« first day (2578 days earlier)      last day (2348 days later) » 

12:02 AM
Stupid online manuscript submission system allows me to upload a pdf but a week later I get an email saying Word docs only. Couldn't they put that in the instructions for authors.
 
 
2 hours later…
2:03 AM
hi
 
 
3 hours later…
5:09 AM
@UlrikeFischer Ooooh :)
@StrongBad Twice stupid: because they don't check the uploads and because they accept only Word documents :):):)
 
5:38 AM
danielsank@snail:~$ tlmgr install modular
(running on Debian, switching to user mode!)
Unknown directive ...containerchecksum 06c8c1fff8b025f6f55f8629af6e41a6dd695e13bbdfe8b78b678e9cb0cfa509826355f4ece20d8a99b49bcee3c5931b8d766f0fc3dae0d6a645303d487600b0... , please fix it! at /usr/share/texlive/tlpkg/TeXLive/TLPOBJ.pm line 210, <$retfh> line 5752.
I tried to install a package and got a rather obscure error message, as shown above.
It seems this has something to do with outdated texlive.
 
@DanielSank In which case, tlmgr update --self just might save the day!
 
@HaraldHanche-Olsen nope, same error :-(
Is there somewhere I can just drop a sty file to install it?
 
@DanielSank You can always stick a copy in your personal texmf tree, I would imagine. Get it from CTAN.
 
@HaraldHanche-Olsen Yeah I wrote the package so I have the sty file...
But where do I put it?
 
@DanielSank I am not a linux user, but my guess would be ~/.texmf/tex/latex/ or a subdirectory thereof.
 
5:47 AM
hmm
~/texmf/ contains two subdirectories: tlpkg and web2c.
 
@DanielSank I think it will work no matter where you put it. kpathsea is quite permissive when looking in your personal texmf tree. But I would recommend mkdir -p ~/.texmf/tex/latex and then copy your file in there.
 
I see.
 
Subdirectories are nice to organise stuff when you have loads of files, that's all.
 
Thanks, lemme try...
Έυρηκα!
It worked!
 
@DanielSank :)
 
5:51 AM
Thanks.
 
@DanielSank Now just don't forget it's there, for when your tlmgr upgrade does work again and you wonder why you're getting an old version of the .sty file.
 
heh
That sounds like good advice, the taking of which means I'm doing it wrong...
 
@DavidCarlisle yeah, exactly. I guess the plan is to define \write to stick a penalty before the itself so it gets broken at when I duplicate the word
 
 
2 hours later…
7:50 AM
@AGoldMan I had a quick look at that last night but I don't see how you can add a penalty without potentially adding a breakpoint that is taken, specifically it would allow you to get a page break between the \write and the word it was indexing so you would only get one \write but with the wrong value
@DanielSank that means you have a tlmgr from 2015 or older but using a tl2017 repository. that doesn't work.
 
 
4 hours later…
11:57 AM
@DavidCarlisle if the penalty is 9999, why would it ever get taken normally?
also I could cheat by using an empty discretionary and setting \exhyphenpenalty to 10000 for all reflows of my box except for when i grab the word
 
@AGoldMan I don't know perhaps the normal page width is 1sp wide as well, and so that's the only possibility:-) I guess it's safe enough. Personally I'd be tempted to document it as a feature, the word appears on both pages and is indexed as such, doesn't seem entirely unreasonable to me (but then I am completely unused to the convention of doubling the word at the break so my intuition is probably wrong)
 
@DavidCarlisle it's not a reasonable feature. that's my feeling at least
@DavidCarlisle what did you think of the code, by the way? aside from the spaces
 
@AGoldMan :looked more or less sensible (didn't read every line:-)
@AGoldMan several of my packages have "features" that some have claimed unreasonable, so that in itself is not a bad thing:-)
 
12:23 PM
@DavidCarlisle luatex 1.06.2 broke synctex ;-(. Just wrote to the luatex list.
 
12:35 PM
@UlrikeFischer must admit I never use synctex (but then my tex files are rarely more then a paragraph long:-)
 
I encountered a strange problem with the x86_64-darwinlegacy binaries of TL2017. Using the tlmgr only the biber binary in x86_64-darwin got updated from version 2.7 to 2.8, but not the binary in x86_64-darwinlegacy (stayed at 2.7). This lead to a version mismatch with the newly updated biblatex package. I completely reinstalled texlive2017 and biber was now completely missing from x86_64-darwinlegacy.
Does anyone know if this expected behaviour (i.e. biber support for x86_64-darwinlegacy is discontinued) or should this reported somewhere? [My current workaround is a link to the biber binary from x86_64-darwin, which seems to work fine.]
 
1:18 PM
@egreg might want to see Wolfgangs answer to that ntheorem question.
 
@daleif Wow!
 
@daleif I left a comment:-)
 
@DavidCarlisle I saw. Though is comment was a bit harsh. I don't think ntheorem even hooks into eqnarray. I know egreg does not like it. But we use it quite a lot and it gives cleaner code than manually adding \qedhere
 
1:38 PM
@daleif I don't think ntheorem should necessarily load amsmath but an answer shouldn't tell people not to use a required package, or at least not without a good alternative in place.
 
2:07 PM
@DavidCarlisle which is exactly why I pinged
 
3:06 PM
InftyReader is quite good at OCR, even for equations: inftyproject.org/en/software.html (five pages per day are free).
I saved the pdf page as a binary image (600 dpi) using Adobe Acrobat Reader (Save As Other - Image - TIFF).
 
3:49 PM
@DavidCarlisle Interesting interaction between LuaLaTeX and bm
2
Q: Macro definition not working using LuaLaTeX and German

Christoph90Im compiling the following MWE with LuaLaTeX in an updated TexLive 2017 installation on Ubuntu: % !TeX program=lualatex \documentclass[a4paper,12pt,twoside=semi,openright]{scrbook} \usepackage{polyglossia} \setmainlanguage{german} \setotherlanguage{english} \usepackage{fontspec} \usepackage{su...

 
4:25 PM
@egreg I'm sure it's not my fault
 
4:47 PM
@DavidCarlisle Maybe ensure " has catcode 12? Take the example from Heiko. ;-)
 
5:08 PM
@egreg packages that make " active in the preamble should have spaces in their name and fail to load.
3
@egreg your shortstack texts look horrible and out of line:-)
 
@DavidCarlisle The customer's always right.
 
@egreg if @JosephWright had written an l3tabular package by now then I'm sure you would have gone on to explain how to do it properly.
 
5:23 PM
@DavidCarlisle Bruno's job ;)
@DavidCarlisle I've got siunitx v3 to write (including sorting out how we want 'well-behaved' code to look ...)
 
6:01 PM
Quack from the shaky bus
I am hungry... 😢
🍕🍕🍕🍕🍕🍕🍕
TeX people left the room...
I could prepare a surprise party
I need balloons
I ate the cake
 
yo'
6:26 PM
@PauloCereda I see the pinapple there!
 
6:59 PM
@PauloCereda No food on the shaky bus?
 
7:16 PM
@PauloCereda I've just eaten a marinara today 🍕🍕🍕
 
7:55 PM
@PauloCereda On RAIStoria they’re talking about the Spanish Inquisition
 
8:08 PM
SO um, has anyone seen the error
"ERROR - Data file 'lit13.bib' cannot be read in encoding 'UTF-8': Can't decode ill-formed UTF-8 octet sequence <E8> at C:\Users\MyName\AppData\Local\Temp\par-4d617474686577\cache-4b419f4e7b1e3e9cdb78a8c63410aec04e2636e8\inc\lib/File/Slurper.pm line 63."
When running biber before?
I've wiped the cache, and I'm using 2.2 so it isn't the old cache error
OK, so line 63 seems to deal with getting the directory name? WTF, I don't have unicode OR spaces in my directory path
 
@egreg On this question: tex.stackexchange.com/q/402398 there's a partial solution with XeTeXinterchartoks although it has downsides. Do you think it's worth posting? (Or is there a way to fix the problems?)
% !TEX TS-program = XeLaTeX
\documentclass{article}
\makeatletter
\chardef\CharBound=\e@alloc@intercharclass@top
\makeatother
\XeTeXinterchartokenstate=1 % enable character classes
\newXeTeXintercharclass\mathquote % create a new class
\XeTeXcharclass `` \mathquote
\XeTeXcharclass `’ \mathquote
\XeTeXcharclass `' \mathquote
\XeTeXcharclass `” \mathquote
% between any character of class 0 and \mathquote
\XeTeXinterchartoks \CharBound \mathquote = {\hspace{1pt}} % purposefully large for testing
\begin{document}
 
8:25 PM
@AlanMunn The kerning related to quotes in Adobe Garamond Pro is interesting. ;-)
 
@egreg But it's a problem with the math+quote, because \textit works fine.
 
@AlanMunn That could work
 
@egreg What I don't understand is why anyone would want to use $F$' in the first place. Either the ' is a prime, in which case it should be inside the math, or the F isn't math, in which case it should be \emph or \textit...
 
@AlanMunn Anyway, thinking to do math with mathspec without manual adjustments such as "f is hopeless.
 
Ok, removing all UTF8 characters has it working, sort of
 
8:30 PM
@egreg I don't use enough math for this to be an issue for me, so I'm not familiar with the problems.
 
@Canageek Is it the newest biber? Nice to see you, by the way!
 
@egreg Thanks, and yes, I updated yesterday
@egreg I've bypassed the problem by removing all the ASCII
and yeah, I've been either writing in Word or in the lab
I hurt my rist (tendonitis) this summer, so wasn't on the computer much
And when I was writing, I was using voice dictation which does NOT latex well
So I would type into Google Docs using my phone, and someone else did the formatting in Word
@egreg Well, I assume version 2.8 is the newest
Yep, it is blowing up every time it sees é in an author name
 
@Canageek Can you make up a MWE?
 
@egreg On it. Remind me of how to embed a bib file in a tex file?
 
8:46 PM
@Canageek \begin{filecontents}{\jobname.bib} ... \end{filecontents}.
 
\documentclass[]{article}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\begin{filecontents}{\jobname.bib}
@article{KillBiber,
author = {Canageek},
pages = {1--666},
title = {{Iä! Iä! Cthulhu fhtagn!}},
year = {2017},
journal = {J. Fail}}
}
\end{filecontents}
\usepackage[backend=biber]{biblatex}
\addbibresource{MWE.bib}
\begin{document}
Delete the title line and it works fine.
\nocite{*}
\printbibliography[heading = none]
\end{document}
There you go
@egreg There you go. Remove the ä and it works fine
 
@Canageek Yes, this fails. Even adding an explicit bibencoding=utf8 to biblatex doesn't help.
@Canageek Join the 21st century and use LuaTeX. :)
 
9:04 PM
@Canageek Works if filecontents is before \documentclass (which is its proper place)
 
@egreg if the latex format would write utf8 as utf8 .....
 
@DavidCarlisle It seems that inputenc gets in the way. ;-)
@DavidCarlisle Not your fault?
 
@egreg well no because the version here doesn't do that:-) (So I can say hello to @UlrikeFischer:-)
 
@egreg Doesn't work when it is two files
 
@Canageek I think it's an encoding problem: the bib file is probably Latin-1
 
9:10 PM
@egreg Notepad++ says it is UTF8
 
@Canageek I agree with @egreg it's more likely your .bib file isn't encoded the way you think. You could try adding bibencoding=latin1 to your biblatex options and see if that works.
 
@egreg But you are right, deleting the file I had on disk made tha twork
 
@Canageek E8 is e-grave in latin 1, do you have any è ?
 
@DavidCarlisle I had both, they've both been find and replaced with latex codes so it would compile
 
@Canageek Detecting encodings is a non-trivial problem from regular latin range text.
 
9:12 PM
@AlanMunn Ok, so how do I manually set it to be UTF8?
 
@Canageek the file being in latin1 seems most likely cause
 
@DavidCarlisle How do I force the file to be unicode?
 
@Canageek well I load it in to emacs and then save as utf8 (the first part of that solution is common to all solutions)
 
@DavidCarlisle I already have it loaded in emacs, how do I save it in a specific encoding?
 
@Canageek How did you create the file in the first place? (I'm with @DavidCarlisle here).
 
9:14 PM
@AlanMunn Copy pasted it from an old file that I created 4 years ago.
 
@Canageek C-<ret> f
 
@AlanMunn As for the MWE one, I...also copied that file but deleted all but the first entry
 
@Canageek which then prompts for how you want to save the file
 
@DavidCarlisle <C-Return> is undefined, what am I doing wrong?
 
@Canageek the mwe here s impossible to debug encoding issues as it is always utf8 by the time it is posted to the site.
 
9:16 PM
@DavidCarlisle Until you compile it, in which case it screws up the filecontents. :)
 
@Canageek you are doing what I wrote not what I meant to write
@Canageek C-x <ret> f
 
@DavidCarlisle And this is how all the 'undocumented features' in your packages arise. :)
 
@DavidCarlisle Yeah, looks like it was latin1. Ok, so what changed in biber that made it fail when this file has worked before? Did it somehow get switched to latin1 or is there an error in the new biber?
 
@AlanMunn quite possibly
 
Yep, saving it as UTF8 fixed it
Strange
As I said, I compiled this file recently before updating biber and it was fine
 
9:31 PM
@Canageek don't know, perhaps perl got more picky (it is usually possible to auto detect latin 1 and fix on the fly eg byte e8 can not be utf8 so it's almost certainly an e-acute so could be interpreted that way rather than an error (eg html5 mandates this permissive parsing)
 
10:30 PM
Finally at home!
What did I miss today?
 
10:41 PM
@PauloCereda nothing
 
@DavidCarlisle oh
 
@PauloCereda Nothing, apart from a hidden feature in bm. An easter egg, I'd say: instead of a math symbol, you can get the info string of a random package.
 
@egreg ooh the author must be brilliant
 
@PauloCereda good analysis
 
@DavidCarlisle :)
 
10:52 PM
@egreg with another package in a similar situation but by an author for whom @PauloCereda's analysis does not apply, the behaviour is much worse: ==> Fatal error occurred, no output PDF file produced!
\documentclass{article}

\catcode`\j=0

\usepackage{frontespizio}
\begin{document}

\end{document}
 
@DavidCarlisle oh no, a fatal error!
 
I am NEVER writing a paper in Word agian if I have a choice
Track change is THE best feature LaTeX doesn't have
(No, diff is NOT a decent substitute, I want to know what WORD I changed, not have half the document light up due to insertions in the middle)
But efff, move one thing and the figures and the captions get all messed up
 
@Canageek I've never written a document in Word:-) (I have it installed and have used it to occasionally read a few, but that's all)
 
Every time I get annoyed at floats I need to go back and use Word for a while
4
David: Yeah, the person who did the initial formatting has never heard of LaTeX, and I had tendinitis. Try writing LaTeX with voice commands.
At that point I was even having trouble from compensatory overuse, and was risking injuring my left hand
 
@Canageek sorry to hear, how are you feeling nowadays? Are you fully recovered?
 
11:06 PM
@PauloCereda I'd guess about 95%? I don't notice it most of the time, but if I use it a lotfor a few days it starts letting me know I need to lay off
Probably 98%
 
@Canageek Nice to hear. I had it about 7 years ago, it was an awful experience.
 
@PauloCereda It wasn't nearly as bad as when I had RSI in high school. This only hurt if I tired to use the hand vs all the time, and only took two months to heal.
@PauloCereda Whereas RSI hurt for much of a year, then took a year to heal and wasn't back to full strength for a year after that
 
@Canageek I can relate during my gaming phase. I had to do some physiotherapy sessions.
Of course, I always wanted to treat the underlying problem. :)
 
@PauloCereda Yeah, it is weird as I wasn't gaming much during that time
laptop out of batteries, got to go
 
@Canageek oh
@Canageek see ya, buddy!
 
11:22 PM
@DavidCarlisle And this one?
\documentclass{article}
\catcode`\y=0
\usepackage{bm}

\begin{document}

\end{document}
! TeX capacity exceeded, sorry [input stack size=5000].
\z@ ->\bm@protect \z@
                      {}
l.449 \let
          \heavysymbol\hm
!  ==> Fatal error occurred, no output PDF file produced!
 
@egreg seems reasonable to me
 
11:56 PM
@PauloCereda Maybe too much trigonometry:
-3
Q: Could be every math problem be solved using only trigonometry?

olegarioHere in Brazil, since the early days of school at Math courses we are habituated to study trigonometry a lot and the reason is clear, since a lot of problems can solved using it. So, I was thinking, if we could reduce every single math problem into a trigonometric problem, trigonometry could solv...

 
@egreg I guess that would make it sine qua non (or cosine, or arcsine...).
 

« first day (2578 days earlier)      last day (2348 days later) »