@UlrikeFischer — I think I’ve fixed up the new language bug, and if I can fix the extensions bug quickly I’ll wait to do that before another release — but if I can’t do that quickly I’ll release some time tomorrow I hope

@WillRobertson fine with me. I'm sorry, but I can't test now, on my way to Japan soon ...

@UlrikeFischer have a good trip

Safe travels :)

Is there an obvious way to align that equals symbol with the first line of a matrix environment within an aligned environment?

(I'd like to maintain the matrix and would prefer not to break it up into several lines which is what has often been suggested in my searches for a solution)

\begin{matrix}[t] although it would be better to use alignedat rather than matrix (aligned also has a [t] option)

Thanks, David. I suspect I have something wrong with my preamble because [t] is just getting typeset as part of the first term [t]y_{1}a_{1}. Will try and figure it out!

@Khallil matrix is for matrices and sets its content as inline math, you are beter to use aligned which sets its content in display math to match the outer alignment,

@Khallil actually I mis-remembered aligned and array have [t] but matrix doesn't

@DavidCarlisle I thought mathtools had it, but that was "only" horisontal alignments, it seems.

@mickep yes crazy really since it's such a thin wrapper around the standard array it could easily have supported that, maybe we could add it one day although always the issue of existing content with [ intended to be typeset

@DavidCarlisle But mathtools has \begin{matrix}[r], so if you add it there, it could be safe.

Oh no, it is matrix*

@mickep would probably need to be keyval so [valign=t,columnalign=r]`

@DavidCarlisle Oh, that is then a bigger change. You can have {matrix**}[key=val]... :)

@mickep \input context

@DavidCarlisle Oh, when you say it, it might actually be possible to vertically align matrices already.

@UlrikeFischer no problem, have a great time!

@UlrikeFischer Have a good journey!

@samcarter thanks. Bär is a bit nervous, he doesn't know if they have chopsticks in his size.

@UlrikeFischer Tell him they do not eat bears in Japan, so he can relax...

And have a nice trip!

@UlrikeFischer That's a very important point to worry about! These should fit etsy.com/listing/1294361774/…

It seems it is popcorn time again :)

@mickep don't let @UlrikeFischer see this theguardian.com/world/2023/apr/03/…

@samcarter I wonder what you mean ;)

@JosephWright :)

@DavidCarlisle It's a well kept secret.

@samcarter ?

@UlrikeFischer I find the non-human tex user discussion quite entertaining. When ever one of these emails pops up, I simply imagine a room full of ducks and Bärs typing away on their latex documents and have to smile :)

@samcarter You could say they should be talking about beings vs non-beings

@JosephWright ... given how easily the texlive list derails, this sounds dangerous. We might be getting mails with philosophical discussions for the new few years :)

Long live array

@mickep Anime lied to me

@JosephWright could you look at tex.stackexchange.com/q/717343/2388? I can't test now.

@UlrikeFischer Will do - beats marking scripts ;)

@UlrikeFischer Yes errors on the second run

@UlrikeFischer I'll see what's gone wrong: must be an oversight in the deprecation code

@Khallil I bet they also had a top-aligned matrix in it.

@DavidCarlisle FYI 2401.09436v1 does not seem to validate in Arlington 1.26.0-RC3.

(I better not ping the traveler...)

@DavidCarlisle Yes, but I dont't want an svg-font. I would like to create a font, which is in the result like TwemojiMozilla.ttf; but the sources of the glyphs are svg-files, better to say ai-files.

@mickep what does it complain about?

@UlrikeFischer I don't remember, I do not have it loaded at the moment. But I can check.

@UlrikeFischer That is what is there. We tested a few context pdf files, and so far the strangest thing was that arlington claims that Resources is required in XObjectFormType1, even though it is optional according to the pdf 2 standard. Adding empty ones is just weird...

@mickep ah, that is known: the document includes a pdf which is not pdf 2.0 and which uses deprecated settings. Not much we can do here.

@mickep you can open an issue ... or if I see an uncompressed example I can do it.

@UlrikeFischer That must happen often, so I guess it would be good to be able to mark included files.

@mickep I ignored warnings if it didn't say it was a hard failure and in some cases Peter Wyatt implied the deprecated features would be undeprecated github.com/latex3/tagging-project/discussions/… In the end I used the arlington checker from the current github sources not the verapdf one, ie the C++ code from github.com/pdf-association/arlington-pdf-model/tree/master/… @UlrikeFischer

@UlrikeFischer I could perhaps open one even without example.

(I should probably just first check that arlington is not really stating somewhere that they require the Resources, but it does not makes sense in general...)

@DavidCarlisle Ah, you are a more advanced user. :)

@mickep some warnings but no "error"

BEGIN - TestGrammar v0.81 built Mar 30 2024 11:21:56 (GNU-C linux debug) PDFix v6.20.0
Arlington TSV data: "/mnt/c/cygwin64/home/pdfa/arlington-pdf-model/tsv/latest/"
PDF: "/mnt/c/cygwin64/home/FrankMittelbach/Accessible-basic-elements-tagging/examples-new/ua-2/2401.09436v1-tagged.pdf"
Info: Traditional trailer dictionary detected.
Info: Header is version PDF 2.0
Info: Processing as PDF 2.0
    4884:               Trailer->Root (as Catalog)->Pages (as PageTreeNodeRoot)->Kids (as ArrayOfPageTreeNodeKids)[3 (as PageObject)]->Resources (as Resource)->XObject (as XObjectMap)->Im1 (as XObjectFor…

@UlrikeFischer Issue identified: I guess I do a dev release ASP then we can do a main release in a couple of weeks to address

@DavidCarlisle I do not see any differences between warnings and errors. Edit: Ah, in your output. I see!

@mickep errors are test failure, warnings are a mild slap on the wrist.

@DavidCarlisle In vera it is not so clear what is a warning and what is not.

@mickep for the verapdf validators I ignore all output except (in the xml output log) the line <validationReport jobEndStatus="normal" profileName="PDF/UA-2 + Tagged PDF validation profile" statement="PDF file is compliant with Validation Profile requirements." isCompliant="true"> I just grep for validationReport.*isCompliant="true" and ignore anything else it says

@micke specifically I validate for UA-2 by checking I get "true" on every line from

# validate all files to PDF/UA-2 (bible last as it takes a while)

for i in  [^b]*.pdf b*.pdf
do
echo ""
echo "======================"
echo $i
echo "----------------------"
/c/users/davidc/verapdf/verapdf.bat  -f ua2 $i 2>&1 | grep -E "errorMessage|validationReport " | sort | uniq
echo ""
echo "======================"
echo ""
done

@DavidCarlisle As long as it validates, that makes sense, indeed.

@mickep if it doesn't validate, my usual plan is to wait for @UlrikeFischer to tell me what to do next

@DavidCarlisle That is a good plan. I send an email to Hans, and obviously ruins his day(s)... :)

The warnings about that included figure are particularly annoying, you can convert it to a jpg and then it all passes with no warnings but clearly there is no sense other than technical adherence to the rules in which including a line graph and text legend as a jpg image is more accessable than including it is scalable graphics and real text.

@DavidCarlisle Indeed, so it it is borken.

@samcarter Also the cyrillic-modern thread on the texlive list is a bit fun. It was in fact discussed on BachoTeX that cyrillic was missing from the gyre fonts (and latin modern), so the upload could be some kind of reaction to that.

@mickep I heard rumours about an internet outage at BachoTeX - this explains the delay between last update of the font and the upload, doesn't it? :P

@samcarter The maintainer was not there. The internet connection did indeed go down. One could use the mobile a bit, but not stable.

@mickep a fun theory destroyed so fast :)

@samcarter Sorry. :)

@DavidCarlisle if I get an opportunity I will ask Peter.

In l3keys you can add keys to groups. Is it possible to check whether a specific keys is in a specific group? I am currently reading the implementation, but it might take a while for me to understand exactly how things are coded ... =)

Or, it would also be fine for me to be able to check whether keys weren't used by \keys_set_groups:nnn ...

It seems that \l__keys_unused_clist contains the unused keys after \keys_set_groups:nnn, but it is an internal token var, so ... I should probably not use it in my package, right?

@JasperHabicht Why do you need to do that?

Where can I find the documentation of the macro \multicolumn? I am reading the documentation of the package multirow and it mentions multicolumn but there's no section that only focuses on describing multicolumn, so I suppose that its documentation is somewhere else.

@rdrg109 Core LaTeX documentation: Lamport's book, etc.

@JosephWright LaTeX: A Document Preparation System?

@rdrg109 \multicolumn is a standard latex command so any latex tutorial eg learnlatex.org/en/lesson-08#merging-cells

@rdrg109 That's the one

Ok, thanks!

@JasperHabicht The unused list is there for various things, but as it's internal you shouldn't be using it - sounds like you want to set known keys rather than keys by group (which is really meant for 'timing' - set one group, then set another group, needed for example for fontspec)

@JasperHabicht It would be possible for us to provide a version of the group-setting code that leaves the unset keys available - but no-one has ever asked before

@JosephWright Not quite. The setting is: All keys can be set globally but also locally (as optional argument to a comman). However, local setting only makes sense for a group of keys for one command, while other keys only make sense to be set for another command.

It does not hurt to set the other keys, but I would like to issue a warning to the user then.

@JasperHabicht I'd simply define them in two subtrees, then for the global setting do a set_known for each subtree

I thought about using subkeys ... but then global setting would require the user to set the keys using these subkeys

@JasperHabicht \keys_set_known:nn { foo / subtree-1 } { <keys> } \keys_set_known:nn { foo / subtree-2 } { <keys> }

@JasperHabicht Like I said, adding a version of the group code that returns the unset keys is easy - I'm doing a release tonight or tomorrow, I could add it in there (it does make some sense for consistency with other key functions)

@JosephWright Let me think about the subkey solution. It might be an idea for the package. I found an inconciseness in the package anyways as some strings that are currently set via the same logic are already replaced when parsing and others when typesetting ... I should fiddle apart this anyways

Then, I could indeed just use two subtrees

But you are right, I could still use subtrees and hide them from the user ... that's true

@JasperHabicht I might add the feature anyway - it's quite trivial to do and would go with consistency-of-language

But first, I need to cook and eat

[Question] Where can I find the documentation on the parameters that are allowed in the tabular's argument? Namely, l, r, c, p, m and b. I found the documentation on l, r and c in LaTeX: A Document Preparation System in section 3.3.3 Arrays, but I couldn't find the documentation for p, m and b.

@rdrg109 array package?

I can see that p, m and b are also documented in learnlatex.org/en/lesson-08#merging-cells, but I am more interested in an extensive documentation more than an introductory explanation.

@JasperHabicht Thanks!

@rdrg109 Well, at least for m and b. p should be documented elsewhere as it is defined in basic LaTeX

@rdrg109 Array package docs

@JosephWright With this approach, if I only use \keys_set_known:nn, then there won't be a warning if unknown keys where used, right? So, it would still be good to have access to the unused keys ...

@JasperHabicht \keys_set_known:nn already has a 'return value' version

@rdrg109 p is standard latex, m and b are from the array package

@JosephWright 👍

\NewDocumentCommand { \JSONParseSet } { m } {
  \keys_set_known:nnN { jsonparse / global } {#1} \l__jsonparse_unused_keys_clist
  \keys_set_known:noN { jsonparse / parse } { \l__jsonparse_unused_keys_clist } \l__jsonparse_unused_keys_clist
  \keys_set_known:noN { jsonparse / typeset } { \l__jsonparse_unused_keys_clist } \l__jsonparse_unused_keys_clist
  \__jsonparse_warning_unused_keys:
}

Something like this then ...

@JasperHabicht Yes

Cool

@JasperHabicht I will though look at adding the same idea for groups

Actually, I could just replace the last line with \keys_set:no { jsonparse / typeset } { \l__jsonparse_unused_keys_clist } ...

But this way, I can issue just a warning

silently ignoring unknown keys

ooh: darwintypeface.com

@mickep ^^ ooh

@PauloCereda real Darwin font: vrt.be/vrtnws/en/2017/02/12/…