« first day (3504 days earlier)      last day (33 days later) » 

7:33 AM
@DavidCarlisle as Petra complained about the missing reference to the author and licence in the readme.md of zref, what should we put there (also for pdftexcmds)?
 
@UlrikeFischer I don't know, is it worth maintaining the fiction of the ho-tex group or should we say Heiko up to whenever then you me and him from then on? Probably simplest to keep it as ho-tex group and simply copy this form from the top of the dtx but lose the %
% File: pdftexcmds.dtx
% Version: 2019/11/24 v0.31
% Info: Utility functions of pdfTeX for LuaTeX
%
% Copyright (C)
% 2007, 2009-2011  Heiko Oberdiek
% 2016-2019        Oberdiek Package Support Group
%
 
@DavidCarlisle Sounds fair
 
@UlrikeFischer then we can make sure the l3build auto tag function includes the readme to keep the date in sync
 
@DavidCarlisle, @UlrikeFischer Odd that \pdf@elapsedtime thing: in l3luatex the change Ive made specifically mentions Lua 5.3. I must have forgotten to push it back tothe older code
 
@DavidCarlisle I just corrected otehr in the readme ;-).
 
7:40 AM
@JosephWright haven't had time to look yet, are there no other functions impacted by that change
@UlrikeFischer how did that get there? Who could we blame?
 
@DavidCarlisle It's the only one I changed in l3luatex, so I guess that's true
 
Should I say "Copyright" or "Authors"?
 
@UlrikeFischer copyright I guess, ctan don't care who wrote it, just who has the right to upload changes in future
 
@DavidCarlisle well it is for the readme that is shown to users.
 
@UlrikeFischer I really don't mind, I'd just use the same text as the dtx, easier to check that they say the same thing then:-)
 
7:50 AM
@DavidCarlisle I pushed a change. If it is okay I can upload.
 
@UlrikeFischer yes sure, thanks
 
8:01 AM
@DavidCarlisle uploaded, but the auto tag code needs to be adapted.
 
@UlrikeFischer yes OK will check later, I guess readmes across the whole collection will want similar updates at some point as well.
 
@DavidCarlisle yes we should then do this at every update. (the next are ltxcmds and pdfcrop ...).
 
8:24 AM
@egreg just stole an 8 year old tick from me :(
 
@manooooh basically you measure the size of the picture each step and then prepend a path with the option use as bounding box at the start of the picture. There will be probably a way to do this automatically in a couple of passes using aux files, and several gurus here that can code it in seconds, but even manually is just quite fast).
 
quack
 
8:39 AM
@PauloCereda I refer you to my previous replies.
 
8:50 AM
@DavidCarlisle oh
 
 
1 hour later…
10:15 AM
@yo' Have you guys considered having a look at similar popup data from other projects? For example texmaker and textstudio? One ought to extract information from their archinve of cwd (or what they are called) files.
I've often wondered for example why other editors haven't copied several of the features found in Auctex for Emacs.
 
@daleif like having a lisp interpreter:-)
 
yo'
@daleif and texstudio provides any such statistical data publicly?
Also, I don't really thing that going purely by statistics is the right way. (Note that this way, \linebreak got into the list, but we all know that you shouldn't really need this command. And similarly with couple more things.
 
10:35 AM
Hi guys, is this a problem of the mirror or it's me?
TL20[romano:~/texlive2020] 36s 1 % tlmgr update --all --self

/home/romano/texlive2020/bin/x86_64-linux/tlmgr: TLPDB::from_file could not initialize from: ctan.javinator9889.com/systems/texlive/tlnet/tlpkg/…
/home/romano/texlive2020/bin/x86_64-linux/tlmgr: Maybe the repository setting should be changed.
/home/romano/texlive2020/bin/x86_64-linux/tlmgr: More info: tug.org/texlive/acquire.html
 
@Rmano You could try another mirror...
@Rmano try this
 
@PauloCereda ¿Can I change mirror on the fly? I didn't know...
 
tlmgr update --self --all --repository mirror.ox.ac.uk/sites/ctan.org/systems/texlive/tlnet
 
@PauloCereda Yep, that worked. Thanks!
@PauloCereda /happy squirrel sounds/ (which I still don't know how they're called in english)
 
@PauloCereda You remind me I want to talk about that with the team for Travis-CI
 
10:43 AM
@Rmano ooh quack
@JosephWright we ducks are very good at reminding other ducks :)
 
11:36 AM
@yo' not statistics, but popups of the various relevant functions provided by various packages.
@DavidCarlisle Coulnd't these ideas be implemented without Lisp?
 
yo'
@daleif ah yeah; implementing package-based suggestions would be a tremendous task.
We provide autocomplete for macros that were already used in the file, so this should do the trick quite.
 
@yo' that's why working together on it might be a good idea. Perhaps send some funding their way (like google does with various projects)
 
yo'
@daleif well, the thing is, there are utterly conceptual complications. For one, it's not entirely easy to know which packages are actually loaded.
 
@yo' log file parsing
It could be added to the list of "nice to have"
 
yo'
11:52 AM
@daleif yeah, but that itself has its other issues.
@daleif yeah. I'll write a ticket about it. However, I can't really promise anything...
 
 
1 hour later…
yo'
1:03 PM
@daleif Another thing is frankly, is it the best thing that you provide all commands from all packages the user (or even some of the packages) loads?
 
1:30 PM
@daleif yes of course, ignore me, I just meant that every editor should have a lisp interpreter built in, not intended as an entirely useful comment,
 
@DavidCarlisle -- Could you take a look at this: Why do arguments to nested \tilde or \breve commands reappear when amsmath is used? Has that problem been fixed yet in amsmath? (My answer got a late upvote, and while checking that out I noticed that I'd said I'd try to get a public list of outstanding bugs, but that's definitely still absent.)
 
@DavidCarlisle Well, lisp might be interesting, but see what happend with xindy. Additionally didn't auctex start like 30 years ago, so a bit of a head start in comparison to other latex code handlers.
 
@barbarabeeton no still there I think
@UlrikeFischer actually the function works on the README.md as you edited it, but you need tagfiles={"README.md", "*.dtx"}
 
@DavidCarlisle -- Okay. Thanks. What about a "public" list of open problems? I'm willing to help compile it, modulo being able to access the information relatively easily.
 
1:45 PM
@JosephWright @DavidCarlisle did you see the question about siunitx and color group?
 
@UlrikeFischer no
 
1
Q: xcolor changes size of numbers of the siunitx package

zlm4When using siunitx in combination with xcolor the size of the numbers defined by \num changes for some instances. See following MWE. The second 1 in the denominator gets smaller if xcolor is used. This only happens with KOMA-Script document classes, not with the standard document classes. The ord...

 
@UlrikeFischer ah new just now
 
@UlrikeFischer Yes
 
@UlrikeFischer it's rather strange
 
1:47 PM
@daleif Whatsits, I think
 
@JosephWright ooh
 
@JosephWright no, simply the additional grouping.
 
@UlrikeFischer Ah, right, yes
 
@JosephWright whatsit, how does that change the font?
 
@barbarabeeton the answer got upvote(s) because this morning there was a question marked as a duplicate (correctly) tex.stackexchange.com/questions/547790/…
 
2:00 PM
@DavidCarlisle -- Okay. That's rather terse. I think that some of the suggestions from the AMS list aren't there, but since I don't have a copy available, ... Is this list mentioned in the CTAN collection? (Would be useful.)
@DavidCarlisle -- A question closed as a dup today is good test material for (i think) #127.
 
2:19 PM
@barbarabeeton no as it's all in a private repository, but we can mine it as issues are raised at github
 
@DavidCarlisle -- I assume that's a reference to the AMS list. What about mentioning the "open problems" list in the CTAN collection? Is that there already?
 
2:43 PM
@barbarabeeton sorry you lost me (I'm following an online seminar at same time:-)
@barbarabeeton the ctan page lists the github issues page if that is what you mean, amsldoc references latex-project.org/bugs which will get you to the same place
 
@DavidCarlisle -- Thanks -- that answers the question(s). Enjoy your seminar.
 
@barbarabeeton C++ it's the day job not for fun:-)
 
@DavidCarlisle Sounds exciting
 
@DavidCarlisle -- Oh, I understood that, and was being snide.
 
3:07 PM
@UlrikeFischer Argh, what a stupid error! I'll fix this asap. Thanks for pointing out!
@UlrikeFischer I got a feature request for duck ski. I vaguely remember that you once showed a ski duck picture? Do you still happen to have this?
 
@samcarter_is_at_topanswers.xyz have you not seen @JosephWright ? ducks don't need skis they can slide on their feet
 
@DavidCarlisle At TUG15 he walked instead of sliding :(
 
user image
6
@samcarter_is_at_topanswers.xyz ^^^ sent ...
 
@UlrikeFischer Thanks a lot!
 
 
2 hours later…
5:04 PM
@DavidCarlisle ooh
 
5:15 PM
@UlrikeFischer I should have the PDF stuff sorted soon
 
@JosephWright you mean l3build?
 
@UlrikeFischer Yes
@UlrikeFischer I'm also thinking about color
 
@JosephWright I was confused at first - (almost) everything is PDF-stuff currently ;-)
 
@UlrikeFischer :)
 
5:35 PM
@UlrikeFischer Try the latest checkin
 
5:45 PM
@JosephWright I got a result folder but it contains the stripped down pdf's with e.g. the fonts as [BINARY STREAM]. (I copied the lua-files to the script folder and then did simply run one "l3build check test")
 
@UlrikeFischer Which PDF specifically? I'm copying before normalisation
 
@JosephWright I run a test (annot) which uses luatex, pdftex and xetex. And I get three files annot.luatex.pdf, annot.xetex.pdf and annot.pdftex.pdf in the build/result folder and they are all three identical to their normalized counterpart in the build/test folder.
 
@UlrikeFischer OK, I'll think some more: I have a bit of an issue in that I can't run the XeTeX tests (my anti-virus does not like it)
@UlrikeFischer I'm sure I'm going in the right direction ;)
 
@JosephWright I'm sure of it. In which file is the code that tries to copy the pdf?
 
@UlrikeFischer l3build-check
@UlrikeFischer L796
@UlrikeFischer Ah, I have an idea
 
6:02 PM
@JosephWright you are copying the wrong file, and pdfext contains the period ...
 
@UlrikeFischer No, that's not it
 
@JosephWright with ` cp(name .. pdfext,testdir,resultdir)` I get (with one engine) the right pdf in the result folder, but how can one rename it so that the next engine doesn't overwrite?
 
@UlrikeFischer I've worked out the issue, give me a couple of minutes
@UlrikeFischer Should be fixed
 
6:28 PM
@JosephWright Now it it works fine I have three different pdf in the folder.
 
quack
 
@JosephWright I assume that all test configs will use the same result folder? So one should be careful to give the tests unique names? (not a problem if I know it).
 
 
1 hour later…
7:39 PM
@JosephWright people still use antivirus? :)
 
@UlrikeFischer @DavidCarlisle A nice line from math.cls (besides the silly name): \newfont{\abdullah}{cmr9 scaled \magstep0}
Not to mention {\bf Abstract:$\mbox{\!}$}
@yo' ^^^^^
 
7:56 PM
@egreg eek:-)
 
8:13 PM
@egreg eek ;-). I found lately this comment in a style:
\RequirePackage[svgnames]{xcolor}%
  \@ifundefined{T}{%
	    \newcommand{\T}[2]{{\fontencoding{T1}%
	      \selectfont#2}}}{}
%    \end{macrocode}
% There seems to be a bug in the T1 encoding of some package
% (unidentified, but possibly \textsf{xcolor}) which
% uses the command {\ttfamily{}\textbackslash{}T1}, which is an
% impossibility (no digits allowed in command names). So we fake
% it here to stop \LaTeX{} complaining, by dropping the first
% argument on the floor.
 
@UlrikeFischer nice!
 
8:40 PM
@UlrikeFischer I'll fix that too
 
@JosephWright I already found a solution. I did set the resultdir variable in the config-files (but if were automatic it would be fine too).
 
@UlrikeFischer I'll automate: probably tomorrow then do a release
 
@JosephWright: Forgot that I have supower powers with tikz-pgf questions. This question got closed when I just intended to vote for it to be closed: tex.stackexchange.com/q/547869/4301
Can you reopen it?
Thanks. That was quick.
 
 
2 hours later…
10:24 PM
Hello everyone,

If I have some multi-lined content saved inside the command \TCBXheader. Then I have

\newsavebox{\headerbox} % https://tex.stackexchange.com/a/360875/2288
\savebox{\headerbox}%
{%
\TCBXheader
}

how can I retrieve the actual height of the box?
 
\ht\headerbox or \dimexpr\ht\headerbox+\dp\headerbox\relax depending what you mean by the height.
 
I actually need to pass that height calculation to the geometry package option headheight.

So, when trying your second option, I found to that the header overflows the top page border
while
headheight = \dimexpr\ht\headerbox+2\dp\headerbox\relax
allocates more space than needed
 
@Diaa why 2 ?
 
@DavidCarlisle just experimenting
 
@Diaa if your content is in a parbox then by default it's centred so more or less half the total height is in the depth so ht+2dp is 50% too big
@Diaa if you use a \parbox[b] as in the linked answer then the depth of the box is the depth of the last row which is usually the depth of a \strut so around 30% of \baselineskip so using 2dp makes it about 4pt too big at 10pt font size.
 
10:35 PM
@DavidCarlisle The linked answer is just for my reference. My actual problem is here tex.stackexchange.com/q/547802/2288
Or I would be grateful if you could help me make the following work

\documentclass{exam}
\usepackage{graphicx}
\usepackage[xparse,breakable,raster]{tcolorbox}

\newcommand{\TCBXheader}{%
\begin{tcbraster}[
raster columns=3, raster rows=1, valign=center,
raster height = 2cm, raster column skip = 0mm
]
\begin{tcolorbox}[left=0pt,right=0pt]
First line\\ Second Line
\end{tcolorbox}
\begin{tcolorbox}[halign=center]
Centered Text
\end{tcolorbox}
\begin{tcolorbox}[halign=right]
\includegraphics[height=2cm]{example-image-a}
noting that I don't need to use a box inside the header since exam needs many runs to fetch the number of questions/marks, and I find that using boxes inside the headers always get me ?? instead of the numbers, which implies that it doesn't store/retrieve info out of the aux file as I guess.
 
I don't now the exam class and don't really know tcolorbox, but I ran your example and get ! Missing number, treated as zero. is that what you expect?
 
@Diaa well the headheight is easy: headheight = 2cm. There is no need to calculate anything.
 
@UlrikeFischer My original document has several rows with variable height
 
@Diaa then why raster height = 2cm ?
 
In this answer to my question https://tex.stackexchange.com/a/547859/2288 , it is said that tcbraster should be inside a minipage to avoid the error of missing begin document.

So, what I expect is a header whose height is calculated and passed to headheight
 
10:48 PM
@Diaa \begin{minipage}\TCBXheader\end{minipage} generates an error as you are missing the width argument but it's not clear why you are using a minipage anyway.
 
@UlrikeFischer I just tried something with my MWE
 
@Diaa no I mean do you get that error (which is simply a syntax error for minipage unrelated to any of your questions about heights of boxes)
@Diaa as Ulrike says the height of this construct is the height of the image and you know in advance that is 3cm
 
@DavidCarlisle I think you are absolutely right. I totally overlooked that error of missing argument for minipage.
 
@Diaa if you get any error at all then don't even look at the pdf output (which is meaningless) how can you overlook an error?
 
@DavidCarlisle but in the main document I don't know the actual height of several rows with variable heights
@DavidCarlisle kind of panic attack
 
10:53 PM
@Diaa you can measure the height and depth as stated. but don't you want it to be less than the height of the image so the boxes look right? I really don't know what this code is trying to do:-)
 
I found that the original document has no syntax errors with the same strange output of overflowing the top border
@DavidCarlisle I am sorry for the inconvenience. I will try to make a more solid MWE :)
 
@Diaa setting text in the preamble really isn't supported fonts are only fully set up at \begin{document} you can sometimes get away with measuring small bits of text in the main document font but if you push it too far it will break. Also of course using boxes in the header is not common as they are fixed but header text usually depends on the page content
 
@DavidCarlisle I think this is the main issue. I don't have dynamic allocation of text vertical space.
 
@Diaa the page head height needs to be fixed, you can calculate it per-document but not per page
 
especially, when the header has text dependent on the main document variables.
@DavidCarlisle I didn't think about this. You just saved an OCD patient hours of working on it :D
 
11:19 PM
I have another question; is it ok to conditionally load a package option like this?

\usepackage[
left = \iftrue 10mm \else 20mm \fi
]{geometry}
does it have any fragility?
it doesn't work with me, so, I wonder if I am doing it wrong or another approach should be used
 
@Diaa er why? iftrue is always true so it's just a slow way of saying 10mm
 
@DavidCarlisle Just for having a MWE :)
I think \PassOptionsToPackage can be used, but I would like to know if another one is available.
 
@Diaa well that's the same as saying 10mm so its Ok whether your real code is OK depends on what it is.
@Diaa it's rather weird though as the actual condition is evaluated in code you do not control at the point the package decides to use the key value. so if your condition is true or false at different points you don't know when it is evaluated so what value it will have
 
I am writing an exam, so when it is a booklet I need to have a left=10mm, otherwise, I need it left=30mm

However, what I wrote gets a distorted output, while

\ifbooklet
\PassOptionsToPackage{left = 10mm}{geometry}
\else
\PassOptionsToPackage{left = 30mm}{geometry}
\fi

works well
 
11:35 PM
@Diaa well there you are evaluating the conditional in your own code so you have more control but in that case it makes no difference as booklet is presumably a constant always true or always false in any given document
 

« first day (3504 days earlier)      last day (33 days later) »