« first day (5069 days earlier)      last day (79 days later) » 

01:40
@cfr use this

timeout 3 lualatex main.tex

it will kill the command after 3 seconds if not completed. Change as needed.
 
1 hour later…
cfr
cfr
02:45
@Nasser ah! thanks. I guess that should work with pdftex, too. my actual file doesn't do this, but my attempt to create a mwe does :(.
03:15
@cfr it should work with any linux command. Here is a page with examples
https://www.cyberciti.biz/faq/linux-run-a-command-with-a-time-limit/
cfr
cfr
03:52
@Nasser thanks, but I found the man page ;). I just didn't know which command. I like this:
> BUGS
       Some platforms don't currently support timeouts beyond the year 2038.
 
3 hours later…
06:53
@UlrikeFischer Any of your current branches meant to be PRs? @DavidCarlisle suggested we need a dev release, so want to see what I can merge
07:07
@JosephWright the pdf77... is a trivial change (only makes an unknown standard an error) and can be merged. The float one is WIP. It will be ready today. But Frank should perhaps see it as it changes multicol
@UlrikeFischer Cool
cfr
cfr
07:36
can I put a modified version of ot1cmr.fd in a test repository provided it is renamed?
@cfr Of course - we do all sorts of stuff!
@cfr Of course - we do all sorts of stuff!
@UlrikeFischer does it add float-support to multicol? :)
08:39
@cfr that's what happens if you run latex at texlive.net, you get cut off after a few seconds
@UlrichDiez I'm not sure you can.
 
2 hours later…
10:45
@DavidCarlisle @UlrikeFischer @daleif I created a question regarding the problem I mentioned: tex.stackexchange.com/questions/726567/…
@Richard we would see it without a ping here (I have already answered in a comment)
11:13
Wy does this error if placed in the preamble of a doc. Fonts not set up?
\ExplSyntaxOn
\coffin_new:N \l_A_coffin
\vcoffin_set:Nnn \l_A_coffin { 4cm } { Ag \\ Ag \\ Ag }
\ExplSyntaxOff
Works fine if placed after `\begin{document}`.
@daleif untested but probably yes. You should not do typesetting in the preamble.
@daleif fonts are not set up fully but the error is (as has always been the case in latex) you can not set paragraph text before begin document. hmode hbox can be typeset/measured in some cases, but \everypar has always given the missing begin document error.
If using coffins instead of \parbox, what extra things are needed to be manually added for it to behave like \parbox, I've notived fx that the text is indented. \@parboxrestore? Pole feature was something I could use in a project, where I have text in two boxes. The text should align on the top baseline, and the lowest baseline in the entire thing should then be placed on the overall baseline.
11:29
@daleif yes a parbox is pretty much just a vbox with \@parboxrestore
@DavidCarlisle That expain it. In my Expl conversion project, I've been using hbox_set with parboxes inside. Should probably use low level coffins instead.
A bit sad that the poles syntax is not independently explained in interface3 as xcoffins is marked experimental. So without looking in the xcoffins manual one cannot use the stuff from l3coffins as the pole notation is never explained.
@daleif On my to-do list
Also, there are still some open issues with how poles work
@JosephWright oh I was just about to blame you:-)
2
@daleif I'll see if I can address this over this week - I've been tidying up some other doc issues (for 2e), so it's a good time to do it
@DavidCarlisle Well of course
@JosephWright So the l3coffins stoff might change in the future? It is more, can I use it or not. Again, the end product here will not be public, so updating is a little funky.
11:34
@daleif I think we may need some refinement still, but we are committed to the code in general, so it would have to be to address serious issues - most cases would clearly have to be unchanged
So very low risk
@JosephWright I'll have a play with it then. Using \hbox_set ... \parbox[t]{}{....} seems rather dirty
@daleif Coffins/'saveboxes' still likely need work - the typesetting side of expl3 is less well developed than the base code fucntions
11:53
Erh, is playing with l3doc, and has \@@par in some code (do not remember why). But the tex on the ins file converts \@@par to \__module-xpar, how do I make it stop? The code is just inside a normal macrocode.
@daleif you can use @@@@ or locally turn off the %<@@=module> processing
@DavidCarlisle Ahh, that us what what thing is doing (I had stolen some code from siunitx.dtx). I just removed the <@@=module> line, I was wondering where a random - and x came from, I had <@@=module-x>.
12:11
Are other packages still messing with the class options? I'm using l3keys to setup class options as key=val, and I seem to still get Unused global option(s) for all the key=val options that were used.
12:26
MWE:
\begin{filecontents}[overwrite]{foobar.cls}
\ProvidesExplClass{foobar}{2024-09-16}{test}{foobar}
\RequirePackage{fix-cm}
\DeclareRobustCommand\normalsize{%
\@setfontsize\normalsize{10.5pt}{14pt}%
}
\normalsize

\keys_define:nn { foobar / class options }
{
A .code =,
A .value_required:n = true,
B .code =,
}

\keys_set:nn { foobar / class options }
{
B = bar
}

\ProcessKeyOptions

\end{filecontents}
\documentclass[A=foo]{foobar}

\begin{document}
a
\end{document}
@daleif Might happen, yes. There was a breaking change in the tracking of unused class options that many key=value packages haven't yet caught up to (and some never will)
What does the message say?
For the MWE
LaTeX Warning: Unused global option(s):
[A].
Note that is this not using any other packages than fix-cm and it is not changing anything.
I had thought that l3keys would have fixed this. And I thought it did, but somehow this warnings has started to show up again.
12:44
@daleif You need to use \ProcessKeyOptions[foobar/class options], if you omit the optional argument it'll try to use the file name as key-set.
13:02
@Skillmon Ahh, that was better.
13:12
@daleif or simpler use the default package name as the key collection
@DavidCarlisle I find it useful to have things split up so I know exactly what they are used for.
@daleif OK, although using foobar for the class options (to match what \ProcesskeyOptions expects, and will be seen in most other classes using this) is still possible even if you use foobar/something for more specific keys elsewhere
@DavidCarlisle Not sure how l3keys then copes with \set_keys:nn { foobar } { something/key = anything }.
@Skillmon is that "not sure" as in polite english idiom for "you are wrong" or literal translated German for "I am not sure"?
13:37
@DavidCarlisle the latter, missing an "I am". But of course "you are wrong" always applies to you :P
 
3 hours later…
cfr
cfr
16:55
@DavidCarlisle ?? overleaf also times out. but I'm inclined to abandon my attempt to create an mwe. it just crashes konsole. only I'd feel happier if I could figure out why in case the actual file can do the same thing. (it doesn't but I've no idea what the difference is.)
@cfr yes, but I don't know what they do, i just meant that specifically I use the linux timeout program
cfr
cfr
@JosephWright thanks.
@DavidCarlisle oh, sorry. thanks.
 
3 hours later…
20:25
user image
11
@barbarabeeton sorry it's only a comma, I know you would have preferred a semicolon
@DavidCarlisle Oh, so you got it. Nice!
@DavidCarlisle -- Well, it's a nice number anyhow. (I doubt I'll ever get to 6 digits, so the divider would be in the right place.)
@mickep I even took your advice and risked a couple of answers so I was on ... 2 for an hour or so this morning:-)
20:40
@DavidCarlisle Living on the edge!
 
2 hours later…
cfr
cfr
22:51
@DavidCarlisle so you can answer now & I can delete mine?
@cfr simpler to go to bed and forget about it:-)
cfr
cfr
if I can't figure out what to send ctan, should I just ask them?
@cfr they don't mind what you send as long it has a readme and some code
cfr
cfr
@DavidCarlisle but karl does care.
@cfr well yes but then you need to ask him not ctan:-) basically only other requrements for texlive are open-source things like having source of any pdf doc
cfr
cfr
22:58
@DavidCarlisle what if the source of some of your docs is generated by l3build?
@cfr I suspect it's up to you, whether to put the generated tex source on to ctan or not. we don't usually put the build script on ctan.

« first day (5069 days earlier)      last day (79 days later) »