Joseph Wright

Fixed in dev?

@davidcarlisle I think a change would match the position we seem to have arrived at more generally, i.e. that `\Declare...` should log any redefinition.

@muzimuzhi True but in other places we have both methods, not least because the expandable functions can have lower performance. So I think for consistency-of-language we likely want this, even...

I'll see if others agree, and if so I'll assign to me (unless @muzimuzhi fancies it!).

We notice your putting in quite a bit of effort here, but we are concerned that this might not be productive. As it stands, it's not really possible to review...

The current definition of `\ref` looks wrong: it's supposed to take two arguments. For me, something like ```latex \NewExpandableDocumentCommand\JAWref{sm} {\IfBooleanTF#1{\@kernel@sref}{\@kernel@ref}{#2}} \text_declare_purify_equivalent:Nn\ref{\JAWref} \text_declare_purify_equivalent:Nn\@latex@warning{\use_none:n} ``` seems to give reasonable results, so I'd...

I think the current solution is one similar to `siunitx` uses: have a swappable definition called something like `\ref (bookmark context)`. I will raise again the idea of 'string contexts'...

I'll see if I can fix this today - we are due a release

Ok, I see why this is happening, but a fix is non-trivial. It's because we are now doing lowercase first, and that means we take the lowercase branch for the...

The '2e part' of the issue could be fixed by changing `\text_titlecase:nn` for `\text_titlecase_first:nn` in the def. for `\MakeTitlecase`. That though doesn't sort the underlying problem - how to cover...