« first day (4861 days earlier)      last day (57 days later) » 

6:54 AM
I am getting a "new" error as follows:

> Unknown option 'sections' for package xcolor.
This error somehow seems familiar, but a search doesn't show anything. And it was working before. Can anyone shed any light on this? Can I remove the "sections" option without breaking anything?
? H

LaTeX has been asked to set an option called 'sections' but the package
"xcolor" has not created an option with this name.
But this option has been here a long time.
 
7:11 AM
@FaheemMitha Example?
 
7:29 AM
@FaheemMitha xcolor now errors on unknown options. That is new. Simply remove the optio it did nothing before as xcolor simply ignored it.
4
 
 
2 hours later…
yo'
9:15 AM
@UlrikeFischer Oh this one definitely has to go to our TL2024 blog post! (Thanks for the email btw, I shall reply today I hope)
 
@yo' it is not completly new (november), and a side effect of the switch to the new latex option handler. The plus here is, that you no longer get option clash errors from xcolor, e.g. this here will load dvipsnames and svgnames and ignore with a warning the load-option monochrome:
\documentclass{article}
\usepackage[dvipsnames]{xcolor}
\usepackage[svgnames]{xcolor}
\usepackage{xcolor}
\usepackage[monochrome]{xcolor}
\begin{document}
aaa
\end{document}
LaTeX Warning: Package "xcolor" has already been loaded: ignoring load-time
(LaTeX)        option "monochrome".
 
yo'
@UlrikeFischer "november" as a month is not in Overleaf's dictionary :-) We know what August 2022 is, then July 2023 and then (hopefully) March 2024 :-)
2
 
9:49 AM
Using the l3draw package, what would be the best way to set a bounding box? In TikZ you can use a \path without stroke and fill which would update the bbox, but \draw_path_use:n { } just does nothing. I could use a box or a coffin of course ...
Probably a path with line width 0pt ... hm
 
@JasperHabicht What's the use case?
 
@JosephWright This is the right question! I want to create small icons that can be used like symbols in text. They should all have the same bounding box. This is probably best done using a coffin.
 
@JasperHabicht Yes, doesn't sound like it's part of the drawing itself at all to me
 
@JosephWright Right, except that if I could set the bounding box, I could specify where the parts of the icon sit within the box
 
@JasperHabicht OK, that sounds like you want an explicit 'set the bb' function, which is not the same as drawing a path with no stroke - I'll look at it
 
9:57 AM
By the way, I found that some macro that can add two points (especially with vector and with polar coordinates) could be useful. PGF has \pgfaddpoint. But I don't really know at what programming layer this is defined
 
@JasperHabicht what do you want the bb to do, isn't it equivalent to just setting the size of an outer tex box?
 
@JasperHabicht l3draw doesn't need explicit functions for this, as we are just using tuples: \fp_eval:n { (0,1) + (1,2) } for example works, so anywhere we want a co-ordinate expression you can just 'do the maths'
 
@JosephWright Right, I did that and created my own macro which just does this.
 
@UlrikeFischer OK. Thank you. I've removed it.
@JosephWright I don't have a short example on hand, though I could make one, I suppose.
 
@JasperHabicht You don't need a macro, you can just put the expression into the args of l3draw
 
10:07 AM
@JosephWright TikZ/pgf has a use as bounding box (plus minus spaces), I guess l3draw should have something similar as well...
 
@Skillmon Indeed. MetaPost has setbounds currentpicture to... that also sets the bounding box. Very useful.
 
But, can I do \fp_eval:n { (0,1) + ( \draw_point_polar:nn { 2 } { 45 } ) }?
 
@JasperHabicht Did you try it? ;)
 
I can! A-ma-zing!
 
@JasperHabicht One of the design aims for l3draw was to avoid needing all the \pgforigin-like stuff, and just have 'it's an tuple expression` as the basic idea
 
10:12 AM
@JosephWright Yes, these tuples are a great idea! It simplifies things a lot. I just had to wrap my head around it. Maybe you could add one example to the doc and also put \draw_point:n there which is the macro that ouputs well-formed tuples, if I am not mistaken.
 
@JasperHabicht I'll look at it
 
@JosephWright I think, one can just use draw_path_use_clear { clip } ...
 
@JasperHabicht Probably best is if I add an explicit { bb } option here
Or perhaps \draw_path_bb:n?
 
@JosephWright I have no idea how this is implemented in PGF. Maybe use their approach?
 
@JasperHabicht I'll take a look at that and at Metapost's interface
@JasperHabicht I take it you are enjoying l3draw overall :)
 
10:21 AM
@JosephWright Totally! =D Sorry, that I sometimes ask simple things. I just need to get used to it.
 
 
2 hours later…
12:26 PM
@JosephWright With the lfs thingy, I hope you are testing it thoroughly if it is going to be shipped already in TL24.
 
12:56 PM
@mickep The changes in texlua are I assume tested already (@MarcelKrüger will know more I suspect): my suggested script would only be ready post-TL'24 release and I presume subject to review before being added to the restricted-allowed list
 
1:07 PM
@JosephWright OK, let us hope that the changes in luatex do not break anything.
 
@mickep My guess is Karl, Luigi, etc., have been testing this for quite a while
 
@JosephWright OK.
 
@mickep In contrast to that, the Lua part is easy :)
 
@JosephWright The lua part of what?
 
@mickep Getting data using lfs: if the security is right at the engine level, I don't have to worry about it :)
 
1:25 PM
@JosephWright Oh, that seems logical.
 
@mickep I suspect it will also be useful for e.g tagging (@DavidCarlisle, @UlrikeFischer, @MarcelKrüger better placed to comment than me)
 
I removed tex.stackexchange.com/questions/710118 because, for a reason incomprehensible to me, I cannot reproduce the behavior I observed two hours ago. If the tex-live maintainer reads here, I kindly ask him/her to kill my issue report, too, before it appears in the archives.
@DavidCarlisle Thank you for having commented in any case.
 
1:48 PM
@JosephWright not sure about tagging, but imho getting the list of files would be useful in various places.
 
@UlrikeFischer I was thinking of MathML generation
@UlrikeFischer Yes, I think so - hence wanting to see if we can make this more-or-less automatic, à la epstopdf
 
@JosephWright the Niels Bohr example (which we haven't published) is the first one with a plausible automatic pipeline that didn't involve hand corrections to mathml generation (but does assume a locally installed latexml)
 
@DavidCarlisle Oh, Bohr.
 
2:14 PM
@JosephWright texlua is not directly affected by the current changes since all the restrictions all not applied there.
 
@MarcelKrüger ?
 
@JosephWright I mean the already merged luatex mentioned in the release announcement only have an effect when started as a TeX engine and not when starting as texlua, since kpathsea restrictions only run in TeX mode (for now).
 
@MarcelKrüger Suggests things won't work then :(
 
@JosephWright The upside is that l3build will not be completely broken.
 
@MarcelKrüger Karl didn't immediately say 'no', so now I'm confused ;)
 
2:29 PM
@JosephWright And most of the changes are basically Lua code, so we can emulate them in a Lua package which will create very similar restrictions to restricted LuaTeX without engine changes (WIP)
@JosephWright Basically Karl said that we can create textlua scripts which can be on the allowlist, we just need to ensure to implement the needed restrictions.
 
2:55 PM
@MarcelKrüger Seems more fragile than I'd hoped: I'd assumed that texlua would detect it was being run from a shell call from the engines and auto-restrict
 
3:11 PM
@MarcelKrüger Do I need to do anything or is it likely you'll have a generic loadable module?
 
3:24 PM
@JosephWright I expect that you do not have to do anything.
 
 
1 hour later…
4:54 PM
@yo' do you guys have info on who to use overleaf on a tablet (android). I wanted to test an mwe I copied from the site, but I does not seem to be able to paste into an OL doc, it does blinks on long presses
 
yo'
@daleif try to first add any text (a space might do actually) and then paste. Would that help?
 
5:25 PM
@yo' no, then I get a spelling popup
@yo' ahh, no space and then paste helped, thanks
 
yo'
@daleif oh I have that off...
@daleif yeah, sorry for the glitch. We hope to bring a paste button...
 
@yo' it's hard to support all devices.
 
yo'
@daleif yeah
 
@yo' i don't usually use OL, but was on tablet only and thought this mwe was strange so wanted to give it a go
@yo' can we read the generated toc file?
 
yo'
5:46 PM
@daleif yep, under the logs
(well, you can download it)
 
@yo' there is a download option for the toc file?
 
@yo' i don't have such buttons under logs. I get the two warnings and can unfold the log. Don't see those buttons. Perhaps screen to narrow.
 
yo'
@daleif looks like a bug to me, sorry! (And thanks, I haven't realized this problem before!) This is a stupid one and I'll file a ticket for it internally. For the time being, you can append ?ide-page=default to the URL within the project and it should help.
 
How would I safely undefine a command that is defined using \NewDocumentCommand and the like? It feels that \let ... \undefined might not be the right tool for it and if I say \NewDocumentCommand{\foo}{ }{ } \let\foo\undefined \NewDocumentCommand{\foo}{ }{ } I get a warning about '\foo code' and/or '\foo defaults' not being undefined. What is the meaning of this?
 
5:59 PM
@yo' that helped. Now I just need a text editor on this tablet
 
yo'
@daleif yeah, Or you can try something like \verbatiminput{output.toc} but it will probably fail for the file not being closed when you input it. That's solvable with a latexmkrc file.
Just please don't reupload the TOC file to Overleaf or if you do, rename it, or bad things will happen!
 
@yo' no need, seems the problem is somewhere else (seems to be a sporadic space in some tocbasic code)
 
yo'
@daleif If I can ask please, do you know the Android version and which browser it is?
 
@yo' its chrome on android 10
 
yo'
@daleif Thanks!
 
6:20 PM
@JasperHabicht I'd say 'you don't' - what's the context?
 
@JosephWright =D Well ... maybe I should do things differently anyways. I created a "modern" version of a package that now uses l3draw instead of TikZ. I want the commands to be the same. Now I want to create a doc for both packages that also shows how to use them with examples. So, I kind of need to switch the definition of some commands mid-file ...
 
@JasperHabicht So you can just \DeclareDocumentCommand to overwrite
 
Well, I would like to grab the commands from the .sty file ... So what I currently do is: first load the old package, copy the commands, undefine them. load the new package, copy the comands, restore the "old" commands ... later restore the "new" commands
Actually, it is only two commands ... So, I could also just copy the defintion into the doc ...
 
6:47 PM
@JasperHabicht ;-) My rule of thumb is do not think about automation only you have at least five cases.
2
 
@UlrikeFischer Yes, it's just ... from single-cases sometimes arise questions which I find interesting to think about =D Actually, this is kind of the evergreen question: how to only grab one or two commands from a package.
SInce I can freely change the definitions in the package, I probably just define a copy in the package with a different name and then I can in fact use \DeclareDocumentCommand. This will be the easiest.
 
yo'
7:13 PM
@JasperHabicht do this?
\let\daleifStoreNewDocumentCommand\NewDocumentCommand
\let\NewDocumentCommand\DeclareDocumentCommand
\usepackage{version2}
\let\NewDocumentCommand\daleifStoreNewDocumentCommand
 
@yo' Well, there was another twist in that, because version2 used \ProvideDocumentCommand =)
 
yo'
@JasperHabicht ouch. And you don't want to do this hack with \ProvideDocumentCommand as that actually changes something...
 
@yo' I just went the other way around: I gave the commands in version2 different names and then check whether the commands are defined, if not, copy these. In the doc, I then can easily use \RenewCommandCopy. It is cleaner anyways.
@yo' In order to have \ProvideDocumentCommand kick in, I had to undefine the commands first. Also, \let cannot be used with these commands. I had to use \NewCommandCopy ... and this eventually let to my question =D ... This approach was obviously too complicated to begin with
 
yo'
Btw, I keep the xparse command as basically wrappers around their expl3 equivalents, so usually my "public" commands are very simple, at most a few conditionals on aruments being present... Then it's quite easy to do these things.
2
 
@yo' That's ... the plan
 
yo'
7:26 PM
@JosephWright :)
 
@yo' Yes, I typically only define the user-level commands using the xparse commands.
... but in the doc file, I am the "user level" right? =)
 
7:48 PM
@yo' oy, what did I do?
 
yo'
@daleif oh nothing, wrong nickname in the command! :-( I'm so sorry for messing this up -- my brain just wired it wrong as we spoke about other things before.
 
@yo' all good, was just, what did I do now
 
@yo' haha, I did not even notice ...
@daleif feel free to use my nickname in your code snippets as well =)
oh, I don't have a nickname acutally ... hm
 
8:08 PM
Hi to all users into chat.
@UlrikeFischer Hi. Ulrike I have the same problem of rendering with TeXworks. I have sent on github/texworks/issue a message to one of the mainteiners of the editor. Do you can do something?
@UlrikeFischer Best regards
 
8:30 PM
```\documentclass{article}
\ExplSyntaxOn
%\NewDocumentCommand \foo { m } {old}
\cs_if_exist:NTF \foo
{
\RenewDocumentCommand \foo { m }
{
\__foo_code:n {#1}
}
}
{
\NewDocumentCommand \foo { m }
{
\__foo_code:n {#1}
}
}
\cs_new_protected:Npn \__foo_code:n #1
{
Some~new~code~#1
}
\ExplSyntaxOff
\begin{document}
Hola \foo{Hola}
\end{document}``` Creo que tengo algo así en uno de mis archivos antiguos
Can any ConText user tell me what the context --luatex file.tex line does (I can't find --luatex in context --help)?
 
@PabloGonzálezL Uses LuaTeX rather than LuaMetaTeX
 
@JosephWright Thanks :D ...with that I can solve a headache with the current version of context.
@JosephWright where is the --luatex option documented?
 
@PabloGonzálezL ER, that I don't know
 
@JosephWright Snifff jajjaj
@JosephWright Question, the documentation for xparse was mostly moved to usrguide...am I right?
 
@PabloGonzálezL Yup
 
8:39 PM
@JosephWright Ok, but not in its entirety, missing the explanation of the starry environments (which differ from the s argument) and comments for verbatim
 
@PabloGonzálezL If we missed bits, that's an oversight: I'm not sure I know which bits you mean
 
1.3
Verbatim arguments
Arguments of type v are read in verbatim mode, which will result in the grabbed argument
consisting of tokens of category codes 12 (“other”) and 13 (“active”), except spaces, which
are given category code 10 (“space”). The argument is delimited in a similar manner to
the LATEX 2ε \verb function, or by (correctly nested) pairs of braces.
Functions containing verbatim arguments cannot appear in the arguments of other
functions. The v argument specifier includes code to check this, and will raise an error if
1.7
Starred environments
Many packages define environments with and without * in their name, for instance
tabular and tabular*. At present, xparse does not provide specific tools to define
these: one should simply define the two environment separately, for instance
From xparse document (p5-6)
@JosephWright I've been meaning to make an issue about this for a while, but I know that documenting it takes a lot of time, ...you could put a new xparse introduction in l3keys2e style in which you make it clear that you need to load it explicitly only if the kernel is older than 2020 and guide the user to userguide.
 
@PabloGonzálezL Sure
 
@JosephWright Today you only need to upload it if you want to use GetDocumentCommandArgSpec, GetDocumentEnvironmentArgSpec or if you want to use log-declaration.
 
@PabloGonzálezL All of which are deprecated ;)
 
8:50 PM
@JosephWright heheh, also in xparse there are some messages This function is experimental ...come on, clearly that is not true :D
 
@PabloGonzálezL WAs true in xparse: the stable stuff is in the kernel
 
@JosephWright Interesting :D
 
yo'
9:02 PM
@daleif Thanks!
 
@DavidCarlisle One question (just out of curiosity) why the line \cs_new:Npn { \IfValueF { \tl_if_novalue:nT } from source2e (p189) takes the parameter p and is not simply \cs_new:Nn (as it does not take argument #1) ?
 
@PabloGonzálezL New intro added
 
yo'
@PabloGonzálezL I don't have source2e open, but I guess that it's an optimization and the argument is simply not repeated, so you need to force the wrong args actually.
 
@JosephWright Ohhh....too fast....:D
 
@PabloGonzálezL There are two schools of thought on Npn vs Nn - Frank and I prefer Npn as it's faster in setup
@PabloGonzálezL As @yo' says, this is slightly faster as you only read the arg once
 
9:11 PM
@JosephWright Ahhhh...OK, I struggled at first to follow the Nn and Npn rules, but after several (many) egreg's answers I stuck with the rule of "use p if you are going to use #1etc.
@JosephWright Do you think it would be wise to place some promotion of the current usrguide document in the new xparse introduction?
 
@PabloGonzálezL I've put the comment similar to l3keys2e in - I think that covers it
@PabloGonzálezL It's not an issue loading xparse - just not needed
 
@JosephWright I know, but, usrguide is much more complete actually, it brings information about NewCommandCopy and ExpandArgs and examples...and almost nobody reads the news, it is simpler to write texdoc xparse when one doesn't remember how to work some argument (maybe it's just me who does it :)
 
9:28 PM
@PabloGonzálezL \cs_new:Npn { \IfValueF ? where's the N argument?
@PabloGonzálezL but as Joseph says using p is faster
 
@PabloGonzálezL Easier would likely be to ask TL to alias such that texdoc xparse opens usrguide, and texdoc -l xparse is needed to get to xparse.pdf
 
@DavidCarlisle :D
@JosephWright Mmm...if I remember correctly there was a list for texdoc...but I doubt I can influence that :(
 
@PabloGonzálezL you can set it locally or ask on texlive (or better, texdoc) list for the defaults to be changed
 
@DavidCarlisle Mmm...not a bad idea :D
@DavidCarlisle Any reason why you didn't put the "show log" button at https://texlive.net/run?
 
@PabloGonzálezL %!tex log
 
9:36 PM
@DavidCarlisle Wow...that's magic
 
@PabloGonzálezL DOCUMENTED FEATURE: YOU SHOULD READ MY DOC
Mar 26, 2012 at 19:37, by David Carlisle
@Canageek moral of the story: never read the documentation, bad things happen
 
@DavidCarlisle Jejeje...Where are all the %something in texlive.net documented?
@DavidCarlisle Blame my davidcarlisle.github.io/latexcgi
 
@PabloGonzálezL there is a blue bit at the top of the page that says "documentation" and gets an underline when you hover on it. It's called a link...
 
@DavidCarlisle Sniffff...jajaj
@DavidCarlisle Copyright 2020–2023 David Carlisle ...mmm 2024?
 
@PabloGonzálezL if I change it, no need to change it just to change the copyright
 
9:47 PM
@DavidCarlisle Just details :D
 
@PabloGonzálezL We've got a PR still to sort that relates to this
 
@JosephWright I saw a thread in the latex2e repo sometime, but now I can't find it, the actual document has some margin notes (only 2) and no index or references like the rest of the LaTeX Project documents.
 
@PabloGonzálezL I think we decided this was more confusing than helpful
 
@JosephWright It depends on your point of view or how you learned or where you got your first examples to learn from, from my perspective it is useful to differentiate the * argument for environments.
 
@PabloGonzálezL We tried that ... didn't work well!
 
10:02 PM
user image
9
 
@JosephWright Hehehe...that's true, but that doesn't mean it won't be there...it's like the writing style in expl3, it's well documented and the vast majority of authors follow it...although others seem to ignore the rules :D
 
10:37 PM
@PabloGonzálezL but there isn't really a starred environment. tabular* is no more related to tabular than tabularx is, it's just a name.
 

« first day (4861 days earlier)      last day (57 days later) »