« first day (2747 days earlier)      last day (2172 days later) » 

12:41 AM
@HenriMenke Thanks! There was some confusion which is in my opinion related. I think what happens is that circuittikz redefines the to syntax (only) if some keywords are used. Among other things, paths can get broken into smaller pieces, which is good for arrows but not so good for closed cycles.
 
 
3 hours later…
4:02 AM
@StefanKottwitz @PauloCereda Please update texdoc.net or take it down. It is not useful to distribute outdated documentation.
 
 
2 hours later…
5:39 AM
@HenriMenke I wrote the script.
 
6:03 AM
* wolfie ponders how many debianites it takes to screw in a lightbulb
<Viiru> wolfie: Somewhere around 600? One screw's the bulb, and the rest
        flame him for doing it wrong.
<part> wolfie: is the bulb free software?
<Tv> Can we vote on whether to screw it or not?
@UlrikeFischer there is a lot of German tourists around here, but I am still shy to try to say a sentence to them. :)
 
6:44 AM
@DavidCarlisle Hmm
@DavidCarlisle Guess I'll take a look at the raw log: the problem with dropping the dates is loosing useful lines ...
 
@PauloCereda You could start with "Ich bin eine Ente". But if anyone then says "Oh Abendessen" better run.
 
@JosephWright while you are tinkering in the web, you could make all logged filenames start on a new line:-)
 
@DavidCarlisle That would be a very good change!
 
@DavidCarlisle I still don't understand how to rearrange a dozen lines ...
 
7:04 AM
@JosephWright do the original line ends go as part of filename normalisation? (I haven't had time to look yet:-)
@JosephWright one possibility is to declare it a feature and we just check in the tlg and move on.
@JosephWright oh, you meant web lines not lines of log!
 
@DavidCarlisle Exactly
 
7:27 AM
@PauloCereda or „Guten Tag, ich komme aus Brasilien. Bitte lassen Sie uns bei dieser Fußball-Weltmeisterschaft gewinnen!“. ;-)
 
8:00 AM
@UlrikeFischer Is there a football world championship this year? (For the ones who don't know I'm Italian: I'm ironic)
 
@CarLaTeX There are rumors about a low level competition far away in Russia. So low level that the best teams didn't get an invitation.
 
@DavidCarlisle I think I know why I can't get stuff to build locally: looking at the .travis.yml file, a load of stuff has to be touch'd first
 
Jan 26 at 15:09, by egreg
@PauloCereda We could invade Switzerland, so they can react and invade us, then we'll surrender.
 
8:17 AM
@DavidCarlisle I wonder if a blog post on primitives, building TL, the Travis-CI set up, etc. might be useful
 
@egreg Good explanation ;)
 
@DavidCarlisle Looking at the raw logs, something is not quite right ... do you think the date information is actually worth having? (Or rather, are the lines containing dates useful?)
@DavidCarlisle Now I see the date lines appearing, I'm wondering if I got this right ... as it is, there is something amiss (data being duplicated), so I'll need to adjust again :(
 
@JosephWright well some are, some less so, but in the raw logs the (./foo.fd lines are on separate lines from the dates in ProvidesFile, but they got merged...
 
@DavidCarlisle OK, so I'll fix the formatting rather than strip the lines back out
@DavidCarlisle Probably this is due to some other 'tidy up' code that has never been down this path before: I can look for File at the start of lines and preserve the space, for a start
@DavidCarlisle Thinking I could do with a dedicated development PC now I'm compiling TL sources ;)
 
8:33 AM
@JosephWright just ran into a package which broke in tl18 as it used \global with xparse commands \global\DeclareDocumentCommand ;-(.
 
@UlrikeFischer Oops: should I search TL for the culprit?
 
@JosephWright I already made a bug report: github.com/antoinelejay/currency/issues/2
 
@UlrikeFischer Ah, cool
@UlrikeFischer Can never have worked: there is loads of non-expandable stuff before the definition kicks in
 
@JosephWright yes that's what I wrote in the report. If he wants a global definition he will have to define the commands differently (but I don't see why it should be necessary. If users define the currencies inside a group they get what they deserve ...).
 
@JosephWright so is it compiling now?
 
9:33 AM
@DavidCarlisle I think so
 
9:43 AM
============================================================================
Testsuite summary for TeX Live Scripts 2019/dev
============================================================================
# TOTAL: 1
# PASS:  1
# SKIP:  0
# XFAIL: 0
# FAIL:  0
# XPASS: 0
# ERROR: 0
============================================================================
@DavidCarlisle ^^^
 
10:21 AM
@UlrikeFischer ooh David Von Karlisle
 
@PauloCereda David von Karleyl looks more german. "von Karlisel" looks ok too (but sounds different).
 
@UlrikeFischer He would have a fake accent nevertheless. :D
 
@PauloCereda He would use google translate and the accent isn't so bad: translate.google.de/#en/de/Dinner
 
@JosephWright so I'm out of a job then:-) (--no-clean is a useful option to Build if you want to speed things up, and cd Work/texk/web2c; make even more so
 
@UlrikeFischer ooh
 
10:27 AM
@PauloCereda Mein Hovercraft ist voll von Aalen
 
@DavidCarlisle Ah, right: I assumed it would only clean up if you wasked
@DavidCarlisle So Build once then avoid it again?
 
@JosephWright yes but sometimes (depending what filedates I just pulled from you:-) it gets out of sync (or I don't trust it) so I redo a full build just in case.
 
@DavidCarlisle Key point is though it should be faster than Travis-CI ... sounds good, particularly for these really tricky changes (which I still don't follow: strange errors I think from C)
 
@JosephWright certainly
 
@DavidCarlisle Cool
 
10:56 AM
@DavidCarlisle :)
 
 
2 hours later…
1:01 PM
+    \exp_after:wN #1
+    \exp_after:wN #2
+    \exp_after:wN #3
@JosephWright we should consider adding \expanded ^^^^
 
 
2 hours later…
3:24 PM
Is there a way to get the length of a piece of text that is not the usual size?
 
3:36 PM
How to correct this point (30:1cm) in addplot axis? it seems the angle is clockwise using this approach axis cs: 30:1cm yields an error.
 
@Ronen \settowidth\mylength{\large hello} ?
 
4:00 PM
@DavidCarlisle :)
@DavidCarlisle That one is done: just a question of getting the final patches into the engines. We could send a .diff to Karl for pdfTeX, put in a proper pull request for (u)pTeX and ...
@DavidCarlisle ... XeTeX?
 
@JosephWright send a PR to the texlive-source upstream repo see if they accept it:-)
 
4:15 PM
I have a datatool question: how do I do DTLforeach to select just a subset of the database using a string comparison? There are examples of selection based on numerical tests, but not string tests.
 
@AlanMunn there is no difference. You can use whatever test work, I e.g. have a \DTLforeach which uses expl3-tests in the body.
 
@UlrikeFischer Ok, I was trying to do it in the optional argument of \DTLforeach.
 
Hm, I'm also having an issue where \par seems to throw off the distance for \textwidth, letting it exist outside the margins..
 
@Ronen as always it's best to ask a question on the question site with a complete example (I could not start to guess what you mean by the above comment:-)
@AlanMunn ^^^ :-)
 
@DavidCarlisle Oi
 
4:22 PM
@AlanMunn should work too for not to complicated comparisions. There is also e.g. \DTLisSubString. But on the whole I try to avoid another set of comparision commands and stick to expl3 if possible.
 
@UlrikeFischer Well that requires having learned expl3... On my todo list. But not today. :)
 
@Ronen oh unless you mean you have <blank line> \parbox{\textwidth}{... in which case the line will be overfull as the box will be indented by \parindent unless it is in some environment such as center which does not indent.
 
Ah no, I'm trying to create underlines on a title page. I just wrapped \noindent into the function/method that creates the signature field with typed name under it
 
@Ronen underlining is so horrible, do you have to? But again I can not guess what you mean, make a small example document and post a question.
 
@UlrikeFischer \DTLisSubString worked perfectly for what I'm doing. Thanks.
 
4:27 PM
It's the best way that I can think of creating variable text-on-an-underline while staying vertically symmetric with a blank underline for a signature
Rather, having the underline stay vertically identical relative to the text on its line. If I truly can't figure it out, I'll put a post up
 
@DavidCarlisle Well yes, that was my thinking too
@DavidCarlisle Will ultimately be a .diff to go into the SVN, but the same idea applies
 
@JosephWright I haven't looked in detail about the tieup, are they just running a one direction svn to git, not the reverse?
 
@DavidCarlisle Yes: it's git-svn as a one-way mirror, run by Norbert I think
@DavidCarlisle I'm largely waiting until we have everything to go in one 'shot'
 
4:51 PM
this question didn't have much of a mwe, but i guessed correctly and my comment solved the problem. however, the problem (an unwise % in a .bib file that bibtex can't handle) is something that might bite someone else. if there's a duplicate, i'd vote for closing, but i haven't found one. does anyone know of a good duplicate, or should i answer?
 
5:03 PM
@barbarabeeton this answer here also mentions that you shouldn't use % inside an entry: tex.stackexchange.com/questions/261261/…
 
@UlrikeFischer -- good find. thanks. i think i will edit that answer to say that a % is passed on directly to (la)tex and can be catastrophic there. (the % was already added as an edit, per the comments.)
 
@DavidCarlisle The interface is already there. All it takes is reading the manuals of unicode-math and fontspec. I know, I know. 😀 — egreg 1 min ago
@DavidCarlisle ^^^^^^^
 
@egreg I still think that using \mathrm (and so a text font) for operators is wrong. I'm with @DavidCarlisle here that \symup should be used.
 
5:26 PM
@egreg do they have manuals?
@JosephWright yes, how is the xetex-tidy plan going?
anyone around who can try l3build ctan in tools ? (@JosephWright)
 
@DavidCarlisle Rumors say so
 
@egreg I think this looks right
0
Q: Problem with package "tools" after Update

Gotti91I have a problem with the package tools after an update of MiKTeX 2.9.6670 64 bit. See the following MWE in which I use the package auto-pst-pdf. While compiling, I'm getting this window: After clicking Install, the end of the log-file looks like this: MWE (TeXnicCenter LaTeX->PDF): \do...

 
@UlrikeFischer An eff operator should have the ligature.
 
Ah, finally I can submit something that isn't full of random unnecessary text.
Thank you for all your help David, by the way.
 
@UlrikeFischer is there anything a miktex user can do quickly to stop miktex complaining about a missing .tex
 
5:41 PM
The number of answers of yours I've seen is basically innumerable lol
 
@DavidCarlisle create the missing .tex
@DavidCarlisle oh you mean the question above? I will look,
 
@UlrikeFischer is that enough? can they write it to the right place? could you post an answer:-)
 
@DavidCarlisle Just checked and I don't find the .tex file on TL 2018. It's there in TL 2017
 
@egreg it was probably there two days ago, I updated tools to fix the array issue, and used the texlive 2018 l3build (to match travis) and it seems to have dropped the .tex from the zip file (it installs it locally)
 
@egreg -- i agree with you there. (@UlrikeFischer -- i think this argument is the clincher.)
 
5:50 PM
@egreg @barbarabeeton then the math font should provide the means to build such a ligature. And if the math doesn't allow this, then ok one can as a work-around escape to a similar looking text font. But the default should be to use the math font for math.
 
@UlrikeFischer -- the math roman font (\symup) is intended to be used for single-letter variable names, as are all letter symbols in plane 1. (at least that was the intent when i fought for it with the unicode technical committee.)
 
6:09 PM
@barbarabeeton well using the chars from normal latin plane would be okay too. But it would be imho a flaw if something so well known as a math function like \sin or \lim could not be typeset correctly with a math font alone. Using a text font feels like a work-around and like the logical thing to do.
 
7:05 PM
@JosephWright @egreg @UlrikeFischer I just ran l3build ctan (after hiding all the test files) using the public tl2018 l3build an .tex gets put in the zip. I guess I'll resubmit to ctan.
 
@DavidCarlisle, I don't understand what recourse I have. \MakeUppercase{} is removing all formatting from the commands put in to it, and functions horrifically. Is there a more robust \begin{} version?
For example, if I change \uppercase{ to \MakeUppercase{ then the centering is lost, and the simple text contained inside it is no longer uppercased.
 
@Ronen you shouldn't be apply \MakeUppercase to \centering just apply it to the text. `\uppercase simply isn't suitable, you can use it in simple cases but not if anyone will need an accented letter in their name for example.
 
Can it not accept a large block of text, across multiple paragraphs?
And if I remove \uppercase{} from \newcommand{\prevDegree}[3]{...} then the #2 command isn't uppercased either
 
@Ronen it could, with care but it would be very odd thing to do, and in your class you are applying it separately to each field
 
Only out of necessity
 
7:16 PM
@Ronen no, that would be the normal thing to do, and it should be \MakeUppercase{#2} not \uppercase{#2}
 
I understand that I need to include \MakeUppercase{#2} to uppercase that portion, but I don't understand why \MakeUppercase{\prevDegree{a}{b}{c}} won't uppercase the arguments
Or rather, won't uppercase the output of \prevDegree{...}
 
@Ronen If you need this you should use a font which has uppercase letters at the places of the lowercase letters. Some open type fonts have an option to do this, see e.g. tex.stackexchange.com/questions/370799/…
 
@Ronen I'm confused why do you want to uppercase the output of that command, all it does is uppercase #2 and pass #1 and #3 unchanged, so if you need to uppercase all three arguments what is for?
 
I happened upon an article that discusses `\expandafter` which I think will solve the issue, as it's a latex compiling order problem, I think.

In the title, "Bachelor of Science" needs to be uppercase, but elsewhere in the text it might be regular case. I would assume that "B.S." would be typed in by the user as an uppercase piece of text, but I realize it might not be wise to rely on the end user for that. #3 is a year, so there's no need to consider case.
So the user would input the name of their degree in the .tex, and the .cls would parse it and present it correctly, without the user learning to read what's in a .cls
 
 
1 hour later…
8:25 PM
Thanks for the help, @DavidCarlisle
 
 
1 hour later…
9:45 PM
@JosephWright I guess I used an l3build without this Include dot files in tree() (fixes #30) for tools :(
 
@DavidCarlisle Cool
@DavidCarlisle Should be in the latest release to CTAN: isn't it working?
 
@JosephWright it worked just now but clearly the version of of tools I sent last week was missing a file (I may have picked up a locally installed l3build, can't reconstruct now:-)
 
@DavidCarlisle Yes: I'd forgotten to say about that and hadn't released it, so it likely got overlooked
 
What could be the reason that unicode-math gives active apostroph a definition? Where would it need it?
\documentclass{article}
\usepackage{unicode-math}
\begin{document}
\catcode`\'=13 \show'
\end{document}
 
@JosephWright I must have seen the checkin log at the time as it came to mind as soon as I saw the question here. Oh well. (I also just deleted all the test files so it passed, an and then sanity checked by diffing against the last release)
@UlrikeFischer ' always has such a definition, even in plain tex
 
9:54 PM
@DavidCarlisle and for what purpose? Is there some place where ' are made active?
 
@UlrikeFischer yes ' mathcode "8000 and its active definiion expands to ^{\prime}
@UlrikeFischer if you ignore all the underscores and l3 stuff the definition is the same as this classic plain tex:
\catcode`\'=13 \show'

\bye
 
@DavidCarlisle Ah I see. I had forgotten that with mathcode "8000 this definition is used. Well this means that it is a rather bad idea of polyglossia to use ' as a shorthand.
 
@UlrikeFischer yes (well it could keep the math definition and do something else in text I suppose)
 
10:14 PM
@UlrikeFischer oh just seen mail on gh issue
 
@DavidCarlisle well there is an \ifmode in the code. But unicode-math is used and when latin is the main language and one use $f'$ one gets a loop. And when latin is an other language one gets error with ' in text.
 
@UlrikeFischer I suppose I could look:-)
 
10:27 PM
@DavidCarlisle Well I'm always in two minds if one should try to correct polyglossia. Perhaps it would be better to look if latin.ldf of babel contains something problematic for xelatex and lualatex.
 
@HenriMenke Ok, I took texdoc.net down.
 
10:46 PM
@UlrikeFischer Quite strangely, the shorthand has a sensible definition only if Latin is the main language.
 
11:11 PM
@StefanKottwitz Updating would have been better but it is not up to me to tell you how to spend your free time. Anyway, thanks for dealing with it.
 
11:48 PM
@marmot The leftright version wasn't too difficult. ;-)
 

« first day (2747 days earlier)      last day (2172 days later) »