« first day (5191 days earlier)      last day (21 days later) » 

00:04
@Skillmon good night!
 
8 hours later…
08:22
@JosephWright wasn't there a \tl_case:nn? It's gone now...
Ah, found my mistake, ignore that ^^^
 
1 hour later…
09:37
@DavidCarlisle @egreg @cfr mail to Michael Sharpe sent.
 
1 hour later…
cis
cis
10:50
@Skillmon I tried a website, which uses
`timeout 15 /usr/local/texlive/2019/bin/x86_64-linux/latex`


Is it possible, that your code https://tex.stackexchange.com/a/734942/46023
does not run there?
11:07
@cis the code shown there runs on texlive.net
@cis if you want a low-level optimised \ResP for maximum performance I'm afraid I'll have to ask for money, I don't feel like implementing an optimised subset of \fp_eval:n... :P
(and l3fp is already extremely well optimised, so I'm not even sure how much μs I could shave of the runtime of the current implementation)
@cis 'A website'?
@Skillmon Only by going LuaTeX-only I think
@JosephWright well, I'm sure I could find spots in which there could be micro optimisations... But otherwise, yes, Lua would be the go-to solution
cis
cis
:67012829 That doesn't surprise me, I think he'll use the latest and most up-to-date.
Unfortunately, the website I use has TeXLive 2019...
Yes:
https://matheplanet.de/matheplanet/nuke/html/latexrender/example2.php
@JosephWright for instance every \expandafter\nextstep\numexpr...; could be made faster using \expandafter\nextstep\numexpr...\relax;
@Skillmon If you do that, take all of the ; out ...
11:16
@cis ask them to update or use a newer tex system. 2019 is pretty old now, especially if trying to run new code.
cis
cis
@DavidCarlisle Yes, I asked. I hope they will update to TL2024 (25?)
@cis do you not have tex locally on your machine?
@JosephWright nope, that would make it slower, because then you'd have to use \expandafter\nextstep\expandafter{\numexpr...\relax}. The ; act as the right-delimiters of the macro.
@Skillmon LuaMetaTeX ...
@JosephWright I know, you could still do this, but with a performance degrade because of \expandafter\nextstep\numexpr...\expandafter;\numexpr...; turning into \expandafter\nextstep\numexpr...\expandafter\relax\expandafter;\numexpr...\relax; (just realised I omitted every \the in my previous messages)
cis
cis
11:28
Yes, I have TL 2024 on the PC.
However, the TeX outputs of my article on the website look more elegant when translated there (as opposed to pasting them as a png file).
@Skillmon I see the performance argument, but if we make a change here, I'd rally rather be able to handle ; in LuaMetaTeX
11:50
@Skillmon Is \numexpr expandable? I guess you want \expandafter\nextstep\the\numexpr...
12:06
@egreg obviously, see my comment in that message in ().
12:23
@Skillmon but you could presumably use (something) instead of ; (it's only an internal separator isn't it ? (@JosephWright)
@DavidCarlisle Yes - it's just it shows up like 1500 times!
@DavidCarlisle yes, but there was a quick and dirty test in the past to replace it with another separator, but that didn't work, iirc.
@JosephWright well get @Skillmon to use a reasonable editor then that's not a problem:-)
@DavidCarlisle wasn't I who tested that, so blame @UlrikeFischer
@Skillmon I would never blame her for anything
 
1 hour later…
13:54
@DavidCarlisle oh no ;-(
14:09
@UlrikeFischer but at least @Skillmon will be happy to hear that there are rumours that you have been using a good editor recently.
@DavidCarlisle she used VIM?
@DavidCarlisle I used an OS with in-built xml-validation ...
 
1 hour later…
cfr
cfr
15:21
@Rmano thanks. I would be interested to know what response you get. but \mathscr is not configured as a maths alphabet, so I'm not sure it is really a bug. at the same time, I haven't followed through what it's doing in enough detail to really understand what it's trying to do. so maybe it could be made to behave a bit more as if were a maths alphabet, even though it isn't one.
15:59
@DavidCarlisle so is set TEXINPUTS=".;dpi300;;" the correct command?
:66996975 \documentclass{article}
\usepackage{hvmath}
\begin{document}
Text $\Pr(A/B) \frac{x^{-2}}{\sum_{i=1}^{n} y - 23\pi}$.
\end{document}

this is hvmi10test.tex, the code that you've given in your answer @cfr
i would actually prefer linux, as you say :-p. it is just that my main laptop containing linux has broken down, and i am using a spare windows 11 laptop
cfr
cfr
16:31
@ApoorvPotnis my sympathies ;). you could try the texmf.cnf method David suggested.
16:59
@ApoorvPotnis probably yes (just one ; at the end though), although I don'thave a windows tex to test
 
1 hour later…
18:20
And I missed my last shot at a palindrome for a while, I got a tick and am now at 69991.
18:47
In any case, why do we need to tex file in the same directory as the other files? The hvmath will correctly load the necessary files from the correct places from the texlice distribution, right?
We don't generally need to put the font files in the working directory, so why in this case?
 
2 hours later…
20:19
@ApoorvPotnis you don't need to, it's just that all the paths include the current directory, so putting everything there should work and if it does you know the font files are not corrupted, then it's just a matter of putting the files in the right places. If it doesn't work when they are all in the same directory, the problem is elsewhere and the files are probbaly broken, not just in the wrong place
@Skillmon you just need to learn to use better downvote management
2
 
2 hours later…
cfr
cfr
21:52
@ApoorvPotnis :67014828 I think double is OK. you just don't need double at the end.
@ApoorvPotnis because otherwise you have to install them before trying them out.
@ApoorvPotnis did you get it to work in the end?
is it possible to make a github action report failure if any job fails, but not cancel remaining jobs if one fails? I can do the first or the second, but I can't figure out how to have both.
@cfr not sure I understand but in a shell if you do command || echo "oops" then the command never actually fails and you just need to collect up the log messages

« first day (5191 days earlier)      last day (21 days later) »