« first day (2431 days earlier)      last day (2800 days later) » 

06:51
@jrh equation and displaymath do different things (equation is for numbered equations and displaymath, which is usually written as \[ is for un-numbered equations)
07:35
@DavidCarlisle I see decodearray should be 'restored'
08:23
@JosephWright didn't you just do that?
@JosephWright did you get this? (from Travis ci) "Emails are so yesterday! (except this one, of course)" .... @CarLaTeX was telling me that the other day:-)
08:48
@DavidCarlisle ?
@CarLaTeX email (lists) :-)
@DavidCarlisle does another one find them not clear? :)
09:05
@CarLaTeX I got an automated message from the travis-ci continuous integration thing this morning saying being notified by email of build failures is "so yesterday" and I could configure notification by campfire or flowdoc or hipchat or IRC (yeh one I recognise:-)
09:32
@DavidCarlisle there is really a notification system call "campfire" ;-). I thought you made this up ;-). Does it use smoke?
@UlrikeFischer I have no idea, I assume so!
@egreg did you see I got a gold badge for mentioning you in an answer?
09:56
The camptown ladies sing this song, doo dah, doo dah
@PauloCereda thank you for your incisive and important contribution to the debate.
@DavidCarlisle <3
@DavidCarlisle at least I didn't open a PR for that. :)
@PauloCereda if I did, would you know what to do with a hipchat notification of it?
@DavidCarlisle Nope. :)
@PauloCereda as we can tell from your choice of song, you're just not hip enough
10:04
@DavidCarlisle oh
@DavidCarlisle enjoy this: it.babbel.com/it/magazine/…
@CarLaTeX Italian ninjas? Wait a minute...
@PauloCereda LOL
10:20
@DavidCarlisle Citing authoritative sources is a key to success.
@DavidCarlisle Oh, well… ;-)
@egreg apparently
@PauloCereda carrying pineapple pizza
@DavidCarlisle LOL
10:44
@DavidCarlisle Yes, I mean 'It's now in TeX Live'
10:55
If you wanted to hinder successful compilation when there are undefined references, how would you do it?
@Skillmon change
\def\@setref#1#2#3{%
  \ifx#1\relax
   \protect\G@refundefinedtrue
   \nfss@text{\reset@font\bfseries ??}%
   \@latex@warning{Reference `#3' on page \thepage \space
             undefined}%
  \else
   \expandafter#2#1\null
  \fi}
to use \@latex@error{...}\@eha instead of \@latex@warning{...}
@Skillmon but you need to do that after the initial runs or forward references won't work
% arara: pdflatex
% arara: bibtex
% arara: pdflatex
% arara: pdflatex
% arara: halt if found('log', 'undefined reference') &&
% arara: --> showMessage(1, 'Oh no!', 'There's a problem, yo!')
@DavidCarlisle ^^
$ arara
Error: Unable to access jarfile /usr/local/texlive/2017/texmf-dist/scripts/arara/arara.jar
@DavidCarlisle but this way compilation would be completely aborted if I use -halt-on-error . @PauloCereda what is arara (never heard of that)? And parsing the log with a scripr is not desirable -- I want a TeX only solution, I'm sorry :)
@PauloCereda ^^^
11:03
@Skillmon Sorry
@DavidCarlisle you are mean :)
@Skillmon isn't that what you asked for? (what else did you mean by "hinder") ?
@DavidCarlisle I need Nicola's .jar trick, it seems. :)
@Skillmon @PauloCereda is probably crying into his pringle tube now at you not knowing arara (google for its author:-)
@DavidCarlisle I am hugging my Pringles tube, that's quite accurate. :)
@PauloCereda or just get a cygwin-specific arara shell script into texlive, hang on...
11:06
@DavidCarlisle sort of... But I only want the PDF to be unusable, not the complete compilation to abort (sorry, I was imprecise about that). If I would throw an error, the compilation would never fix the label issue.
jrh
jrh
@DavidCarlisle Oh, makes sense. Thanks!
@PauloCereda @DavidCarlisle well arara is on my machine (guess it was distributed with the LaTeX installation), but it has no --help and no manpage, so it is unusable!
@Skillmon sorry I can't guess what you want:-) by default you get a warning and ?? in the output, what do you want that is more severe than that but less than an error?
@DavidCarlisle my bad, nothing to be sorry about
@DavidCarlisle Oh no worries, I can poke Nicky. :)
11:09
@DavidCarlisle I think I can built something with the \setref macro :) thank you for that hint
$ arara
  __ _ _ __ __ _ _ __ __ _
 / _` | '__/ _` | '__/ _` |
| (_| | | | (_| | | | (_| |
 \__,_|_|  \__,_|_|  \__,_|

arara 3.0 - The cool TeX automation tool
Copyright (c) 2012, Paulo Roberto Massa Cereda
All rights reserved.

usage: arara [file [--log] [--verbose] [--timeout N] [--language L] |
             --help | --version]
 -h,--help             print the help message
 -l,--log              generate a log output
 -L,--language <arg>   set the application language
 -t,--timeout <arg>    set the execution timeout (in milliseconds)
@PauloCereda ^^
@PauloCereda vvvvv
@DavidCarlisle ooh what did you do?
#!/bin/bash
posixararapath=`kpsewhich --progname=arara --format=texmfscripts arara.jar`
jarpath=`cygpath -w $posixararapath`
java -jar "$jarpath" "$@"
@DavidCarlisle Thank you!
@DavidCarlisle: also, I wrote the pretty printing functions because of your fancy Cygwin environment. :) Everybody wants their tests to be all green. :)
@PauloCereda see if you can get arara4 out before I switch to the ms linux subsystem:-)
11:13
@DavidCarlisle ooh a race
@DavidCarlisle I could release the binary, as it's already frozen, and leave documentation behind. No one reads that, and I can harvest more points in the main site.
@DavidCarlisle: you know you are the culprit for this: github.com/cereda/arara/issues/62
@PauloCereda I prefer GNU make, I think, over arara!
@Skillmon No worries. :)
@Skillmon: David is just teasing me, as usual. :)
@PauloCereda I got that
@Skillmon David is mean
@PauloCereda he is making fun of ducks, as everybody :)
@DavidCarlisle why do you use MS the first place?
11:19
@Skillmon as always. :) We ducks are very naïve, hence a good target for bullying.
@Skillmon sort of related to who pays me
@DavidCarlisle I'm sorry to hear...
@DavidCarlisle when is the aux-file written or when are the labels written to it? If I throw an error in \AtEndDocument the labels aren't written.
@Skillmon no I could use linux and wine or remote access to windows when needed but I prefer it this way I have a windows PC but do everything in xwindows, all my terminals open are xterm I have a couple of emacs and a cygwin texlive so the fact that it's windows doesn't really interfere its convenient when you are testing windows builds to have a windows machine locally.
@DavidCarlisle emacs ewww
@Skillmon the aux file is opened for writing in begin{docuemnt} and labels are written at the point you have \label it's closed for writing then re-opened for reading in end{document}
@PauloCereda usually at least two
11:26
@DavidCarlisle Two sessions?
@DavidCarlisle If I put a \@latex@error in \AtEndDocument all the aux contains is a \relax (in a file which only uses \section{test}\label{sec:test}\ref{sec:test} in between begin and end document
@PauloCereda one the cygwin locally and one running on a real linux box somewhere in the building
@DavidCarlisle ah cool. Fedora, I suppose? :)
@DavidCarlisle you could do like Ormandy who made the Microsoft Malware Protection Engine run on Linux, just so he could use the superior tools available on Linux for fuzzing and stuff.
:)
@Skillmon if you quit after the error with x while the file is open for writing then how much actually written to the file depends on the details of your OS file buffeering.
@Skillmon that's why microsoft now ship an ubuntu linux as part of windows 10
11:29
@DavidCarlisle that makes sense :) Didn't consider that
@DavidCarlisle so no real Linux then. :)
@Skillmon all out windows products are made with tools lik egnu make etc with cross platform scripts (mac/linux/windows) that's why we use cygwin so much (the end products are native windows binaries not cygwin ones but all the build is cygwin based when on windows)
@DavidCarlisle is there any MinGW setup too?
@DavidCarlisle how hard is it to port a Linux C program (which uses GSL) to native Windows, without the use of cygwin? I successfully build against cygwin but haven't tried without it and have to be able to do so for my thesis (if possible, if not, the cygwin solution is alright I guess).
@Skillmon sorry too close to day job for chat (also not really my area:-)
11:37
@DavidCarlisle ok :) I was just curious, as I never tried that (and never programmed anything natively for Windows)
@Skillmon just stick your C in visual studio do the secret magic masonic handshake then press "build" and it'll all work first time....
@DavidCarlisle sounds tempting to try :)
@DavidCarlisle but seriously an OS that doesn't allow you to put a dot infront of the file name (if using the gui)... How good can it be?
@Skillmon hmm gui what's a gui :-)
@DavidCarlisle that's the stuff you see if you start a freshly installed Windoof until you cured it from that curse :)
@DavidCarlisle That guy from Street Fighter
11:42
@PauloCereda tell me more about him, I never played Street Fighter (Tekken ftw)
@Skillmon oh yes the bit before you start up xterm and type make? I tend to ignore that bit
@Skillmon Neither have I. :) I grew up playing MK. :)
@DavidCarlisle Windows has only one reason to exist on my machines and that's the games which have no Linux port. Is it easy to run a Xorg on Windows using cygwin?
@Skillmon I just install cygwin's X along with the initial setup so it's just a matter of checking a box to install X
$ X --version
Welcome to the XWin X Server
Vendor: The Cygwin/X Project
Release: 1.19.2.0
OS: CYGWIN_NT-10.0 dc-bantham 2.8.0(0.309/5/3) 2017-04-01 20:47 x86_64
OS: Windows 10  [Windows NT 10.0 build 15063] (Win64)
Package: version 1.19.2-1 built 2017-03-09
@Skillmon What about this: v v v
\documentclass{article}
\makeatletter
\usepackage{xcolor,eso-pic}
\def\@setref#1#2#3{%
  \ifx#1\relax
   \protect\G@refundefinedtrue
   \nfss@text{\reset@font\bfseries ??\AddToShipoutPictureFG*{\AtTextLowerLeft{\rule{\textwidth}\textheight}}}%
   \@latex@warning{Reference `#3' on page \thepage \space
             undefined}%
  \else
   \expandafter#2#1\null
  \fi}
\begin{document}

  \section{abc}\label{b}
  \ref{b} \ref{c}

  \newpage blub
\end{document}
11:56
@UlrikeFischer I kind off like that approach. I only fear that some might not recognize that before hitting a print button :)
@DavidCarlisle does it capture super? So that I can run a decent window manager in that X and don't get head ache on the simple tasks of life, like managing around 20 terminals, instances of pdf-viewers and the like?
@Skillmon It's so long since I tried running a normal X window manager I can't remember I just run it with no root window which means essentially I use teh windows gui as the windows manager but I regularly have several dozen xterm and a dozen or so xpdf and 5 or six emacs with no issues
@JosephWright Hangouts not working from Safari
@egreg Possibly not: I've grabbed the latest Google Chrome
@egreg Just popped up saying your were trying to join: perhaps give it another shot
12:20
@Skillmon I think this here nobody can ignore v v v
\documentclass{article}

\usepackage{xcolor,eso-pic}
\usepackage{etoolbox}
\makeatletter
\def\@setref#1#2#3{%
  \ifx#1\relax
   \protect\G@refundefinedtrue
   \nfss@text{\reset@font\bfseries ??}%
   \@latex@warning{Reference `#3' on page \thepage \space
             undefined}%
   \immediate\write\@auxout{\noexpand\global\noexpand\booltrue{openrefs}}%
  \else
   \expandafter#2#1\null
  \fi}

\newbool{openrefs}
\boolfalse{openrefs}

\AfterPreamble{%
 \ifcsundef{r@test-for-undefined-labels} %
 {% first compilation never succeeds
@Skillmon (but it needs one compilation step extra, to write the boolean to the aux.)
@UlrikeFischer wow, that's nearly perfect! Thank you very much for this awesome idea. The extra compilation is ok, I guess.
I must just shout out a big Oh! for the joy of drawing figures with MetaPost!
12:55
Sorry for asking stupid questions again. Is it possible to get the timestamps of a file in LaTeX without external scripts and within one run?
@Skillmon You mean \time?
@JosephWright but \time gives me the current time in LaTeX, what if I want to know the timestamp of another file? Like I have my file a.pdf and my file b.tex. Is it possible without shell escapes and with just (La)TeX to get the time the file a.pdf was written the last time?
@JosephWright or even the file b.tex was altered the last time...
@Skillmon \pdffilemoddate in pdfTeX
@JosephWright Ouch I swear I read lemonade in there.
@JosephWright thank you very much. Is there any place where one can search for such commands? Like I searched for it before asking and didn't find anything and if I search using the command name I instantly find stuff related to it :(
13:11
@Skillmon It's a pdfTeX primitive: texdoc pdftex
@JosephWright well, the problem is, what if I don't know that there is a primitive defined by pdfTeX and just search for what I need using the web search engine of my choice.
@Skillmon I would start the search on tex.sx: The first hit after searching on tex.SX for "file date" leads to the filemod package ...
13:36
@Skillmon Probably we'll add support to expl3 (tends to happen with most low-level things, and I'm working on file operations at the moment)
 
1 hour later…
14:56
@JosephWright please add it (and if you're at it, please add a compare between timestamps -- like \ifnum)! If I knew what I was doing I'd help programming this, but I never took a look at expl3 sources and don't know TeX primitives well enough, I think.
15:21
@UlrikeFischer thanks for hinting at the filemod package, it is perfect for my intentions.
16:14
@DavidCarlisle Has pointed out Unicode 10 is out
We obviously all need 🧙 from unicode.org/emoji/charts/emoji-released.html
Oh cool, a downvote :-)
@Johannes_B I've the feeling that does not happen so often on TeX.SE as on Math.SE. Do you understand why you got it?
@mickep I guess the user doesn't agree with my answer :-)
@Johannes_B Well, obviously, in some way. :)
@JosephWright I can remove my poor answer if you write another one. :-)
16:59
@Johannes_B I might write something: the problem with profiling loading is it misses almost all of the real payload
@JosephWright If i have facebook open and a youtube video playing music, i don't really care how long it takes to compile the pdf :-)
17:12
@DavidCarlisle Bug report about EPOCH!
@DavidCarlisle 'Not our fault', of course ;)
@JosephWright oh no
@JosephWright yay
17:29
@DavidCarlisle Updated to Unicode 10
@JosephWright yes just got home and saw the checkin message thanks, I updated my unicode.xml at github xml-entities as well
@JosephWright yep, replied:-)
@DavidCarlisle Looks good: just 'interesting' in the context of our parallel EPOCH discussion ('It's not any use ...' ;)
@JosephWright has UKTUG got $100 spare to sponsor 🦆 ?
@DavidCarlisle Was thinking of doing something very 'GitFlow' like and suggesting an approach to EPOCH in l3build in a feature branch
@DavidCarlisle You've seen the account summary
@DavidCarlisle We've just ordered 500 TeX Live DVDs ...
@Johannes_B I'm not sure why you sent me that link. It is in no way related to what I've asked...
@Johannes_B well it is interesting though...
@DavidCarlisle Ooh, we could sponsor a TeX one ;)
@DavidCarlisle On other matters, is my answer to tex.stackexchange.com/questions/376840/… OK?
17:50
@Skillmon I got the downvote for that answer.
18:04
@JosephWright yes although it's longer than the answer I normally use of "moving argument" = "somewhere where things go wrong" and \protect = "whatever local hack is needed to fix things in this context"
@DavidCarlisle Yes, I did wonder about saying there is a reason 'moving' is somewhat loosely defined
18:25
@JosephWright -- do you care about typos? (e.g. "needs" in para.3 should be "need".) i can edit or report if you do care.
Ok. But why did you ping me in that link posting? So I can give you an upvote to make everything alright again?
19:15
@Skillmon It could be that @Johannes_B pinged the wrong person there, that it was supposed to go to @mickep.
Oh, so that was the question. Thanks @TorbjørnT.
 
2 hours later…
21:42
@barbarabeeton Please do
@JosephWright -- i gather that means "edit". okay, will do.
@JosephWright -- done. very small really, but i'm fussy.
22:03
Hello! I am editing my lecture notes and use the amsthm package. As I missed some of the lectures, I would like to set the counter of the first theorem on the page to some number. Only the first one, e.g. 12.7. Other should follow that numbering. How could one set the counter in this way?
It has just worked with \setcounter{section}{12} and \setcounter{defi}{7}.
but, I have a different numbering for other theorem styles, as defi refers to the definition style and anme to remark. Any way to set all the counters at once to get a consequtive numbering?
22:21
@Kirill No, if they are independent.
@egreg they should be dependent: if I set nothing manually, they follow each other.
@Kirill Sorry, but without an example…
should I post a question or would you like to see it here? @egreg
@Kirill Didn't you post one?
@egreg 5 min ago, but I will add an example now
22:49
@Kirill As I said, if all theorems are independently numbered, you cannot do them all at once; besides, what number should they get?
@egreg the number can be chosen free now, say 12.7 for defi. Then follows anme with 12.8, satz with 12.9 etc.
@egreg What I mean - if I set nothing, they are consequtive. So, there is a common counter. We must change him.
@Kirill no you have declared them all with independent counters
@Kirill Well, you don't show how you define them, how can we help you?
Maybe I reduced to much for a minimal example?
I usually have a template, and all "my" theorems are consecutivly numbered. But the setting the counter breaks this consecutivity.
@egreg actually I wrote it in the question, don't I? Or, again, have I thriven something away for a minimal example, that is necessary?
@Kirill no, when you declare a new theorem you can either give it its own counter (the default) or use an optional argument to specify an existing counter. You have not used that argument so your theorem types are numbered independently
23:04
I am confused. In my template they are numbered consecutivly. @DavidCarlisle
@Kirill just try the declarations you posted, then try some \begin{defi}..\end{defi} followed by a \begin{anme} you will see the numbers are not related
@Kirill no it isn't the last two theorems are both numbered 1, one is defi one is anme
@DavidCarlisle Ok, I get it, my fault, sorry. Only the theorems of one type are consecutivly numbered.
@Kirill that's the way you have declared them, so you need to set each type, you could have declared them to use the same counter, but either way you need begin end
@DavidCarlisle I am using begin and end again :) What could I do then? Put anme in "definition" style? But it would be great if I could have defi as it is, and anme in italic.
What would be a right structure?
@Kirill the way to declare theorems to use an existing counter has just been posted, declare the first one as you had it then for the rest instead of the last optional [section] which means restart every section, use the first option [defi] which means to use the same counter as defi
23:18
So, now I am getting an error, that the defi counter is not defined, although it should be.
I have another question to this later.
I have fixed the first the problem from 1.18, but now I have 12.1 for defi, and 12.1.1 for anme, although I told anme to follow defi.
So, I have fixed everything.
It was all about the order of definitions and position of []
The second qeustion would be: what should I do in the case of e.g. 10 definitions? If we name them with A,B,...,J, should I tell B to follow A, C to follow B,... J to follow I, or should B follow A, C follow A, ..., Z follow A?
23:35
@Kirill the first optional argument doesn't mean "follow" it means "use this counter" so as I said above if your first definition (say defi) does not use that option it will define a defi counter then all your later ones use [defi] to say count in the same sequence
@Kirill texdoc amsthm is only 15 pages long and would have saved you a lot of time:-)
GustavoMezetti tells me the same. But I cannot understand how can I manage all the packages I use. Geometry, graphicx, hyperref, array and LaTeX Companion have much more than 15 pages. But I use tens of them. @DavidCarlisle
@Kirill but you are not asking about those packages you are asking about amsthm!
Will it be an error to say C to use a counter of B when B uses the counter of A? Mathematically it should be transitive, is it?
@Kirill yes of course it will be an error as there is no counter B
@DavidCarlisle oh I thought every definitions becomes a counter, and if B depends on A, then B just copies the value of the counter of A.
btw, @DavidCarlisle what role mathematics plays in TeX?
I am studying it, but asking myself: if I get my mathematics study, will i be able to improve TeX, or should one get a computer science study to do that?

« first day (2431 days earlier)      last day (2800 days later) »