« first day (4779 days earlier)      last day (88 days later) » 

cfr
1:01 AM
@AlanMunn I was particularly thinking of something like two languages both written in Cyrillic. I can't tell any of them. Sometimes I can't even distinguish Welsh and English. I guess I'm actually thinking mostly about individual words.
@AlanMunn Meaning one flag or two?
 
@JasperHabicht -- Distinguishing between Slavic languages rendered in Cyrillic by spelling differences is very much like distinguishing between Latin-alphabet languages, and I suspect this is also true for Arabic-family languages. But quite a few non-family-related languages use the Cyrillic (e.g. some languages of the "stans") or Arabic (e.g. Farsi) scripts. Two scripts that I find particularly intriguing are Cherokee and Inuktitut.
 
 
2 hours later…
2:42 AM
@cfr Two flags on the same pole.
 
cfr
3:11 AM
@AlanMunn I hope the dragon was on top ;).
@barbarabeeton Some of the Celtic languages are mutually intelligible (or almost so) when spoken, but not when written.
I guess that would be a common pattern.
 
3:33 AM
@cfr Sadly no. But Americans have very peculiar views about their flag, and one of the "rules" is that it has to fly higher than any other flag. (That's likely a common rule, in many countries, but I don't think most people elsewhere would know or care.) So I suspect whoever was flying them didn't want to offend people. When I was 5 I apparently horrified people by dragging an American flag on the ground behind me in NYC. :D
 
4:11 AM
@AlanMunn -- I'm not sure what the situation is with respect to flags of other nations, but there is an official "flag code" in the U.S. The American flag is always to be flown higher, and preferably to the right of, any state or local (U.S.) flag. It isn't to be allowed to touch the ground, and when it's lowered from a flagpole, it's to be folded first lengthwise in fourths, then into triangles, ending up with the stars on the outside. It shouldn't be used as, or applied to, a garment.
 
@barbarabeeton Yes, that's what I'm referring to. But in most other countries nobody really cares about this stuff in my experience.
 
cfr
@AlanMunn I know ;). That's why I said it.
@barbarabeeton Good grief. I only knew people were ... er ... strangely enthusiastic.
 
(cont'd.) When a flag is worn out, it should be destroyed by incineration (so I've always wondered why flag-burning is illegal). There are more rules. Should be easy to find if you look for "U.S. flag code". I grew up in Maryland, which has a quite magnificent heraldic banner. It was a game for my Girl Scout troop to check how it was flown at state parks; more than once, we found that it was upside down. I grew up during WWII, and all this did mean something then.
@cfr -- Times change. I've been imprinted, I fear.
 
cfr
@AlanMunn England's flag has strong associations with racism. The Welsh flag is rather more subversive. The current Westminster government wanted to make some huge Union Jack display on some new building in Cardiff, but they gave it up. They also wanted it on all the vials of covid vaccine, but abandoned that, too.
@barbarabeeton I'm not a fan of the Union. I would never display the Union Jack. It means something to me, but the meaning is mostly negative. The Welsh flag is different. It means something positive to me but not like that the US flag seems to to Americans. I mean it's a symbol. It's not the thing itself.
@barbarabeeton I don't think it is a WWII thing either. I'm not saying it's not partly that, but there's just a vast difference in the value Americans put in their flag which doesn't really translate, regardless of generation. There are also differences between England and the other nations, though, which complicates things. There was very, very little bunting around here during the coronation, for example. There wasn't the same attempt to silence protesters, either.
The Union Jack is very heavily associated for me with monarchy and rule from Westminster. For me, it also symbolises loss of identity due to Brexit.
[Wales voted for Brexit, but we're currently least keen on the monarchy.]
[The 'Prince of Wales' thing is a particularly sore point ....]
 
 
2 hours later…
6:45 AM
@cfr Just remember that this strange Westminster family is of German decent, they changed their name due to some historic events (WWI).
 
cfr
7:14 AM
@Skillmon Indeed. I'm not sure it greatly helps in this particular case. But in general, yes, we support Germany when they play England. (I'm pretty sure my exchange partner thought we were being polite.)
 
 
2 hours later…
9:27 AM
@cfr Interesting - as someone born in England, I'd strongly favour the Union Flag over the English one (but then I guess I do think the union is a sensible idea: in geographic terms, having multiple nation states on the island of Great Britain feels like an indulgence to me. I am though in favour of a much more federal structure.)
 
9:47 AM
@cfr (And yes, I recognise that this feels very different for people with strong Welsh/Scottish identities)
 
10:04 AM
Isn't it nice if code uses intuitive variable names like \...ht for width and \...wd for height? :P
  \wd\beamer@areabox=\beamer@animht
  \ht\beamer@areabox=\beamer@animwd
This makes the code so much easier to read :)
 
@barbarabeeton awwww
 
@samcarter The toolset I need to work with during dayjob has such a bug, for one of the GUI element types width and height are switched, but only during get, not for set... Also reading the colour of a GUI element gives the colour as (signed) integer in 0xRRGGBB, but with the sign-bit flipped. So this small flip in a package doesn't seem too bad!
 
@Skillmon that sounds like fun :) Does the test case use a square gui element?
 
@samcarter matters get worse: no unit testing possible...
(not sure whether the developers of that toolsetting had any test suite for the backend stuff -- we're stuck on an old version for production, and we won't update to the new version because we want to switch completely away from that provider)
 
10:19 AM
@Skillmon Oh, no wonder they need a programming wizard killer rabbit for the job :P
4
 
@samcarter though my error rate is low on edits, I'm not perfect @DavidCarlisle, and sometimes introduce new bugs... Testing is fun.
 
You still need to learn that the new word for "bug" is "feature"
 
@samcarter maybe I can use that phrasing on the next validation round.
 
10:54 AM
@samcarter that's why we have feature requests
 
11:45 AM
@JosephWright is that error expected? (I came across that coding in @MarcelKrüger luamml)
\ExplSyntaxOn
\debug_on:n { check-declarations}
\ExplSyntaxOff
\documentclass{article}
\begin{document}
\ExplSyntaxOn
xx \hbox_set:Nn \c_zero_int {abc}
\ExplSyntaxOff

\end{document}
! LaTeX Error: Inconsistent local/global assignment
 
@UlrikeFischer I think so: you've not used a proper box assignment - we don't support 'raw' TeX box management in the expl3 functions
 
@JosephWright ok.
 
12:47 PM
@DavidCarlisle I just rewrote an old TikZ answer of mine with pict2e. I'm blaming you for this :-)
2
 
@campa that is not how you spell "thanking"
 
@DavidCarlisle blaming, thanking... po-tey-yo, po-tah-to
 
@campa ask @UlrikeFischer about the difference
@JosephWright although the natural response to the error isn't to change the assignment but change \c_zero_int to 0
 
@DavidCarlisle is there a difference??
 
2:06 PM
@DavidCarlisle Yes fair but it's simply not supported behaviour to use an arbitrary number here
 
@JosephWright in purely L3 code you can hold that line but it's not so easy if using l3 code that's got to interact with existing 2e packages. i suppose a better idiom would be to define \l_scratch_zero_box to be zero and use that rather than c_zero_int
 
@DavidCarlisle We already have \l_tmpa_box
 
@JosephWright yes but can you be sure that's 0 (if adding some l3 based hook that's got to interact with \setbox\z@...
 
@DavidCarlisle You can be sure it's not!
 
@JosephWright I know:-)
 
2:17 PM
@DavidCarlisle This comes up with checking active if you use \<a-letter>_ as the first arg.: really not likely to be an issue in non-expl3 code
 
@JosephWright Marcels code is inserted inside latex2e code which uses 0 (\z@) already, so it had to be either \setbox\z@ or \hbox_set:Nn 0 or something like that. (on the whole \setbox\z@ it probably the best option).
 
@UlrikeFischer \setbox\z@ I'd say - best to be realistic about where to and not to use expl3 constructs
 
yo'
@DavidCarlisle interesting -- Thanks! I think you're rigth that that we're running it through PS, but I'll double check.
Actually we just run dvipdf with $dvipdf = "dvipdf -dNOSAFER -dALLOWPSTRANSPARENCY %O %S %D";
 
2:47 PM
@yo' you do? ouch, do you use a modified graphics.cfg do that graphicx, hyperref etc all default to dvipdfm \special not dvips?
 
yo'
@DavidCarlisle You ask me? I have no idea :D But I don't think so :)
Actually, the example at overleaf.com/latex/examples/psfrag-example/tggxhgzwrzhn doesn't work, na? UPDATE: It does, I'm just stupid
 
@yo' no you are OK (and the comment in the template document is wrong) dvipdf is the script that comes with ghostscript that just does dvips, gs. It is not dvipdfm dvi driver.
$ type dvipdf dvipdfm
dvipdf is /usr/bin/dvipdf
dvipdfm is /usr/local/texlive/2023/bin/x86_64-cygwin/dvipdfm
@yo' ^^^
 
3:06 PM
Mamma mia, here we sock again? tex.stackexchange.com/q/702840/82917
 
@campa It is interesting how they manage to get the expl3-naming always a bit wrong.
 
@campa I suspect so but haven't flagged as such yet, it's (just) possibly someone else
@UlrikeFischer and types
 
3:33 PM
Olé
 
3:50 PM
Does TeX LaTeX StackExchange have a new downvote-policy? I don't understand why Incrementing value for unspecified optional argument got 8 downvotes? Is somebody undermining the donvote-policy of TeX LaTeX StackExchange?
 
@UlrichDiez It's a question from the user who is circumventing a question ban - they are not popular as a result
 
@JosephWright What user? What question ban? Even if this question might be not popular, I don't see the question itself being spam or rude or abusive on purpose. So I think the reasoning for closing and not even showing is probably not precise.
 
@UlrichDiez when you saw the tenth version of the same question you will not longer think that
 
@UlrichDiez There is a user who was question-banned: they post lots of very similar questions without proper context, etc. I can't say much about the circumvention but it's clear that they are not taking the ban in the way it's supposed to be taken: an opportunity to learn what is expected. So the votes are because the pattern is clear - this is abuse of the system
 
@UlrichDiez it's the 100th account for the same (suspended) user he burns through several accounts per day
 
3:59 PM
@UlrikeFischer @JosephWright @DavidCarlisle Now I see. Thanks for the clarification. So it seems somebody is deliberately flooding TeX LaTeX StackExchange with duplicates of the same question, not noticing answers already provided?
 
@UlrichDiez well not getting answers as the account is suspended and the post deleted, so simply asking it again under a new account, changing a few bits here and there to try to look different
 
@UlrichDiez The questions are all similar, don't really have context, etc. Some get removed by mods, but I can understand people downvoting heavily
 
@JosephWright "Questions" of similar structure - did somebody train a chatbot on annoying the TeX LaTeX StackExchange community?
 
@UlrichDiez I guess this is a possibility.
 
@UlrichDiez No, I'm sure it's a real person - it's just they are not listening to the feedback
 
4:12 PM
@JosephWright The reason why I suspected that somebody might wish to undermine the downvote-policy is that I am a bit paranoid: In the past I noticed discussion about intervention by someone from the company with moderator privileges which seemed not in line with the friendly policy of TeX LaTeX StackExchange. I was fearing that there would continue to be disagreements along these lines, and I'm glad that that doesn't appear to be the case.
 
@UlrichDiez Sure, I get it: we are dealing here with an outlier user, and there needs to be a non-standard approach as a result
 
@UlrichDiez no, chatbots don't usually switch to profanities once they get exasperated
 
@DavidCarlisle I did not see the other posts - only the one which caused me to ask. However, I realize I did not show the best diplomatic skills in the way I asked. Please accept my apologies. I'm not up to speed at the moment - I can't get more than three hours of sleep without the tinnitus waking me up...
 
@UlrichDiez no apologies necessary, as posts are deleted, unless you see them go past it's not likely you would know and seeing posts hit -6 or so within minutes of posting from a "new user" is highly unusual for this site.
 
4:34 PM
@DavidCarlisle @JosephWright Thank you for your understanding.
 
5:14 PM
@UlrichDiez I don't know who downvotes, it isn't me. But I don't see any reason to answer the questions either. This is now going on for months and there is no progress visible or any effort to improve or to learn.
 
@UlrikeFischer A downvote is added automatically also when people flag the post.
 
 
1 hour later…
6:24 PM
@UlrikeFischer I changed it to \setbox in luamml. There are some other changes needed in luamml, so I'll try to give it some love over the weekend.
 
@MarcelKrüger I made a fork in my account and changed there already small bits (including a missing catcode setting, I suddenly had 10R in the pdf ;-)). And started to add some examples and document more. If you have some time to explain stuff to me that would be nice. I'm still wondering how to influence the mathml result, e.g. how to do \sin correctly.
 
6:45 PM
Great. If you want I can also give you permissions in my repo. In `algorithm.tex` I tried to start documenting a bit how the internal table representations work, once you know that you can use `\luamml_annotate:nen` to replace how a specific part gets converted. E.g. for `sin` you want an `<mo>sin</mo>` (which gets represented as `{[0] = "mo", "sin"}`) to be used for the nucleus (since you don't want to overwrite e.g. superscripts), so you would set something like
```
\luamml_annotate:nen{1}{
nucleus = true,
 
7:09 PM
@MarcelKrüger that doesn't work. E.g. this errors with ! Missing $ inserted.: (I'm testing directly with all testphase code by purpose to check for clashes):
\DocumentMetadata{uncompress,pdfversion=2.0,testphase={phase-III,math,table},xmp=false}
\documentclass{article}
\usepackage{unicode-math}
\usepackage{luamml}
\begin{document}
\ExplSyntaxOn
\luamml_annotate:nen{1}{
 nucleus = true,
 core = {[0] = 'mo', 'sin'}
 }{\operatorname{sin}}
\luamml_set_filename:n {test-mathml-2.txt}
\ExplSyntaxOff
$\sin x$

\end{document}
 
You need that in the place where it's inserted, so
\DocumentMetadata{uncompress,pdfversion=2.0,testphase={phase-III,math,table},xmp=false}
\documentclass{article}
\usepackage{unicode-math}
\usepackage{luamml}
\begin{document}
\ExplSyntaxOn
\cs_set_protected:Npn \sin {
  \luamml_annotate:nen{1}{
   nucleus = true,
   core = {[0] = 'mo', 'sin'}
  }{\operatorname{sin}}
}
\luamml_set_filename:n {test-mathml-2.txt}
\ExplSyntaxOff
$\sin x$

\end{document}
 
@MarcelKrüger ah ;-)
@MarcelKrüger and the third argument is what you want to typeset?
 
Yes, the third argument is what should be typeset, the second argument is the Lua table used to overwrite and the first argument is the number of nodes in the third argument. (There is also a version which omits the first argument and calculates that automatically, but that is much slower if you are using pdfTeX instead of LuaTeX)
 
@UlrikeFischer I probably need to play this game too:-)
 
7:32 PM
@MarcelKrüger is there a possibility of not having to do that but simply inferring it from the run of character tokens in a \mathop in the final mlist?
 
@DavidCarlisle well I'm actually looking how to combine this. luamml provides a mathml representation which one can use as AF, that part is quite easy. The main problem is how to fine-tune the mathml it creates (e.g. \sqrt produced msqrt and not mroot.)
 
@UlrikeFischer yes sure which is why in my experiments I actually quite liked having the mathml in an external \jobname-mathml.xml as (a) I could schema validate that it was correct mathml and (b) hand tune it as needed (I write better mathml than tex4ht:-) so for the first run I used the generated mathml but after that, I could tweak....
 
@DavidCarlisle it can write the mathml into a file (or many files). That is actually what happens in the example above. But if you add \luamml_flag_structelem: it also adds mathml-structures and if we want to use that too it is important that one can fine-tune.
 
@UlrikeFischer yes that's what I mean if the system reads mathml from a file if it is there and generates mathml if not (preferably with some checksum sanity check) then it's easy to customise the mathml just by editing the generated file
 
@DavidCarlisle That's the plan and why I mentioned that for \sin this is only an example how to do overwrites and not how this should actually be implemented.
 
7:49 PM
@DavidCarlisle but I don't think one can use a mathml from a file to generate structure elements. That would only be useful if one adds it as AF. Or can one @MarcelKrüger?
 
@UlrikeFischer No, I don't see a nice way for that either. It would require a lot of work to basically map back.
 
@UlrikeFischer iI just \input all the mathml in one go at the start of the run so each fragment is just a def'ed macro you can put it back into the pdf any way you want
@UlrikeFischer oh sorry, I see what you mean, yes OK
 
@DavidCarlisle Of course we could go full circle and also support MathML input in LuaLaTeX, then it would be much easier to also support associate structure elements. Then we could generate MathML in the first run and in later runs use the MathML as source. Of course that would take a lot of work.
 
Hi folks. I suddenly started getting a warning and an error for a simple file. I could swear I haven't changed anything TeX related here for a while.
Package xcolor Warning: Package option `usenames' is obsolete and ignored on in
put line 266.

! LaTeX Error: Unknown option 'sections' for package xcolor.
I did a search. Didn't come up with anything.
 
@FaheemMitha well we updated xcolor. It now warns that usenames is obsolete since 2006, and it errors if you use unknown options (like sections).
 
7:54 PM
I commented out xcolor from the preamble, which I don't use anyway, and it works. But I probably have something similar in other places, which do use xcolor.
@UlrikeFischer Oh. So before it just ignored unknown options, and now it gives an error?
 
@FaheemMitha yes
 
@FaheemMitha there is no problem with xcolor itself. But you should clean up the options you use with this package.
 
@UlrikeFischer Yes, I understand.
@DavidCarlisle OK. This must have been quite recent. It was working a few days ago.
 
@FaheemMitha yes the update was last week or so.
 
@FaheemMitha every other package errors, xcolor did not more or less by an accident, but someone who shall remain nameless so we don't blame her fixed that
 
7:55 PM
@UlrikeFischer OK. Thank you.
@DavidCarlisle I see. OK, shall fix.
 
@UlrikeFischer see, I am not mean and did not blame you?
 
@DavidCarlisle yes ;-) and yes ;-(
 
8:18 PM
0
Q: LaTeX Error: Mismatched LaTeX support files detected

jo87casiI am continuing to get an error about Mismatched Latex Support files in Texmaker. In particular it says: (LaTeX) Loading 'l3backend-pdftex.def' aborted! (LaTeX) (LaTeX) The L3 programming layer in the LaTeX format (LaTeX) is dated 2023-02-22, but in your TeX tree the files require (LaTeX) at leas...

we could add something like this before giving up, so people know where to look
{
\makeatletter
\everyeof{\noexpand}
\newlinechar=32
\typeout{%
  \@@input{"|kpsewhich --all expl3.sty"} %
  \@@input{"|kpsewhich --engine=pdftex --all pdflatex.fmt"} %
}}
@JosephWright @UlrikeFischer ^
 
@DavidCarlisle sounds like a good idea. Even if the user don't understand it, it would make it easier for us.
 
 
1 hour later…
9:42 PM
@DavidCarlisle Sounds good to me
@DavidCarlisle Make a PR?
 
9:53 PM
@JosephWright yes OK later
 
 
1 hour later…
11:02 PM
oh I see you do most of that already `\begin{texttt}
% kpsewhich --all --engine=\meta{engine} \meta{format}[-dev].fmt` so just a matter of adjusting things a bit to oreturn more info
 

« first day (4779 days earlier)      last day (88 days later) »