Morning all!

@AlanMunn @DavidCarlisle looks ok for me. I had to force pdfcrop to use gswin32 as without it is tries to use an external perl and there is (again) a revision mismatch. pdfcrop --verbose --debug --gscmd=gswin32c test-utf8.pdf output.pdf. Your output looks as if it is shifted, you have more border on the left and the bottom than I have. What happens if you add \usepackage{graphicx} to force the papersize setting?

@JosephWright I guess expl3 allocation is wrong for platex as well? github.com/texjporg/ptex-manual/issues/6#issuecomment-544111717

@DavidCarlisle I guess: we've been working somewhat in the dark

@UlrikeFischer which gs version is your gswin, for me gs --version says 9.27

@JosephWright we could do with an english language manual:-)

@DavidCarlisle On the spaces-at-end-of-filenames business, I was thinking of stripping spaces from extensions. Windows ignores them at the end, but does allow for example "foo. tex", but that seems so unlikely to be useful that I think we could forbid it. OK?

@DavidCarlisle Space trimming a whole file name is easier: I wonder if " foo.tex" is really useful ...

@JosephWright yes

@JosephWright I don't think you can access files with " in their names via web2c at all, if you access fo"oo.t"ex via the web2c utilities it will open foo.tex

@JosephWright oh that was a space at the beginning rather than the quotes, sorry

@DavidCarlisle Yes, sorry, I mean a file called [space]foo.tex: as a 'Windows person', I never imagine " being in the file name

@JosephWright personally I'd bring back \zap@space and remove all spaces, but trimming front and back is quite reasonable.

@DavidCarlisle Yes: you can't create file names on Windows containing " (at least, not without causing yourself issues)

@DavidCarlisle :)

@DavidCarlisle I think the gs pdfcrop is using if I call it as above is the one in texlive ltpkg/tlgs/bin, and this has version GPL Ghostscript 9.50 (2019-10-15).

@DavidCarlisle I'll go for start-and-end, and then document the approach in l3file

@JosephWright an edge case is foo .tex as if you trim the full name the space stays but if you access it as {foo } and the macro layer trims the supplied string then it goes.

@DavidCarlisle, @UlrikeFischer Once that is done, I guess back to pdfTeX and \pdfmajorversion: it's going to be non-trivial ...

@DavidCarlisle Hmm, good point

@DavidCarlisle I guess we should agree on this on the team list

@JosephWright yes, I remember it from last year. And I seem to have broken my pdfs after changing the tagpdf ;-(.

@JosephWright I think just document that. (you get similar if you \input{zzz/ file} as opposed to \input{file} if the zzz folder is in TEXINPUTS

@DavidCarlisle Ah, right, yes: I'll make some adjustments in a bit

@JosephWright \zap@space treats {zzz/ file} and { file} equally :-)

@DavidCarlisle But we may never need l3alloc ... (though perhaps a longer-term Secret Plan outcome ...)

@JosephWright yes I was assuming secret plan

@DavidCarlisle Ah, but it's not in l3alloc, is it, it's actually in package mode anyway :)

@JosephWright I know nothing:-)

@DavidCarlisle Oh come on, just because I re-wrote more-or-less everything is no excuse ;)

@JosephWright can one run additional tasks in the build folder when doing l3build checks? Intermixed with the checks? I would like (sometimes) to test the pdf for syntax errors after l3build has done a successfull check.

@UlrikeFischer We'd need a new hook for that

@DavidCarlisle This is a bit more complicated than it looks: for upTeX, it's fine, but in pTeX we have an issue as we use \chardef or \mathchardef to store the max register, and we can't use either ...

@JosephWright as Phil commented in tl list we are not to a byte or two these days, we could use \def

@JosephWright perhaps I can do it differently, I can run first e.g. l3build check -eluatex and then a script which loops over the pdfs.

@JosephWright or probably more compatible use an allocated count register to hold the value

@DavidCarlisle Well it's an int

@DavidCarlisle Yes, that would be much easier

@JosephWright 2e is less fussy about these things

@DavidCarlisle Still tricky with boxes ...

@DavidCarlisle Oh, they have the Omega things ... I should probably extend l3names first

@JosephWright oh you mean the definition command to use in the allocation, not just the top value, hmmm

@DavidCarlisle Yes, exactly

@DavidCarlisle I agree on the 'top' business, but there's also the \int_const:Nn code ...

@DavidCarlisle I'll do the primitives first, then work on this

@JosephWright what does \omathchardef do, is that usable for this (or \kchardef)

@DavidCarlisle Yes, it seems to be: I'll start on some adjustments

@DavidCarlisle \kchardef is only upTeX

@JosephWright so it is

@HironobuYAMASHITA we could use \omathchardef instead of \chardef in register allocation?

@HironobuYAMASHITA ah I see the github comment now, OK you fix this in the platex files, so we leave the base code as it is? (no problem to change it there as well if you prefer)

@JosephWright github.com/texjporg/platex/issues/3#issuecomment-544117355

@DavidCarlisle :)

@DavidCarlisle I'm adding all of the primitives now, then will adjust our allocators

@DavidCarlisle Don't worry, expl3 won't load with Aleph: no \pdfstrcmp

@DavidCarlisle Let's be honest, adding primitives is more fun than Brexit!

> \Brexit=\Brexit.
l.1 \show\Brexit

@DavidCarlisle ^^^ TeresaMayTeX ;)

@JosephWright yes I was thinking of the 2e allocation

@DavidCarlisle Primitive list expanded, but I'm still not sure what most of them do ... (\hfi???)

@JosephWright \hfi is \hfill with two less l

@DavidCarlisle We really need to have a proper conversation with the pLaTeX people and see what could be moved from platex.ltx to the kernel proper

@DavidCarlisle Oh, of course!

@DavidCarlisle Primitives done, will look at allocators later (stuff to do)

@JosephWright I think I'm going to try to build a public bidi release...

I live the way I type; fast, with a lot of mistakes.

[paulo@cambridge ~] $

@JosephWright LOL

@PauloCereda You could come and live in Milan: "se vai piano non sei di Milano" (the original proverb is "chi va piano va sano e lontano"), since everybody goes fast here, and in the other parts of Italy they make fun of us due to it :)

@CarLaTeX but I thought piano piano se va lontano :)

@josephwright @davidcarlisle I noticed here just now; 65536 registers of e-(u)pTeX are available only when both of two conditions are met: (1) e-TeX extended mode is enabled, and (2) Omega primitives are built-in. The condition (2) excludes pTeX-ng by Clerk Ma (github.com/clerkma/ptex-ng), which is largely based on upTeX and e-TeX but not on e-pTeX.

@JosephWright The proceeding will start shortly

@PauloCereda That's the Roman version :)

@CarLaTeX What have the Romans ever done for us? :)

@HironobuYAMASHITA oh thanks for the info we love having yet another slightly incompatible engine to support:-)

What have e-TeX extensions ever done for us

@DavidCarlisle woo

@PauloCereda They were Trump's friends, you know? (When Trump met our president, he said that US and Italy are friends for millenia)

@CarLaTeX oh

Have been, sorry for my English

@PauloCereda and of course every Italian comedian made fun of Trump because of it :)

@CarLaTeX ooh

@UlrikeFischer Adding graphicx doesn't change anything for me.

@UlrikeFischer @DavidCarlisle Seems like it was a ghostscript bug indeed. bugs.debian.org/cgi-bin/bugreport.cgi?bug=927429

@AlanMunn yes sounds like it, it also explains why I don't see it.

@UlrikeFischer The current version is 9.50 which is quite a few iterations from the 9.27 that we were testing with.

@AlanMunn Why don't you have 9.5 like me?

@UlrikeFischer I assume that's the version that shipped with MacTeX, and it doesn't get updated with tlmgr I don't think.

@UlrikeFischer I usually don't mess with it myself. I'm updating now though.

@AlanMunn imho mine on Windows got updated with tlmgr, (but it is a bit of a pain, as the resources sometimes interfere with the external gs and I have to update this too.

@UlrikeFischer And now I know why I don't mess with it... homebrew won't install new packages over things it didn't install originally. And there doesn't seem to be a Mac release on the ghostscript page.

Good afternoon everybody, hi.

@StefanKottwitz @JosephWright @Martin Scharrer Good afternoon from Sicily to you, dear moderators. I simply wanted to report to users who use TeXworks necessary to compile in LaTeX, that two small errors have been fixed by me reported as the link attached: github.com/TeXworks/texworks/issues/864 Best regards.

github.com/TeXworks/texworks/commit/…

Good afternoon to the new enter into chat. Best regards.

@UlrikeFischer @DavidCarlisle Updating to version 9.50 fixes the problem.

@CarLaTeX well, after all, some very good American movies are Italian :-)

@yo' -- I assume you're not referring to "spaghetti Westerns".

@barbarabeeton I am of course! Especially Spencer&Hill

@yo' -- I'll not criticize your assessment; you undoubtedly go to the movies more often than we do.

@yo' :):):)

@CarLaTeX just been to a pizza place. Can you believe it? No pineapple on the menu.

@DavidCarlisle -- I do hope they didn't ban anchovies. I know not everyone likes them, but they should be available for people who do.

‏ ‮bidi gone to ctan

!looc

@mickep you wrote that backwards, that's fake bidi:-)

I've been invited to the GitHub Sponsor beta ...

@DavidCarlisle ?looɔ

@DavidCarlisle Been a bit busy today, but should have some expl3 changes tomorrow for the (u)pTeX alloc business

@JosephWright I got invited last week and in the first few hours got swamped with random chit chat about people asking each other what to do, I unwatched it, I might sign in to teh financial forms or I might not, haven't decided yet

@JosephWright been out all afternoon as well (pizza mentioned above was at same one we use at dungeon meetings:-) family theatre trip in to town

@mickep U+202E is your friend:-)

@JosephWright this quotes around filenames stuff is tricky....

@DavidCarlisle Ooh, unbelievable! A very good pizzeria!

we are back in germany, the Bär travelled as stowaway: