« first day (3623 days earlier)      last day (1389 days later) » 

12:15 AM
Some of the people present can tell me what function the `xindy.mem` file?
$ pwd
$ du -lh xindy*
0 xindy
1,1M xindy.mem
3,2M xindy.run
It's just out of curiosity (I don't know the extension at all)
2 hours later…
2:22 AM
@DavidCarlisle -- Oh gad! Next thing you know, they'll be trying to emulate the "tunes" "played" on line printers. That was one of the fun challenges that I remember from when I worked at the Brown Comp Lab. (Has anyone asked for the sound of high-speed card readers yet? The one attached to the RCA Spectra the Math Society had was fondly known as the "20th Century Limited".)
2:33 AM
I am trying to insert figures in latex but there is much space below and above the figure
and hence a single figure takes up much space (like a page)
\caption{Ant colony patterns.}
2:51 AM
@BAYMAX -- Instead of \begin{center} ... \end{center} (that adds space), just use \centering. There may be other causes, but that one's easy to fix. Another thing that's easy to check for is something at the beginning of the next page that's large, or has forced a page break.
hm,I changed that but seems like no difference
May be there is something else!
3:06 AM
@BAYMAX -- Is it possible that your graphic has an extra margin that needs to be cropped?
Yes that can be the case too
Thanks! I will try to look!
2 hours later…
5:11 AM
@barbarabeeton The vice Commodore 64 emulator has an option for simulating the noise from the 1541 5"1/4 floppy drive. I hated that at the time (really noisy, in the room shared with my sisters) but I love it now that you can tune it down with the volume setting...
2 hours later…
7:10 AM
Clearing LaTeX sins from Overleaf help/documentation happening just now :-)
In which TeX Live was [utf8]{inputenc} first time obsolate?
@yo' ltnews28 April 2018
@yo' I have it in final 2018, but not final 2017.
@yo' so a fully up to date tl2017 or tl2018 from the start
@DavidCarlisle @UlrikeFischer ok thanks both
hhm I disagreed with @UlrikeFischer it's in my tl2017 but maybe I'd updated it:-)
@yo' Try this in your overleaf image
$ /usr/local/texlive/2017/bin/x86_64-cygwin/latex  \\show\\DeclareUnicodeCharacter
This is pdfTeX, Version 3.14159265-2.6-1.40.18 (TeX Live 2017) (preloaded format=latex)
 restricted \write18 enabled.
entering extended mode
LaTeX2e <2018-04-01> pre-release-1
Babel <3.18> and hyphenation patterns for 84 language(s) loaded.
> \DeclareUnicodeCharacter=macro:
#1#2->\count@ "#1\relax \wlog { \space \space defining Unicode char U+#1 (decim
al \the \count@ )}\begingroup \parse@XML@charref \def \UTFviii@two@octets ##1##
7:23 AM
@DavidCarlisle ah {\raggedright\tt\meaning\DeclareUnicodeCharacter\par} :D
@UlrikeFischer oh my tl2017 is running LaTeX2e <2018-04-01> pre-release-1 by default which I guess is not a standard build:-)
(this is easier for me than to look into the logs...)
@yo' well yes that's Overleaf for you:-)
@DavidCarlisle I have LaTeX2e <2017-04-15>, but is this not a bit old? Isn't the fall release missing?
@DavidCarlisle I don't have access to the OL dev console unfortunately
7:25 AM
@yo' well I simply typed "ä" and looked if it print something ;-)
@yo' there is a view logs menu item but it's so well hidden......
@UlrikeFischer it would take me longer to type ä than \\show\\DeclareUnicodeCharacter
@UlrikeFischer that might also work. Like if Czech didn't have enough of these funny letters :D
@UlrikeFischer ltnews shows two releases in 2017 but they were January and April, so Fall fell early that year
Hello, wořld! did the trick. Thanks both!
@DavidCarlisle I can switch the keyboard layout with a click and then it is easy to find the z:
7:34 AM
@UlrikeFischer so can I (catch it accidentally sometimes) but then I don't know where any of the letters are, unless it's @ and " swapped which is the US keyboard.
7:48 AM
@DavidCarlisle no the letters are too important to be hidden in the top line, the Umlauts are beside the L and the P, only the ß in the top.
8:34 AM
@JosephWright in the following example, the second entry doesn't appear in the bookmarks as hyperref removes it. Do you have any idea what one could do here? (it came up in the context of greek numerals, the \ss is only an example for a text command.


@JosephWright oh, expanded works: \section{\expanded{\text_uppercase:n{-\ss-}}}
8:54 AM
@UlrikeFischer \text_uppercase:e ?
@DavidCarlisle no, that no longer uppercases, I then get an ß.
@UlrikeFischer Makes sense: fully expanded text is then available
9:26 AM
in listings, what's the easiest way to highlight a part of a line and draw an arrow to it?
9:36 AM
@Anush tikzmarks
@samcarter_is_at_topanswers.xyz does that do the highlighting too?
I guess I just use @@ round everything in listings so it still works?
I want to highlight $\Var{\tilde{f}_{a_i}} from bpa.st/6Q3Q and draw an arrow to it
@samcarter_is_at_topanswers.xyz I could pose it as a question if it isn't too trivial?
@Anush It is not about being trivial or not, a question is always better because then other user with the same problem can also benefit
@samcarter_is_at_topanswers.xyz done
10:06 AM
@JosephWright and now it loops ;-(. 100000 lines of \q__text_recursion_tail ->\q__text_recursion_tail . Do you have any idea how to correct the hyperref command so that it doesn't break at 1000?

\text_uppercase:n{\HyPsd@greeknumeral{1000}} %999 works fine

@UlrikeFischer I'll take a look
@UlrikeFischer Oh ... LGR :(
@JosephWright yes, but it loops also if one removes babel.
@UlrikeFischer I'll track it down: probably I've missed some condition
@JosephWright you mean it is not hyperref which should be blamed? I just added a few \relax behind this \number#2 and it seems to behave again.
@UlrikeFischer Well there should never been an infinite loop, so I must be missing an end-of-loop condition somewhere
@UlrikeFischer Meanwhile, waiting on @PhelypeOleinik before sending the big update
10:22 AM
Microsoft Excel turns 35 today
@JosephWright ah, well the relax avoids the error, but now I get a ! Missing = inserted for \ifnum., if I had an idea how this numbering system worked I would rewrite it in expl3 ;-(
@UlrikeFischer :)
@UlrikeFischer I think there are some errors in the hyperref code
@UlrikeFischer It has
    \ifnum#2<10 %
      \ifnum#2<100 %
but I don't see at all why those \number#2 should continue to expand - if I alter them to \number#2 % everything is OK
10:55 AM
@UlrikeFischer, @JosephWright : Nested links wrapping around page break + pdflatex-dev/lualatex-dev produce fatal error: ! error: (pdf backend): 'endlink' ended up in different nesting level than 'startlink'. This used to work with current engines. MWE to follow.



Visit \href{ctan.org/tex-archive/macros/latex/contrib/mwe}{mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe \hypersetup{urlcolor=red}\href{ctan.org}{on CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN} mwe mwe mwe mwe mwe}!

ooh CTAN
We could sing to the spam song thingy from the MP spam sketch!
I need my coffee, sorry.
@PauloCereda You are right. That would fit wonderfully!
@AlexG <3
@UlrikeFischer Looking at my loop code, there's not much I can do: \HyPsd@@GreekNum doesn't terminate the number search, so there is an issue unless there is a following non-expandable token. I guess I could switch to scan marks for the loop ...
@JosephWright well I think the hyperref code should be improved. Actually I'm just in contact with the babel-greek maintainer, he should take it over, that's why I'm looking at it, and the question of uppercasing came up in this context.
11:04 AM
@UlrikeFischer I'm not sure it's a bug in the text expander: as I say, with an unterminated number or whatever there is always an issue at the level of token-by-token work (I guess I could ask Bruno how he handles that for unravel)
@AlexG Ah, right: you've caught things just in time (I was supposed to be uploading today)
:55703318 Yes, both dev versions (as of yesterday). Links don't even need to be nested, by the way:



Visit \href{ctan.org/tex-archive/macros/latex/contrib/mwe}{mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe mwe on CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN CTAN mwe mwe mwe mwe mwe}!

@AlexG Got it, thanks
@AlexG I've reported to the team: I'll hold off our format update (Scheduled for today) until we address this
@JosephWright Should be out in a few hours. l3build makes it so easy to adjusting version numbers when properly set up :-)
11:27 AM
RIP Quino.
@JosephWright @AlexG I found the reason, but need to investigate a bit more.
The problem is that hyperref adds a box around the first page, but this is old code that shouldn't be active, so I need to find out why it kicks in now.
@UlrikeFischer So a hyperref issue not a kernel one?
@UlrikeFischer Where should I be looking?
@JosephWright I can repair in hyperref by removing the old code, but it is perhaps also a kernel problem.
@JosephWright @PhelypeOleinik that is the problem ^ ^ the package test fails.
@UlrikeFischer Hm... It only defines \ver@<actually loaded package>. Maybe it should define the requested package to expand to the same thing...
@UlrikeFischer Ah, right: so the shipout part is fine
@UlrikeFischer Tricky, I suppose
@UlrikeFischer, @PhelypeOleinik Fix in hyperref for the moment, so we can release, then address later?
@UlrikeFischer @AlexG So one can fix for the present with
which shows that the underlying reboxing, etc., is OK
11:42 AM
@JosephWright yes, I can make a hyperref release, but it will take a few hours (and we would get the unicode option ...). And I doubt that many people will fall into this, but we could search for ifpackagelater{atbegshi for potential problems.
@JosephWright @UlrikeFischer This should do:
  \expandafter\protected@xdef                %     <-- protected...
     \csname ver@\@currname.\@currext\endcsname{#1}% Loaded package
  \expandafter\protected@xdef                %     <-- protected...
     \csname ver@\@currpkg@reqd\endcsname{#1}% Requested package
    \typeout{Document Class: \@gtempa\space#1}%
    \protected@wlog{Package: \@gtempa\space#1}%   <--- protected
@PhelypeOleinik That works too :)
@PhelypeOleinik OK, so best to hotfix, merge to master and develop, and for me to move the tags as I've not actually released yet?
@AlexG Brilliant timing :) If @PhelypeOleinik had got REVTeX out last night I'd already have shipped
@JosephWright I think so. This seems like the right thing to do, as the package that is requested should be seen as “loaded”...
@JosephWright Phew, ;-)
@JosephWright All according to plan ;-)
11:52 AM
@JosephWright well a hyperref fix would have worked too. But I love latex-dev and people using it ;-)
@AlexG btw: with the next pdftex engine it wouldn't have failed, the nesting error is then no longer fatal.
@PhelypeOleinik You going to handle it, or should I?
@JosephWright It should be just that, so feel free to do it. I'm having my share of fun with version strings :-|
@UlrikeFischer Yes, I remember some discussion in some mailing list about that matter a while ago.
@PhelypeOleinik Ah, nowI have to work out the \IncludeInRelease business
12:08 PM
@JosephWright @DavidCarlisle I'm just looking for more outdated stuff in hyperref, that can go too. The subfigure package is in obsolete on ctan. Do you think I can dare to remove the related code from hyperref?
@PhelypeOleinik I'm sorting now then will trigger a rebuild for the tags: once you have REVTeX out, let me know and I'll upload the main release, then wait a day for the -dev-0
@UlrikeFischer I think so: really anyone updating should be fine
@JosephWright well it would break a document using subfigure. I mean people can install the package manually, and in miktex I still have it, so it is not gone. So it is really a policy question: I don't want to patch all sort of unsupported old packages, but someone could complain. (I could copy the code in an "old-hyperref-patches.sty", then they could add it manually to their document.)
12:24 PM
@UlrikeFischer Perhaps one for the list then: I'd say if something is moved on CTAN to obsolete it really can't be supported at all, but others may not agree
@JosephWright Ah, right, there's that... I wonder if there's a way to have docstrip do that in way that requires less... manual labour
@PhelypeOleinik Talk to @DavidCarlisle about \IncludeInRelease ;)
1:17 PM
Just curious, what is the proper method for coloring a math operator. My naïve attempt fails
! Limit controls must follow a math operator.
\reserved@c ->\nolimits
Editing LaTeX on three monitors is something I should have tried before. One large monitor in landscape mode with the full page filling the entire monitor. One 24" monitor filled with Emacs and one 24" monitor only showing the latex output (I compile via latexmk -pdf -pvc ... on the CLI)
@PauloCereda you got mail ...
@daleif Well, no wonder you need a full monitor, if you try to edit in a sub-OS called Emacs :)
All my question is answered here: tex.stackexchange.com/a/318698/3929
What is the fasted way to check if a number ends with three zeros?
1:23 PM
@Skillmonlikestopanswers.xyz You seem to need to talk to M-x doctor
@daleif looks more complicated there:-)
@DavidCarlisle doesn't that change the font to the text font?
@UlrikeFischer oh no, a "fastest way" question...
@Skillmonlikestopanswers.xyz ok, I didn't want to start a competition, but it shouldn't be too slow ...
@DavidCarlisle In this case it doesn't matter the color isn't there for good. I'm just missing with a manuscript and is coloring the macros that I add as it makes it easier to see the ones I have changed
1:25 PM
@Skillmonlikestopanswers.xyz and an expl3 solution is prefered even if it not the fastest.



@UlrikeFischer doesn't get faster than that.
@daleif no
@UlrikeFischer taking the rest of division by 1000 and checking if it's zero? Integer division should be fast...
@Rmano would have been my next suggestion.
@UlrikeFischer you can get catcode safe in the above by using \expandafter\ifendsinthreezeros@\the\numexpr#1\relax in the above instead.
@daleif \textcolor{red}{zzz} is just \leavevmode{\color{red}zzz} and {\color{red}zzz} would also work (you just need the color restore from \aftergroup to stay inside the \mathop)
@Skillmonlikestopanswers.xyz but then \@M ends in three zeros (which might or might not be what @UlrikeFischer wants)
1:34 PM
@DavidCarlisle well, yes, but that's the case in every expl3ish integer test, it evaluates an integer expression.
@Skillmonlikestopanswers.xyz blame the OP though for sloppy question, if integer expression needed should ask if it's divisible by 1000 :-)
@DavidCarlisle @Skillmonlikestopanswers.xyz yes sorry, I was sloppy. I want numbers in the sense of everything that tex would accepts as a number.
@UlrikeFischer in that case put the \expandafter and \the\numexpr#1\relax in the above. If you want something more expl3-like you can use the following instead:
\prg_new_conditional:Npnn \ulrike_if_three_zeros:n #1 { TF, T, F, p }
    \if_int_compare:w \int_mod:nn {#1} { 1000 } = \c_zero_int
@UlrikeFischer the argument grabbing code will be much faster though.
@DavidCarlisle ahh, thanks, that makes sense
@UlrikeFischer /sad quack
1:43 PM
@PauloCereda there are adventurous ducks ;-). They will go on a trip again ... perhaps this time with a tracker attached.
@UlrikeFischer no please, tracking costs a fortune... keep them
@UlrikeFischer q for you on the main site :-)
@DavidCarlisle context ...
@UlrikeFischer Indeed :)
@UlrikeFischer partly but I think asking mostly about luatex and open source tools to use the pdfs once made
2:02 PM
@samcarter_is_at_topanswers.xyz ...so it will be changed in beamer? is there an issue for following this? Thanks!
Grr, I hate latexrelease guards
Hmm, you see some strange things when you edit other peoples stuff, this one I don't understand: word\emph{\index{word}}, other places in this manuscript I've seen \textbf instead of \emph. Not sure if some of it has come from Scientific Workplace, it is clear that this document has been edited in SWP at some point.
2:18 PM
@JosephWright I wonder who put them there.
2:29 PM
@JosephWright collective responsibility
@Skillmonlikestopanswers.xyz Frank and me, mostly:-)
2:47 PM
@JosephWright ctan.org/incoming
@PhelypeOleinik Cool: just need to get the bugs in my change sorted then can tag and upload
@JosephWright Those often tests are loads of fun
@JosephWright I have put them in a variable, so I can just l3build save -e{etex,xetex,luatex} $rollbacktests
@PhelypeOleinik regression tests that are expected to fail if you change anything.....
@DavidCarlisle I heard they were called \CheckSum :-)
3:06 PM
@PhelypeOleinik if you were as old as egreg you'd recall a time when those checksums were useful and warned you that a janet-bitnet gateway had eaten the second half of the file....
3:27 PM
@Rmano I have the changes in a branch github.com/samcarter/beamer/tree/bold but have no idea if beamer should do this or not. @JosephWright any opinion?
@PhelypeOleinik All fixed, building now. I should ha e gone with a branch as we’d only ha e one final combined commit!
@JosephWright “v” key is broken? :-)
3:43 PM
@PhelypeOleinik Or he's moved to Scotland.
@Skillmonlikestopanswers.xyz /hugs back
4:01 PM
@samcarter_is_at_topanswers.xyz Sounds like a possible, but risky - we'd likely break documents
@PauloCereda /pats back
@UlrikeFischer the expl3 variant should also test whether #1 isn't 0 (I forgot)!
@JosephWright so blaming @UlrikeFischer for the changes in the kernel instead?
4:22 PM
@AlanMunn :-)
@samcarter_is_at_topanswers.xyz @UlrikeFischer maybe you can post the workaround you suggested to me to help the OP of tex.stackexchange.com/questions/564794/… ?
It's a viable stopgap while the thing is sorted out (notice that as it is now it does break documents --- Ubuntu 20.04 is being offered for upgrade already, and that means a lot of users going for TL17 to TL20...)
@Rmano In my experience an incorrect fontseries in a presentation is the least thing to worry in an unbuntu upgrade - I still have some open problems from the update to ubuntu 18 ...
@samcarter_is_at_topanswers.xyz Oh yes, that is true (this is why I never do that if it's not a quiet period, and lately such periods are a really scarce asset). But normally you can trust LaTeX ;-)
@PauloCereda :-( he left Mafalda to us, which is a great gift.
5:19 PM
@Rmano \renewcommand\bfdefault{bx} before loading the class should work.
@UlrikeFischer Wow, even easier.
@JosephWright @samcarter_is_at_topanswers.xyz it is really a bit unclear. I mean in \SetSymbolFont{operators}{bold}{OT1}{cmss}{\bfdefault}{n} one can simply use bx, in \SetSymbolFont{operators}{bold}{OT1}{\sfdefault}{\bfdefault}{n} exchanging \bfdefault by \bfseries@sf makes sense, but in \DeclareMathAlphabet{\mathbf}{\encodingdefault}{\mathfamilydefault}{\bfdefault}{n} not so really ;-( One the whole everything would be easier if the fonts would have always both, b and bx.
@Rmano new version of my name ;-)
@UlrikeFischer Ooops, fixed! /blush
@Rmano don't worry, everyone seems to write it wrong now and then.
Aug 24 at 7:55, by Ulrike Fischer
@DavidCarlisle I acquired various names this summer, I already found Urile Utrike Urlike and now Ulri ;-)
@UlrikeFischer :-) I'm used too to a random number of double consonants in my surname...
5:33 PM
@Rmano my surname normally only looses the "c" ;-)
@UlrikeFischer Does it ever migrate to your first name?
@AlanMunn sure, Ulricke has been seen too ;-) (Or as @Rmano just wrote, Ulriche)
6:42 PM
@UlrikeFischer ooh-rike
7:40 PM
@UlrikeFischer No idea if \DeclareMathAlphabet{\mathbf}{\encodingdefault}{\mathfamilydefault}{\bfseries@sf}{n} makes sense or not, but without I did not get bold font.
@JosephWright interesting interaction between pdfx and biblatex :)
A: Error after MiKTeX reinstall: Text line contains an invalid character

David CarlisleThis is an unfortunate side effect of using the pdfx package. A simpler example is \documentclass{article} \usepackage{pdfx} \begin{document} \iffalse ^^^^200b \fi \end{document} You could make the catcode of character 30 normal (12) rather than invalid (15) \documentclass{article} \usepa...

@DavidCarlisle ups.
7:58 PM
@UlrikeFischer It seems a bit aggressive of pdfx to make the control characters catcode 15
Hello Tex community, I am rather new to the LaTeX playground. What do you all recommend as the best mode to use LaTeX? For example, right now I am using WinEdt for databases and Overleaf for the main processing. Does this matter much or should I look into investing in a more reliable system?
@MagisterAdest what do you mean by winedt for databases? (winedt is an editor you presumably have a local tex system miktex or texlive backing that up?)
@DavidCarlisle I am using miktex, yes. I mean to ask what is the best editor to use? Thank you for clarifying
@MagisterAdest The best editor is the editor you are happy with :)
@MagisterAdest well that's a question on which there is no single answer. personally I use emacs but I have used that more or less every day for over 30 years so hard to change now. We have an answer on site giving pros and cons of several editors for tex
8:04 PM
Databases such as a saved glossary for any work I met do in the future, and another doc saved with all the books I have read so that I can put them in a Bibliography later
@MagisterAdest there are several reasons to use an online system such as overleaf (one which doesn't apply in your case is that it saves installing a local system at all) and several reasons for using a locally installed one, but to be honest I can't see why you would edit the bibliography files locally but your main text document online?
@DavidCarlisle Honestly just because my professor currently only uses WinEdt and he got me started using that for the bibliography files. You have a good point with this, luckily with a quick move it can easily be transferred to the online system.
@DavidCarlisle pdfx is a bit agressive, but basically it is doing this (but then doesn't restore the encoding correctly:



@MagisterAdest one of the highest voted questions on site is:
Q: LaTeX Editors/IDEs

hayalciWhat editors/IDEs are available for easing the process of writing TeX/LaTeX documents? Please state some useful features like code completion, spell checking, building final DVI or PDF files, etc. This question is undergoing a systematic refurbishment, see Let’s polish the Editors/IDEs question ...

@DavidCarlisle Thank you very much for helping a newbie slowly climb the path to documentational excellence
8:11 PM
@UlrikeFischer oh we set it to 15, hmmmm
@MagisterAdest On bibliography files: if you are a grad student doing research, it's generally more practical to have a single .bib file rather than one per document, and this is a big advantage of having a local installation compared to Overleaf.
@MagisterAdest But editor choice is very personal, so don't feel pressured to use emacs just because @DavidCarlisle uses it. There are some crazy people here who use Vim too, and people like me who use neither (I use TeXShop on the Mac).
Seems the same @DavidCarlisle: tex.stackexchange.com/questions/565002/…
@AlanMunn yes but only if you process your tex files locally as well, if you copy it up to overleaf to do the tex processing
@DavidCarlisle Yes, that's what I meant.
@AlanMunn Biblatex allows for remote bib files. So one could host the global file one github or similar and then use this in one's overleaf projects
8:21 PM
@DavidCarlisle yes, but I don't understand what pdfx is doing. It is trying to reset it, but the code is a complete mess.
@UlrikeFischer biblatex has changed to avoid it:-)
@samcarter_is_at_topanswers.xyz True. That's a neat idea. Not maybe beginner level though.
@Rmano thanks I voted it as duplicate although that new one doesn't show eough code to see what has made control-^ invalid in that case.
@AlanMunn It's certainly easier than making some letters bold :)
8:55 PM
@DavidCarlisle I found the error in pdfx: It restores \inputencodingname to utf8 and then calls \inputencoding{utf8} but this is doing nothing, as it thinks the encoding is already correct. He must have realized at some point that the code is wrong, and as workaround he made all chars >127 active in a loop. But the catcodes below are then still wrong.
@UlrikeFischer ah you should get push access to the repo:-)
@DavidCarlisle there is no repo ....
@UlrikeFischer you could make one from the ctan files, that went well for bidi...
9:17 PM
@DavidCarlisle ;-). the input encoding handling in this package is quite enerving. He switches quite often between 8bit utf8 raw encoding. But I guess I should write him about this, it is not good if documents have basically a random catcode state when the package is used.
Hello everyone,

I would be grateful if I could get a help to understand how to make paracol makes the same output as I did with tabular in the following document

\usepackage[a4paper, top=1.9cm, left=2cm, right=2cm, bottom=2.2cm]{geometry}


paracol doesn't make the columns have the same widths of those of the tabular
9:34 PM
@Diaa p{0.34\linewidth} makes a column .34\linewith+2\tabcolsep wide
@Diaa the internal line with for line breaking is the argument to p but there is \tabcolsep padding on either side
@DavidCarlisle I see. However, I don't know how to handle it though :)
If you don't mind, would you please show me how? And also how to use \columnratio to get the same expected output?

@Diaa you could set \tabcolsep to 0pt so there was no padding or use p{\dimexpr.34\linewidth-2\tabcolsep}
@DavidCarlisle I realized now that my original code sets \tabcolsep to zero already
@Diaa so it does, I didn't actually run it:-) I'll have a look in a bit....
@DavidCarlisle Thanks in advance
9:52 PM
Btw, today is my Overleaf 1st anniversary!
@yo' Ive been at NAG nearly 23
@DavidCarlisle This means you must be almost as old as @egreg!
@yo' nah still a youngster both in real life and at nag: several people worked here longer (10 years longer in one case at least)

I have an inquiry: recently, does overleaf update the packages database regularly?
because the biggest reason I gave up on overleaf a couple of years ago was its old database.
I mean if a package has been updated today, when should I find this updated version on overleaf?
@Diaa It depends on what you mean by "regularly". I think they are much more up to date than they used to be but no, I don't think yesterdays updates are available. @yo' could tell you more authoritatively.
10:04 PM
@Diaa Currently we're on TL19. We'll have TL20 very soon.
@yo' To understand you correctly, do you mean that the packages database update is done on a yearly basis?
@Diaa yep, that's exactly the case. But for faster updates, you can simply get the packages from CTAN and put it in the project.
Remember that we have to preserve stability -- it must never happen that a project that worked suddenly stops working because we did an update.
@yo' I totally understand. Thanks for your clarification.
@Diaa You're welcome.
@Diaa being (on average) about 6 months behind is pretty good really and way ahead of typical journal systems, or arxiv for example
10:09 PM
There's actually some discussion about making the TL image available in our beta programme sooner (June rather than September), but that's still up to discussion.
@DavidCarlisle That's what we think too :-)
@yo' of course it would have been better still in a previous era when the latex format didn't change for 10 years:-)
(except that previewing the pdf over dialup paying by the minute wouldn't have been much fun)
@DavidCarlisle well, a release once a year from a DVD: You would simply do a verbatim copy of the version on the DVD. No testing, as there would be no way to correct bugs/incompatibilities anyway.
@DavidCarlisle As far as I remember, in Jan 2019(I don't quite remember), the TL version was 2017 which caused me a lot of troubles.
@DavidCarlisle arXiv plans an update after over 3 years.
@Diaa That could be. TL2018 came in Sep 2019 and TL2019 came in Jan 2020.
There were other things to do with the merge of Overleaf and ShareLaTeX.
@yo' you could add some overleaf insider knowledge here perhaps tex.stackexchange.com/questions/565013/…
10:18 PM
@DavidCarlisle I'm not at work anymore, but I commented :-)
@yo' thanks
@yo' seems like I gave the right answer:-)
@DavidCarlisle and yes, it's technically impossible for two users to see a different set of aux files. It would be necessary for LaTeXMk to fail in a way that would alternate a good and a bad version, and the two users would have to hit Recompile in an alternating fashion too :)
I mean, you can be unlucky enough to get a server error and a crippled compilation, but this should really only happen for borderline-timeouting documents in a case that the compile server is killed before the compilation is finished.
@yo' I suspect the OP has two projects his original (with the error) and a copy made for public viewing (so with a clean aux file and no error)
@DavidCarlisle Yeah, that's why I suggested what I suggested :-)
Jan 14 '15 at 21:56, by David Carlisle
@PauloCereda never believe what they tell you
10:28 PM
10:52 PM
Hi again everyone - is it okay to share a read-only link to overleaf.com in my question instead of posting the whole code?
@VinayPandey not really. Your question is archived forever here so unless you plan never changing the overleaf project the question will not make sense to later readers
@VinayPandey That's a good question. Probably not, as David points out.
But I feel that a cannonical Q&A on this on meta would make sense... (cc @JosephWright)
Anyway, I'm rushing off, see you another day!
@yo' have fun (or sleep, or whatever)
@DavidCarlisle yeah, that's it :D
@DavidCarlisle are you in UK ? it must be pretty late here. Thanks, I guess I am going to post another of my stupid questions on StackExchange
11:03 PM
@VinayPandey I normally don't follow such links.not so much because the link gets invalid (that too) but because such projects consists of many often quite large files. It takes too much time to wade through them.
@VinayPandey yes, it's tomorrow now:-)
@VinayPandey 999 times out if 1000 you can make a small example more suitable for posting here which is anyway kinder to people offering to debug it for you who have to trace the code and a preamble that loads 1000 unrelated packages doesn't help.
11:28 PM
Got it, thanks for clarification @DavidCarlisle and @UlrikeFischer
I am in California, and work daily with Londoners. I was wondering if it is too late for you @DavidCarlisle

« first day (3623 days earlier)      last day (1389 days later) »