@UlrikeFischer This case is extreme, but I almost always have some I can't resolve. In those cases, I fix the ones I can guess and generally leave the others broken. I guess I just don't know how to track them down.

How can I use two different values of some fontdimen in a single equation?

@Jack What are you actually trying to do?

@cfr Here's the result, with very minimal work amunn.github.io/msu-thesis/msu-thesis.html

@AlanMunn I seem to remember it was the packages I wanted to use which made it so complicated. But I can't remember what I was trying to use it for. I suspect I was trying to convert to Word via some intermediate format.

@AlanMunn Do you use arara? If so, do you know if it's possible to set TEXMFHOME as part of a directive?

@cfr I do use arara but for very limited things so I don't know it well enough to know if you can do that. You can definitely run arbitrary shell commands in a rule so in principle I'm sure you can. But @PauloCereda would be the definitive source.

@AlanMunn Yes, I should ask @PauloCereda. I'd like a way of using arara for examples/packages, but I avoid it currently because I don't want to risk contamination from TEXMFHOME.

@cfr lower a superscript past the usual minimum height and be able to pack it in one command

@Jack I'm off to teaching, but that sounds like a hackish solution for a problem that might have a better solution. What is the real use case/problem?

@mickep I want to overline a symbol squared and it's a bit too tall and lowering it also overlaps the superscript

@Jack Cramped?

@mickep overline already makes it use cramped and i want it to go a bit further

@Jack But you also want consistency, I guess? Through the document that is.

@UlrikeFischer I see so far reports of issues with PR1 notable by their absence ...

@JosephWright sounds good!

@Skillmon Indeed

@Skillmon I probably need this week to get a coherent branch: 'week 0' at work so I will be quite busy

@JosephWright 'week 0' == next semester is about to start? (at least that's the case for my sister in law)

@Skillmon Yup

@Skillmon Teaching starts next week, is welcome Events for new students

@JosephWright yes ;-). I have adapted the pdfresources and tagpdf to use the new properties (if they exist) and will probably upload this week and then look if something breaks ...

@cfr I would really need an example. As I said it is actually rare, in the LaTeX documentation I found only two related to footnotes and they are there because Frank tried to fix the footnotes in the variable environment and actually lost them ... (But footnotes, more precisely footnotemark/footnotetext are the most probable candidates for problems).

@UlrikeFischer I'd seen

@UlrikeFischer Meanwhile, I'm trying again to solve the x vs e conundrum by showing we can just switch to e everywhere :)

@JosephWright I would love that. I got various errors while switching because I confused ee and xx ...

@UlrikeFischer I think it works, and I think we might not have the performance issues too widely: when you look at e.g. \cs_gset:Npx and think about hashes ...

@UlrikeFischer My plan is to get a PR sorted this week and push for this before the fall 2e release, so we can tighten up in the 2e expl3 modules too

@DavidCarlisle Argh - Unicode have used a 'minor' release to change the formatting of some of the data files!

@JosephWright nice. Backwards compatibility and semantic versioning are just minor issues, no need to think about stuff like that when you need to include another 100 emojis.

@Skillmon I see what's happened: it probably 'looks' minor as most people like trim whitespace as a matter of course from all lines, but changing XXXX;YYY to XXXX; YYY makes a difference to a TeX delimited argument ...

@Skillmon github.com/latex3/unicode-data/commit/…

@AlanMunn \@percentchar

@JosephWright main UnicodeData.txt didn't change format I think?

@DavidCarlisle No, but EastAsianWidth.txt did: I got a test failure (now fixed)

@DavidCarlisle github.com/latex3/unicode-data/commit/… shows the diffs

@DavidCarlisle Also LineBreak.txt

@JosephWright yes, just checking I hadn't missed anything in unicode.xml update (which currently is UnicodeData + Blocks + MathClass +historic iso/tex/stix data

@DavidCarlisle It looks like someone 'tidied up', but really this was not a good idea ... at least should have been flagged up and really not hidden in a minor release

@JosephWright have you found a spec for the data file ; format? TR25 comes with MathClassEx.txt with an interesting line 003B;P;;;semi;ISONUM;;SEMICOLON

@JosephWright my initial attempt at parsing that failed to notice the 3rd ; is data

@DavidCarlisle eek, really? without quotes to excape it?

@DavidCarlisle Haha, wonderful!

@DavidCarlisle :)

@DavidCarlisle Ooops

@JosephWright the MathClassEx version is supposed to be an expanded version of MathClass with additional fields including the actual character (hence the ;) but it all seems to have been done by hand without any thought about parsing to use the information

@DavidCarlisle Er, someone didn't think this through properly

@cfr @AlanMunn I think it's possible, I will try to draft something and see if it works. :)

@PauloCereda That would be great, though I was more asking if it could be done already than expecting an extension!

@cfr hmmm let me check the documentation, maybe there's something already implemented... hold on

@cfr which engines would you like to run? pdflatex?

I think the answer is I should put something into my own functions. I asked in case there was something I was missing - some standard way to get hyperref to spit out the line numbers where destinations are referenced. But if there isn't, there isn't ;).

@PauloCereda pdflatex almost always, yes.

@cfr cool, thanks!

@cfr Tradtionalist. :)

@AlanMunn I can get the fonts to do what I want more easily in pdfTeX. The only disadvantage is support for some accented characters in Welsh, but pre-composed glyphs don't always exist anyway :(.

(And compilation is faster.)

I found the final broken hyperlink by dumping \hyperlink arguments to the log file ;). @UlrikeFischer Comment above lost the 'ping'.

@cfr That's true. But I'm totally over dealing with font encodings. And I like that I can use any of the fonts I've collected over the years.

@AlanMunn Honestly, I like that in theory, but I almost always end up using Latin Modern in practice.

With a few ADF fonts when I need something specifically different.

@cfr My goto font is Minion Pro. But I also make posters and for those a wider range of fonts is quite useful.

@cfr Myriad Pro is a much nicer sans font for posters than Helvetica, for example.

@AlanMunn I've made one poster recently. Never needed to do it before ;). I like having italic small-caps, real small-caps etc.

@cfr I hope you used tcolorbox with its poster option. Although given it's you, you probably did the whole thing in forest. :D

@AlanMunn I think I deliberately left the fonts alone. Whatever the class I used was, I tried not to touch the fonts or page layout and hoped that would be roughly what was expected. I figured whoever wrote the class must be less clueless than me.

@AlanMunn I seem to have used tikzposter. Why?

And, actually, I did change the fonts. I used cfr-lm ;). What I didn't change was the size.

@cfr Having tried the various poster classes over the years nothing has come close to tcolorbox in being both fully featured and really simple to use.

@AlanMunn If I ever have to do it again, I'll try to remember that. I only have free Minion Pro, so I tend to only use their source code font, which I like for use on screen. Latin Modern isn't a great choice, but I think the result was OK.

@AlanMunn I probably just searched CTAN for 'poster' and went with the first reasonalbe-looking choice.

@cfr I like sans font for posters. E.g.:

@AlanMunn To be honest, the tcolorbox option wouldn't have worked for me. I was specifically looking for something much more template-like. I had no clue how big the damn thing should be. (They said they'd send details to people whose posters were accepted; the details turned out to be 'PDF preferred; image accepted'. Should it be A4? A3? A0? I had no idea.

@AlanMunn It does use sans.

@AlanMunn This definitely looks nice. I think diagrams help, too. I just couldn't think of any useful ones. (Most people had data & diagrams, but I just had reasons to work with.)