« first day (4860 days earlier)      last day (54 days later) » 

3:52 AM
Fellas, its been a while
Does using \setmathfont{STIX Two Math} have the newest version of the font loaded in with overleaf?
I have texlive version 2023, and the last update to the font seems to be from 2021
4 hours later…
8:16 AM
Another day, another expl3 release ...
Will be a full set today - pretty rare nowadays
8:54 AM
@FelixMaldonado I don't think there has been any significant change to Stix Two for years
@DavidCarlisle thinking about testament etc: which structure should get these names? I mean e.g. a chapter has a Sect structure around the whole text, and a structure around the chapter heading.
@UlrikeFischer I suppose same is true of any document with sections, whether to tag the section headings as sections or to tag the whole nested section structure (for which html5 added <section> to go with the classic <h2> tagging which only marked headings). the xml geek in me says to tag the nested structure.
@UlrikeFischer we are probably tagging both but having nice names on the nested structure seems better to me
9:10 AM
Some of the talks might also be interesting for non-physicists: indico.cern.ch/event/1334738/overview
@DavidCarlisle öh. Which one in the nested structure? The current Sect and Part?
@UlrikeFischer yes I think so, although I don't really know what the "normal" thing is for tagged documents
9:31 AM
@DavidCarlisle I doubt very much that there is a "normal" thing here and think we can do what we think looks nice.
@UlrikeFischer use 🦆 and 🧸 as tag names
9:55 AM
user image
@DavidCarlisle ^^^
10:06 AM
@UlrikeFischer good for restaurant menus
2 hours later…
11:50 AM
@UlrikeFischer ooh
@DavidCarlisle oh no
@PauloCereda aparently you are a shy retiring person with no internet presence who is hard to find.
@DavidCarlisle ooh
@DavidCarlisle ?
@PauloCereda did you see the arara question?
@UlrikeFischer "I tried to find a way to contact Paulo Cereda (the author of 'arara'), but could not find that so quickly on the Internet."
@DavidCarlisle Hard, but much harder: I just tried to find an address to report an issue for colortbl and could not find it in the doc ;-) Should I add this in the dpc github or in the tagging-project? tex.stackexchange.com/a/710042/2388
11:57 AM
@DavidCarlisle hm no, link?
@DavidCarlisle danke!
@UlrikeFischer clearly the bug is having colortbl in a repo to which you don't have push access so you can't fix the errors for which you mst be responsible.
@DavidCarlisle "I read for the first time about 'upmendex' in the book 'The Latex Companion - Third Edition" ;-)
@UlrikeFischer colortbl I think
@UlrikeFischer I might move it to its own repo rather than dpctex
12:14 PM
@UlrikeFischer actually I'm confused, array.sty doesn't have the socket tagging code, you patch that in from latex-lab? but I could/should add it to colortb directly?
@UlrikeFischer ah yes it's in latex-dev hmm ok
@DavidCarlisle yes, for this is the format test. In latex colortbl will simply error, but in latex-dev there is a chance to get it working (but naturally there can be more traps ...).
> \tbl_update_multicolumn_cell_data:n=\protected\long macro:
#1->\int_compare:nNnTF \g__tbl_col_int =0{\UseTaggingSocket {tbl/row/begin}\int
_gset:Nn \g__tbl_col_int {1}}{\int_gadd:Nn \g__tbl_col_int {\g__tbl_span_tl }}\
tl_gset:Nn \g__tbl_span_tl {#1}.
@DavidCarlisle ^^^?
@JosephWright exactly
@DavidCarlisle I'm obviously being dense: can't see the issue
@UlrikeFischer format test probably not completely safe (user may have loaded one of the dated old array) but colortbl could simply add the assignments to the current definition
@JosephWright not there, here:
   \tbl_update_multicolumn_cell_data:n {#1}
   \def\@addamp{\if@firstamp \@firstampfalse \else
                \@preamerr 5\fi}%
   \@arstrut \@preamble
which has % on all but one line, so looked wrong, but an \ExplSyntaxOn is actually in scope from some lines above
@JosephWright which ought to have been obvious as there as an L3 csname but I missed that....
12:34 PM
@DavidCarlisle hm you could also test the package date, but I leave the finer details to you...
@DavidCarlisle Ah
@UlrikeFischer or this?



1:01 PM
@DavidCarlisle looks good.
@DavidCarlisle I tried to tag the documentation it seems to work fine with the exception of one error at the \rainbowline example.
ah an naturally longtable breaks with pdflatex, so I used lualatex.
2:08 PM
@UlrikeFischer pushed to github
@UlrikeFischer shame about the following commit but I guess it had to be done.
3 hours later…
@JosephWright I blame @barbarabeeton for not speaking English
@DavidCarlisle :)
@JosephWright I guess I'll do same for color.sty before the next release
@DavidCarlisle Sounds reasonable (as in: it's the sensible position even if it's extremely annoying)
@DavidCarlisle Keeping the same commit message?
@JosephWright :-)
@JosephWright I guess I'll drop in on pdfua
5:05 PM
@DavidCarlisle It on now?
@JosephWright yes stating 5 past
@DavidCarlisle Link (Skype)?
@JosephWright done
5:47 PM
@JosephWright bah Lua script :)
@PauloCereda who has a jvm available these days?
@DavidCarlisle indeed :)
OpenJDK 64-Bit Server VM (build 11.0.21+9-post-Ubuntu-0ubuntu122.04, mixed mode, sharing)
@PauloCereda my windows box ^^
@DavidCarlisle ooh wsl
$ java --version
java 21.0.1 2023-10-17 LTS
Java(TM) SE Runtime Environment (build 21.0.1+12-LTS-29)
Java HotSpot(TM) 64-Bit Server VM (build 21.0.1+12-LTS-29, mixed mode, sharing)
@PauloCereda cygwin ^^ ( well windows which causes you problems)
5:54 PM
@DavidCarlisle oh no, cygwin
6:11 PM
@JosephWright not fair, you have to give me time to take a look before merging if you ask for my review! (github.com/latex3/latex3/pull/1464#issuecomment-1953066194)
I have a scanned print-text with formulas.

Is there any recommended OCR-software?
(“Optical Character Recognition”)

For example OCR to LaTeX...
6:31 PM
@Skillmon Wasn't sure if you'd had time - I wanted something out today in case I need another round before the freeze
6:45 PM
@cis mathpix
`Gestalt besitzen: $n=2^\mu p_1{ }^{m_1} p_2 p^{m_2} \cdots p_k{ }^{m_k}$. Dabei bedeuten $\mu$ eine nicht negative ganze Zahl, die $m_i$ sind entweder 0 oder 1 , und die $p_i$ stellen $k$ verschiedene Primzablen der Form $2^2+1$ dar. (Eine ganz andere Sache ist es freilich, herauszufinden, wie die Konstruktion auszuführen ist.)
Machen wir ein paar Proben: Bei $\mu=0, k=1$ ergeben sich für $\nu=1: n=5$ und bei $\nu=2: n=\overline{17}$. Da 17 eine Primzahl ist, kann man das 17 -Eck mit Zirkel und Lineal konstruieren. Bei $v=3$ und $v=4$ erhält man die Primzahlen $n=257$ bzw.`
@cis ^
@cis it's commercial software but scarily good, and the free version (which is all I have) is fully functional just limits the number of scans you can do a month (I forget the details, I have never hit the limit)
@JosephWright Will look now, maybe I spot something :P
7:11 PM
@Skillmon :)
7:42 PM
@JosephWright looks good.
@Skillmon Phew!
@JosephWright I spotted two judgement call optimisations, both not really worth the hassle.
@Skillmon Optimisations can go in after the TL freeze :)
8:04 PM
@JosephWright oh my
@mickep I can see the reasoning
@JosephWright I'm not too surprised. But also do not really see the point.
@mickep Probably a personal project that was easy enough to send to CTAN
@JosephWright Yes, that makes sense.
And maybe it helps someone with a comparison.
@mickep Author is experienced with CTAN anyway ...
8:19 PM
@JosephWright I was about to write high and low, but maybe not fair.
More like some big.
8:36 PM
@DavidCarlisle Wow..... Ok, I will try.
@cis the above I just clipped on snip screenshot and highlighted your image and that latex appeared in my clipboard to post back here
Yes, I see. And this works.
I will try to translate a complete book:
93 pages, 4 MB. But as pdf maybe ~100 kB.
9:01 PM


        \draw_transform_scale:n { 0.5 }

        \draw_path_rectangle_corners:nn { 1cm , 1cm } { -1cm , -1cm }
        \draw_path_rectangle_corners:nn { -1cm , -1cm } { 1cm , 1cm }
        \draw_path_use_clear:n { stroke }

^^^Bug? The tro rectangles don't overlap
@JasperHabicht \listfiles?
ah, fixed in current version
@JosephWright Yes, it is fixed. Overleaf is not up to date ...
I really need to always double check ...
@JasperHabicht :)
9:55 PM
@JosephWright The first one only is actually faster if the string representation of the keys is below 25 tokens, or else it becomes slower, the second I'm not even sure whether it's really faster, I'd have to benchmark it a lot.

« first day (4860 days earlier)      last day (54 days later) »