96d3791773e22f2e7c7823e5b6
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: reorder sections

This big commit does a series of plain cut+pastes to reorder all the
=head2 sections within the file.

This changes the order from semi-random into roughly the order the
various XS keywords would appear within an XS file, and then within an
XSUB declaration/definition.

No changes have been made to the text: simply that all lines from a
particular '^=head2' up until the next head2 have been cut+paste as
a single unit.

No attempt has been made yet to make the text consistent with the new
ordering; that will be done by the subsequent commits of this branch.

The previous ordering in this file was:

=head1 NAME
=head1 DESCRIPTION
=head2 The MODULE Keyword
=head2 The PACKAGE Keyword
=head2 The PREFIX Keyword
=head2 The OUTPUT: Keyword
=head2 The NO_OUTPUT Keyword
=head2 The CODE: Keyword
=head2 The INIT: Keyword
=head2 The NO_INIT Keyword
=head2 The TYPEMAP: Keyword
=head2 Initializing Function Parameters
=head2 Default Parameter Values
=head2 The PREINIT: Keyword
=head2 The SCOPE: Keyword
=head2 The INPUT: Keyword
=head2 The IN/OUTLIST/IN_OUTLIST/OUT/IN_OUT Keywords
=head2 The C<length(NAME)> Keyword
=head2 Variable-length Parameter Lists
=head2 The C_ARGS: Keyword
=head2 The PPCODE: Keyword
=head2 The REQUIRE: Keyword
=head2 The CLEANUP: Keyword
=head2 The POSTCALL: Keyword
=head2 The BOOT: Keyword
=head2 The VERSIONCHECK: Keyword
=head2 The PROTOTYPES: Keyword
=head2 The PROTOTYPE: Keyword
=head2 The ALIAS: Keyword
=head2 The OVERLOAD: Keyword
=head2 The FALLBACK: Keyword
=head2 The INTERFACE: Keyword
=head2 The INTERFACE_MACRO: Keyword
=head2 The INCLUDE: Keyword
=head2 The INCLUDE_COMMAND: Keyword
=head2 The CASE: Keyword
=head2 The EXPORT_XSUB_SYMBOLS: Keyword
=head2 The & Unary Operator
=head2 Inserting POD, Comments and C Preprocessor Directives
=head2 Using XS With C++
=head2 Safely Storing Static Data in XS
=head3 MY_CXT REFERENCE
=head1 EXAMPLES
=head1 CAVEATS
=head2 Use of standard C library functions
=head2 Event loops and control flow
=head1 XS VERSION
=head1 AUTHOR DIAGNOSTICS
=head1 AUTHOR

and is now:

=head1 NAME
=head1 DESCRIPTION
=head2 The MODULE Keyword
=head2 The PACKAGE Keyword
=head2 The PREFIX Keyword
=head2 Inserting POD, Comments and C Preprocessor Directives
=head2 The REQUIRE: Keyword
=head2 The VERSIONCHECK: Keyword
=head2 The PROTOTYPES: Keyword
=head2 The EXPORT_XSUB_SYMBOLS: Keyword
=head2 The INCLUDE: Keyword
=head2 The INCLUDE_COMMAND: Keyword
=head2 The TYPEMAP: Keyword
=head2 The BOOT: Keyword
=head2 The FALLBACK: Keyword
=head2 The NO_OUTPUT Keyword
=head2 The IN/OUTLIST/IN_OUTLIST/OUT/IN_OUT Keywords
=head2 Default Parameter Values
=head2 The C<length(NAME)> Keyword
=head2 Variable-length Parameter Lists
=head2 The PREINIT: Keyword
=head2 The INPUT: Keyword
=head2 The NO_INIT Keyword
=head2 Initializing Function Parameters
=head2 The & Unary Operator
=head2 The SCOPE: Keyword
=head2 The INIT: Keyword
=head2 The C_ARGS: Keyword
=head2 The CODE: Keyword
=head2 The PPCODE: Keyword
=head2 The POSTCALL: Keyword
=head2 The OUTPUT: Keyword
=head2 The CLEANUP: Keyword
=head2 The PROTOTYPE: Keyword
=head2 The OVERLOAD: Keyword
=head2 The ALIAS: Keyword
=head2 The INTERFACE: Keyword
=head2 The INTERFACE_MACRO: Keyword
=head2 The CASE: Keyword
=head2 Using XS With C++
=head2 Safely Storing Static Data in XS
=head3 MY_CXT REFERENCE
=head1 EXAMPLES
=head1 CAVEATS
=head2 Use of standard C library functions
=head2 Event loops and control flow
=head1 XS VERSION
=head1 AUTHOR DIAGNOSTICS
=head1 AUTHOR

Commit: 1d00b081479c5377bae150e4570f36a9aeca7898
https://github.com/Perl/perl5/commit/1d00b081479c5377bae150e4570f36a9aeca7898
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add group headers

Following the previous commit's reordering of the all the =head2
sections, demote most of the =head2 headers to =head3, and add some new
=head2 headers which group together related headers.

Also add some =head3's for a few missing keywords.

Subsequent commits will flesh out the new sections.

Commit: a43c5fe3a961e5783be5a82c7d15939a8f1fdb60
https://github.com/Perl/perl5/commit/a43c5fe3a961e5783be5a82c7d15939a8f1fdb60
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add a new introductory part

Four commits ago, I removed most of the general text sections in
perlxs (i.e. the ones not specifically about a particular keyword).

Now this commit adds a completely new introductory part to perlxs, about
1200 lines long. It represents an attempt to write a background to what
XS and XSUBs, SVs, typemaps etc are, in a complete and modern way.
The existing reference section for each keyword follows it.

I tried to avoid getting too tutorial-like (that's what perlxstut is
for), but I may have crossed the line in various places. In particular
it has a new section which could have been titled "all the bits of
perlguts you need to know in order to write non-trivial XSUBs without
having to actually read perlguts".

Commit: c8d62e378edc4989826c2d1bb5c8c96cbf842fe9
https://github.com/Perl/perl5/commit/c8d62e378edc4989826c2d1bb5c8c96cbf842fe9
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add BNF definition section

Add a section which semi-formally tries to define the syntax and
structue of an XS file, using a BNF-like format.

See http://nntp.perl.org/group/perl.perl5.porters/268701 for the
discussion of this part.

Commit: 56fcc2c4a457daf37608ba364885e5c87d058ae4
https://github.com/Perl/perl5/commit/56fcc2c4a457daf37608ba364885e5c87d058ae4
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update MODULE/PACKAGE/PREFIX

Rewrite the POD for these three keywords, and in particular, treat
them as one declaration, rather than three unrelated keywords.

Commit: 24cf9e414dec614367a1e562def4483f6e152f19
https://github.com/Perl/perl5/commit/24cf9e414dec614367a1e562def4483f6e152f19
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update file-scoped directive text

Populate the new

=head2 File-scoped XS Keywords and Directives

section, partially by cannibalising (and then deleting) the old

=head3 Inserting POD, Comments and C Preprocessor Directives

subsection. This commit only adds text about directives; subsequent
commits will update the various file-scoped keywords.

Commit: b58fedb6792d8cbe2aa2a8f42c627a61a725ac12
https://github.com/Perl/perl5/commit/b58fedb6792d8cbe2aa2a8f42c627a61a725ac12
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update REQUIRE, VERSIONCHECK keywords

Commit: bb0c675c6078654398d1508e80f10d3df72a0330
https://github.com/Perl/perl5/commit/bb0c675c6078654398d1508e80f10d3df72a0330
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update PROTOTYPES: keyword

Commit: 24511a807c8f3783b1ead99a0a79a8fbe0a92324
https://github.com/Perl/perl5/commit/24511a807c8f3783b1ead99a0a79a8fbe0a92324
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update EXPORT_XSUB_SYMBOLS, INCLUDE(_COMMAND)

Commit: 9a7fd52d7dc4f7392249bf25433bc6b039c67ef1
https://github.com/Perl/perl5/commit/9a7fd52d7dc4f7392249bf25433bc6b039c67ef1
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update TYPEMAP: keyword

Commit: ed7511118cab32427e83885c8d3c13f9284ad49e
https://github.com/Perl/perl5/commit/ed7511118cab32427e83885c8d3c13f9284ad49e
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update BOOT: keyword

Commit: 31a0f7600502f8cbf1a2c743bf28c9679c4356bc
https://github.com/Perl/perl5/commit/31a0f7600502f8cbf1a2c743bf28c9679c4356bc
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update FALLBACK: keyword

Commit: d206967dc8d4d925d4ea4ba248905b035de70999
https://github.com/Perl/perl5/commit/d206967dc8d4d925d4ea4ba248905b035de70999
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update XSUB Structure + Declaration

Populate the new

=head2 The Structure of an XSUB
=head2 An XSUB Declaration

sections

Commit: 67c737d941d8119ef05b6b10c324a58c306bb565
https://github.com/Perl/perl5/commit/67c737d941d8119ef05b6b10c324a58c306bb565
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update section 'An XSUB Parameter'

Add some initial text for this new section, and also add a new
subsection "XSUB Parameter Placeholders".

Commit: 0d99ab087a7394e1c41e94cd185bb769d323b7b2
https://github.com/Perl/perl5/commit/0d99ab087a7394e1c41e94cd185bb769d323b7b2
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update IN_OUT etc section

Commit: 3d9d4154df3daab932864ffea5e71cd283685cd4
https://github.com/Perl/perl5/commit/3d9d4154df3daab932864ffea5e71cd283685cd4
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M XSUB.h
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update default, length, ellipis params

Rewrite (and retitle) these three subsections:

=head3 Default Parameter Values
=head3 The C<length(NAME)> Keyword
=head3 Variable-length Parameter Lists

Commit: d3d53839d730c57a7af193acf1b77c41a9ca86d0
https://github.com/Perl/perl5/commit/d3d53839d730c57a7af193acf1b77c41a9ca86d0
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: Input part, PREINIT sections

Add text for the new '=head2 The XSUB Input Part' section, and rewrite
the existing entry for the PREINIT keyword.

Commit: ffc1d0c18a58589634b4e9419ddcc583fdd8bcc9
https://github.com/Perl/perl5/commit/ffc1d0c18a58589634b4e9419ddcc583fdd8bcc9
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update 'The INPUT: Keyword' section

This commit completely rewrites this section and subsections:

=head3 The INPUT: Keyword
=head4 The NO_INIT Keyword
=head4 Initializing Function Parameters
=head4 The & Unary Operator

It de-emphasises the INPUT keyword and suggests using ANSI XS signatures
etc instead.

Commit: 08f74af3ad78ecbe183faf6d6347c0f249322c0d
https://github.com/Perl/perl5/commit/08f74af3ad78ecbe183faf6d6347c0f249322c0d
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update 'SCOPE: Keyword' section

Commit: c83fa60d2d997af27fbb493f4751c062e128593a
https://github.com/Perl/perl5/commit/c83fa60d2d997af27fbb493f4751c062e128593a
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: init part, INIT sections

Add text for the new '=head2 The XSUB Init Part' section, and rewrite
the existing entry for the INIT keyword.

Commit: 6a9feedd1e708f175aafd1966bc08c4e8721f975
https://github.com/Perl/perl5/commit/6a9feedd1e708f175aafd1966bc08c4e8721f975
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: code part, autocall, C_ARGS

Add text to the new

=head2 The XSUB Code Part
=head3 Auto-calling a C function

sections, and rewrite the existing

=head4 The C_ARGS: Keyword

section

Commit: bd34cf6dd5b3ba3445380d28b69729828c8d536b
https://github.com/Perl/perl5/commit/bd34cf6dd5b3ba3445380d28b69729828c8d536b
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: CODE, PPCODE

Rewrite these sections:

=head3 The CODE: Keyword
=head3 The PPCODE: Keyword

Commit: 55c519eed4d8530cc2cea0a4c88141914e923070
https://github.com/Perl/perl5/commit/55c519eed4d8530cc2cea0a4c88141914e923070
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update NOT_IMPLEMENTED_YET: keyword

This keyword formerly wasn't documented. The docs now say "this is what
it is, but don't use it".

Commit: efa2b9691a69dcae95dae8dc2cd924d1ddad9e9d
https://github.com/Perl/perl5/commit/efa2b9691a69dcae95dae8dc2cd924d1ddad9e9d
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: output part

Add text to the new

=head2 The XSUB Output Part

section, and rewrite the text in these existing sections:

=head3 The POSTCALL: Keyword
=head3 The OUTPUT: Keyword

Commit: ccc212fea05064bd36b3f62abc68eb765d513e07
https://github.com/Perl/perl5/commit/ccc212fea05064bd36b3f62abc68eb765d513e07
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: cleanup part

Add text to the new

=head2 The XSUB Cleanup Part

section, and rewrite the text in this existing section:

=head3 The CLEANUP: Keyword

Commit: 7105474d75807c06bf5cde0ee2dd3ca89978bc1f
https://github.com/Perl/perl5/commit/7105474d75807c06bf5cde0ee2dd3ca89978bc1f
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update generic intro, PROTOTYPE

Add text to the new

=head2 XSUB Generic Keywords

section, and rewrite the text in this existing section:

=head3 The PROTOTYPE: Keyword

Commit: 4ac6fde262bdedbcb17994d43338ab33ae5b90b3
https://github.com/Perl/perl5/commit/4ac6fde262bdedbcb17994d43338ab33ae5b90b3
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: mention package name types

Explain that a 'C' parameter type in an XSUB declaration can actually
be a Perl package name or similar, e.g.

Foo::Bar
f(Foo::Bar obj, char *s)

Commit: 54c5a85a6a73a59585106f3492eaf40c57a8c010
https://github.com/Perl/perl5/commit/54c5a85a6a73a59585106f3492eaf40c57a8c010
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update OVERLOAD, add T_PTROBJ

First, add a new subsection

=head3 T_PTROBJ and opaque handles

to the TYPEMAPs section explaining how this typemap can be used to
map between Perl objects and C library handles. It provides a
fully-worked example of wrapping a simple arithmetic library.

Then completely rewrite the

=head3 The OVERLOAD: Keyword

section. In particular, it now refers to the new T_PTROBJ example and
shows how it can be extended to use overloading.

Commit: 73bc0cb2e21ac9b4369b6a297768a1f4b78aa319
https://github.com/Perl/perl5/commit/73bc0cb2e21ac9b4369b6a297768a1f4b78aa319
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: document ATTRS

This keyword was undocumented, even though it had been added 25 years
ago.

Commit: 1014eeebf463e85a1a09362ee6816a75190385f1
https://github.com/Perl/perl5/commit/1014eeebf463e85a1a09362ee6816a75190385f1
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add "Sharing XSUB bodies" section

Populate the introduction to this new section.

Commit: 7f37717be76e4739f3ebe2ecdc1c4ebb25801adb
https://github.com/Perl/perl5/commit/7f37717be76e4739f3ebe2ecdc1c4ebb25801adb
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: ALIAS

Rewrite this section:

=head3 The ALIAS: Keyword

Commit: b7be9ba32169f342d031a877d3de4ed936dd7ae3
https://github.com/Perl/perl5/commit/b7be9ba32169f342d031a877d3de4ed936dd7ae3
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update INTERFACE, INTERFACE_MACRO

Rewrite these sections:

=head3 The INTERFACE: Keyword
=head3 The INTERFACE_MACRO: Keyword

also demote the second to be a head4 child of the first. Then expand
the T_PTROBJ example to use INTERFACE as an alternative to ALIAS.

Commit: aabcc04c244e36b3b4841fa903fde89a64720388
https://github.com/Perl/perl5/commit/aabcc04c244e36b3b4841fa903fde89a64720388
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update: CASE

Rewrite this section:

=head3 The CASE: Keyword

Commit: f0c901c4b4783aac3facb1af3147936cf3ad157f
https://github.com/Perl/perl5/commit/f0c901c4b4783aac3facb1af3147936cf3ad157f
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: add "Using Typemaps" section

Populate this new section (except for the T_PTROBJ subsection, which had
already been added by an earlier commit within this branch).

Note that the "Common typemaps" subsection could probably benefit
from some further expansion by someone familiar with which built-in
T_FOO entries are useful.

Commit: 4e5f4d27b56552cfc52fc2ae1d7cf4283a683646
https://github.com/Perl/perl5/commit/4e5f4d27b56552cfc52fc2ae1d7cf4283a683646
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod
M t/porting/known_pod_issues.dat

Log Message:
-----------
perlxs.pod: update "Using XS With C++" section

Rewrite this section:

=head2 Using XS With C++

Disclaimer: I've never written a proper C++ program. I had to
(literally) dust off my 34-year old copy of Stroustrup(*) and also do
some Googling. Hopefully what I've written is sane.

(*) This was bought back in the days when people used to to learn things
by buying books, and when I thought that I ought to know something about
this newfangled C++ thing. I never got round to reading all of it: I
discovered Perl around the same time, which looked to be a lot more fun.

Commit: 9710ed344e3785c2a627aae3447ab566c2b51eb4
https://github.com/Perl/perl5/commit/9710ed344e3785c2a627aae3447ab566c2b51eb4
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update MY_CXT section

Revise the text in this section:

=head2 Safely Storing Static Data in XS

Commit: 8dd67ae9c1c919fd057753ddfc5c75ab96e43f5c
https://github.com/Perl/perl5/commit/8dd67ae9c1c919fd057753ddfc5c75ab96e43f5c
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update EXAMPLES section

Rewrite this section:

=head1 EXAMPLES

Basically, delete the one big example in this section and instead
provide links to various other examples already present in this document
instead.

Commit: fcf22295819a95a523e4fa2513ba8e3d8dc45200
https://github.com/Perl/perl5/commit/fcf22295819a95a523e4fa2513ba8e3d8dc45200
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs.pod: update CAVEATS, AUTHOR, A DIAGNOSTICS

Tweak the final few sections of perlxs.pod.

Commit: d08740e8b89a42c1dbcd19b20cfa463fe9f66bde
https://github.com/Perl/perl5/commit/d08740e8b89a42c1dbcd19b20cfa463fe9f66bde
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod
M dist/ExtUtils-ParseXS/t/001-basic.t

Log Message:
-----------
perlxs.pod: standardise version numbers in the doc

After the big rewrite, various bits of text which describe when a
particular feature was introduced or changed have ended up using a
random mixture of Perl, ParseXS and xsubpp version numbers.

This commit standardises on xsubpp version numbers. These are mostly
the same as ParseXS, but this handles the cases before ParseXS was
split off from xsubpp. Perl versions suffer from not exactly matching
when an xsubpp version number was incremented, and not matching what is
used by the "REQUIRE:" keyword.

This commit also adds a short new section which tries to explain how the
three sets of version numbers are related.

Commit: 4f3cd13616550ed1e2d63d00b99fe960960657db
https://github.com/Perl/perl5/commit/4f3cd13616550ed1e2d63d00b99fe960960657db
Author: David Mitchell <[email protected]>
Date: 2025-12-02 (Tue, 02 Dec 2025)

Changed paths:
M dist/ExtUtils-ParseXS/lib/perlxs.pod

Log Message:
-----------
perlxs: tweaks following code review

The many commits in this branch have completely rewritten perlxs.pod.
This commit applies all the minor tweaks suggested by reviewers.

(It was far too much like hard work to try and update each individual
commit with the various changes.)

Compare: https://github.com/Perl/perl5/compare/b16fbe94419f...4f3cd1361655

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications

[Perl/perl5] 676faf: regmatch(): improve savestack ix debuggingoutput by David Mitchell via perl5-changes

Tue, 02 Dec 2025 11:05:40 +0000

Branch: refs/heads/davem/re_debug
Home: https://github.com/Perl/perl5
Commit: 676faf69f7f048f041c0569cb843f62e6e605762
https://github.com/Perl/perl5/commit/676faf69f7f048f041c0569cb843f62e6e605762
Author: David Mitchell <[email protected]>
Date: 2025-12-01 (Mon, 01 Dec 2025)

Changed paths:
M regexec.c

Log Message:
-----------
regmatch(): improve savestack ix debugging output

Under

use re Debug => 'STATE';

S_regmatch() produces various bits of debugging output regarding
savestack manipulation; e.g. when the current index is saved; or when
the savestack is popped back to that index on success or failure etc.

Unfortunately that output is patchy and inconsistent. For example it
often talks confusingly about "Setting an EVAL scope" even when an EVAL
hasn't been executed. And it doesn't log all savestack manipulation.

After this commit there is a standard set of debugging messages, such
as:

savestack: snapping ix=14
savestack: regcppush filled in range [14..18)
savestack: regcppop freed in range [14..18)
savestack: freeing in range [10..14)

which are consistently applied.

There should be no change in functionality unless that specific
debugging is enabled.

Commit: 6dee5e6b86cbdbd86719636ad33a5aced48764fe
https://github.com/Perl/perl5/commit/6dee5e6b86cbdbd86719636ad33a5aced48764fe
Author: David Mitchell <[email protected]>
Date: 2025-12-01 (Mon, 01 Dec 2025)

Changed paths:
M regcomp.sym
M regexec.c
M regnodes.h

Log Message:
-----------
regex engine: split EVAL_postponed_AB state

(This commit makes no practical changes in behaviour except for
debugging output.)

Currently, one of the regex engine stack states is EVAL_postponed_AB.

When executing something like /(??{'A'})B/ where A and represent
general subpatterns, the engine executes the eval code, which returns
the string 'A', which is compiled into a subpattern. Then the engine
pushes an EVAL_postponed_AB state and runs the subpattern until it
reaches an END op. Then it pushes *another* EVAL_postponed_AB state and
runs the B part of the pattern until the final END. Then before
returning success, it pops EVAL_postponed_AB off the stack (twice),
executing any cleanup required. Similarly during failure, the
EVAL_postponed_AB_fail action will be executed once or twice (depending
on whether it failed during A or B).

This commit splits that state into two,

EVAL_postponed_A
EVAL_postponed_B

The first is pushed before running A, the second before running B.
The actions currently remain the same and share the same code; i.e. this
commit just does the equivalent of:

- case EVAL_postponed_AB:
+ case EVAL_postponed_A:
+ case EVAL_postponed_B:
... cleanup code ....

But it makes the code easier to understand, makes debugging output
clearer, and will allow in future for the cleanup behaviours to differ
between A and B.

This commit also fixes up a few debugging messages and code comments
which were still referring to 'EVAL_AB', which was renamed to
EVAL_postponed_AB some years ago.

Compare: https://github.com/Perl/perl5/compare/676faf69f7f0%5E...6dee5e6b86cb

To unsubscribe from these emails, change your notification settings at https://github.com/Perl/perl5/settings/notifications