@matmarex I believe you wrote the PagerNavigationBuilder, Can you see any potential pitfalls with this?

No objections from me. It seems that PagerNavigationBuilder won't even need any changes, since it already allow customizing all of the query parameter names. Only IndexPager::getNavigationBuilder() and the related methods need some updates to allow them to be customized (maybe moving the strings to constants that can be overridden in subclasses, or something).

A few thoughts on this:

• Moving these names to a constant/declarative method seems a good idea in principle. Constants would be the simplest, I think.
• I also considered adding a method that would allow subclasses to specify a prefix (or suffix) for all the relevant parameters. This has the advantage of not requiring subclasses to change if the list of parameters is changed; and it generally hides these parameters from subclasses. The problem is, subclasses might need the parameter names to do stuff, and obtaining the full name given only the prefix would be awkward at the very least.
• There seem to be a bunch of pager classes that could use the newly-introduced constants instead of hardcoding parameter names. Fixing these is not required, except for the public interfaces in core (IndexPager, TablePager, CodexTablePager, perhaps ContributionsPager. The problem with hardcoded strings is that we can't deprecate them, but we could perhaps add a line to the release notes.
• One thing that complicates this refactoring is WebRequest::getLimitOffsetForUser(), which hardcodes both the 'limit' and 'offset' parameters. Clearly it would be easy enough to add two parameters to override these names, but that doesn't help with the fact that this method doesn't belong in WebRequest. It might be a good time to move it elsewhere, but where? I think it only has a purpose in classes that deal with pagination; the problem being, these classes don't necessarily have a common ancestor (SpecialPage and IndexPager being the two most common). Perhaps it could be moved to a new trait?
  • I'm also generally really concerned about what other places might be making the assumption that 'limit' and 'offset' always have these names. And this is for everything from MediaWiki code to the CDN (for example, if there are rules to detect expensive queries).