7:39 AM
@BuddhiniAngelika Currently that's complicated since it requires the use of \newunderlinetype (see the lua-ul manual, especially the definition of \strikeThrough which already supports color if you're interested) but direct support for this is planned for the next release of lua-ul.

8:24 AM
@FaheemMitha If it is the example on page 3/4: did you remember to use \usepackage[referable]{threeparttablex}? That whole part of the manual assumes the use of that option.
For me adding this start, the example compiles just fine
\documentclass[a4paper]{article}
\usepackage{longtable,booktabs}
\usepackage[referable]{threeparttablex}

8:36 AM
@daleif Yes, as mentioned a bit later, I figured out that indeed I forgot to add that option.
But assuming you are the author of that documentation, I'm still a bit puzzled why you didn't just write a complete compilable example, instead of leaving off the preamble, and the final end document bit. I think people (me at least) want to quickly try out examples (assuming they are available) to see if a package is a good fit. Spending time trying to fix the examples to compile isn't exactly a welcome activity in this context.
And it's trivial to fix the examples. I could even provide a patch if you want, though it's probably redundant.
Well, trivial once you know what's going on, that is.
One thing to remember is that TeX error messages are often misleading if not totally useless gibberish unless one is an expert, in which case it might be obvious. But most of us are not experts. The number of people on the planet who can really understand TeX error messages and what they mean is probably in the low thousands, if that much.
@daleif And yes, it's the last example in the documentation. I included a Gist link.

8:54 AM
@FaheemMitha it's very common (and not bad practice) to show verbatim just the environment or command under discussion, not a complete document every time.

@DavidCarlisle For the reasons I mentioned, it's IMO better practice to include a complete document. It's that what this site recommends? That's Minimal (or Minimum) Working Document if I recall correctly. Adding a preamble correctly isn't always trivial, for one thing.

@FaheemMitha No. A package document is completely different to a question here. The fact that we ask for MWE in questions here doesn't imply anything about package documentation

@DavidCarlisle I don't see why it's "completely different". It seems to me entirely the same thing. Unless the idea is to make users lives more difficult.
It reminds me, many years ago now, of an exchange I had with a Django project developer.
The bug reporting link was embedded way inside some document, and hard to find.
I pointed out this was nuts, but according to the developer it was by design, to make it harder for people to report bugs. Because of course the first aim of any free software project should be to make it harder for users to report bugs.
But he said that if people read the documentation carefully, they'd see the link.

@FaheemMitha because here the MWE is needed to confirm what code is being used and what the issue is as the description by the OP is often wrong. In a package documentation the point of the verbatim code is to highlight the specific command being described, not lose it in dozens of lines of boilerplate.
3
@FaheemMitha see for example texdoc amsmath (amslguide.pdf) that has dozens of verbatim code sections, none of which is a complete document

9:48 AM
Does anyone know if @StefanKottwitz is around lately? I'm asking because texdoc.org is still on TL2021.

10:07 AM
@FaheemMitha In the context of a package documentation, it kind of make sense because there tend to be tens/hundreds of snippets in the documentation and repeating the preamble each time is kind of inefficient.
It's not like that API documentation of other programming languages show a compilable complete code every time anyway.

6 hours later…
3:51 PM
@DavidCarlisle did you see the report about hyperref and utf8x?

@UlrikeFischer no I just got in, will check

@DavidCarlisle I know what happens (it is the code we put in to expand the protected utf8 chars in the bookmarks). But I'm not quite sure how to resolve it.

@UlrikeFischer disable utf8x if hyperref loaded

@DavidCarlisle ;-).

@UlrikeFischer i would only do the bare minimum of support for utf8x

3:58 PM
@DavidCarlisle well imho it only needs a \def\HyPsd@expand@utfvii{}, so the main question is how to detect it (and if we really want to support it ...)

@UlrikeFischer can we get uf8x characters to be\protected (and would that help?)

@DavidCarlisle we then would need code to expand them too -- that is the main point of the command: to unprotect them for the bookmarks -- and I don't think that we want that.

If anything, I did suggest last time to make some package using inputenc and includes all the (mostly math mode) characters supported by utf8x
Anyway...
Can this question floats - How to use verbatim in figure caption (\captionof)? - TeX - LaTeX Stack Exchange be reopened? Because not all solutions in the target is applicable in this case (in particular, the cprotect solution needs some special handling with 3 nested cprotects)

@user202729 well then make the package. Nobody hinders you.
2

@UlrikeFischer (sigh) Yes I know.

4 hours later…
8:29 PM
The cure might be good, but for the wrong disease. ;-) Don't blame David when it's not his fault! :-Degreg 45 secs ago
@DavidCarlisle ^^^^^^

@egreg I assume it's @UlrikeFischer's fault?

@DavidCarlisle Unfortunately not! It's amsart's fault.

@egreg That's the location of the error, not who to blame:-)

8:50 PM
@egreg In my opinion numbering each line of cases is one disease. The second one is to put the numbers to the left. The third one to put everything in a table. And so on...

@mickep Numbering separately cases is silly, but people do it. :-( And overuse tables, I agree. Two column tables where the left column is just a title should be normal text, probably subsections.

@egreg Yes, maybe that would be the best answer (for the readers).

9:17 PM
@mickep -- Numbers to the left in amsart is intentional to avoid the possibility of having a number in the same place as an end-of-proof marker. It was an intentional style decision, based on about a hundred years of published journals. And I'm sure the possibility of putting numbered equations in a table wasn't considered at the time. Times change; opinions may change. Complain to the AMS via tech-support. (Though I don't predict any action.)

9:48 PM
@JosephWright @MarcelKrüger I have a problem with l3build. I was trying to upload a new version of tagpdf and two png are broken inside the zip. They only contain PNG
.