3:39 AM
Anybody has an idea why compiling the PGF manual does not converge randomly? Pass: travis-ci.com/hmenke/pgf/jobs/263688500 Fail: travis-ci.com/pgf-tikz/pgf/jobs/264275185 And I'm building literally the same commit both times!

2 hours later…
5:40 AM
@DavidCarlisle teh

2 hours later…
7:23 AM
Have they ever tried LaTeX???

7:36 AM
@EnthusiasticEngineer Once I sent a .zip file to a collegue of mine and she replied to me: "I'll open it later, I have to ask for help to unzip it". That's why people use Word :)

7:48 AM
hey folks, I need to use a non-integer font size as default. is \def\normalsize... the best way to do this?

8:27 AM
@EnthusiasticEngineer One should not pay too much opinion to random stuff on the Internet.
Mind you, some of the answers do have a point. E.g.
> And in the real world, nobody cares, that your documents are beautiful. The content will make the difference.
Though clearly "nobody" isn't quite correct. And I'm happy that Knuth did care.
And anyway, it's not like one has a choice. It's either TeX, or something much, much worse.

8:44 AM
@CarLaTeX Exactly, It was a period I used to work in an engineering firm, the direct boss of mine who was an engineer and head of engineering design unit could not open Excel! You are right, people's skills is so much important...
@FaheemMitha Yes but some of the answers there have much confidence behind them that make me surprised! We have lots of perfectly maintained and currently developing ones, we have LaTeX 3 being developed by the lovely experts, there are many many TeX users; and answers there firmly insist: LaTeX is Dead!!!!!

@EnthusiasticEngineer Yes, to write in LaTeX you need some minimal computer skills not any people have

@CarLaTeX Skills and Interest!

@EnthusiasticEngineer I expect most of the people writing those answers don't know much of anything about TeX. They may not even use it.

@FaheemMitha Exactly

Though they're right that TeX does have substantial overhead. The expedient and ugly choices are easier, at least for small documents.

8:49 AM
@EnthusiasticEngineer Yes, if you don't want to learn you will never learn. My mother learnt to use Whatapp at 80 years old, and she has only an elementary school licence, but she is curious and wants to learn new things

@FaheemMitha Exactly the same as the other aspects of life; most of the people are interested in pop music and less are into classical and opera! LaTeX has its unique place in unique people's heart! :))
@CarLaTeX good to hear that... yes some people want to learn and many do not.
And about my linux learning: I am also little by little getting used to my Linux OS! First it was hard, now I am very confident about it... learnt some codes, etc... In long term I hope I be able to shift to it!

@EnthusiasticEngineer Good, I would like to try, too, as soon as I have time to do it :)

@CarLaTeX first I was afraid to do so, but now I am getting more used to it... it is like my eyes are getting comfort with darkness! :))

@EnthusiasticEngineer I have some vision problems, and I like dark themes, I set them in any editor I use

@CarLaTeX it has good features... and if you are busy, it does not take you much time to install any distro you like. about one hour...

8:57 AM
@EnthusiasticEngineer Good to know :)

@HenriMenke the problem seems to be somewhere in the first 106 pages. From the licence text on there a page more.

9:19 AM
@UlrikeFischer But why is it non-deterministic?
@UlrikeFischer There are no random numbers being used before page 106.

9:35 AM
@HenriMenke I don't know I'm on my smartphone currently and could only look for clues to restrict the problem a bit. I can look more later.

9:55 AM
Would you guys know or say that there are or should be different page layouts for documents that are primarily target as electronic only/never printed? Like, narrower margins, different page size, etc? Is there any such norm in the wild?
I always have a feeling I should use narrower margins than the defaults, for example.

10:10 AM
@wilx When I prepare stuff for me to read on my iPad, I try to use as much space as I possibly can. :)

@PauloCereda Do you size the page to exact size of the screen of the iPad?

@wilx I actually never did that, but I tend to use A5 instead of A4. :)

10:27 AM
@wilx why should be worried about it when PDF files normally fit to screen of any desired size?

@EnthusiasticEngineer So that the font size is reasonable for the screen size?

HEAR, HEAR: TAOCP volume 4, fascicle 5 is out!

@wilx I am not aware of the technical side, but I have heard that (and this is what I experience while reading papers and books in PDF format) when a text is converted to PDF, everything is dynamic from the margins, text size, etc. So, when you fit it to screens with different sizes, everything changes with this regard...
That is how we can display perfect PDF presentations in our laptop and on a big screen...

This is a short excerpt from an extensive work that still lacks various queries, design tools, etc.

That has its reason why I use pgfplotstable. You just have to believe me.

I'm not the best TeXer ever, but I know what I'm doing!

10:49 AM
@DavidCarlisle ooh picture

@EnthusiasticEngineer ?
@PauloCereda an important environment
Oct 4 '16 at 22:18, by David Carlisle
@Canageek 1st rule of tex support: never believe users when they describe what happened.

@DavidCarlisle do you want to kill me again? :((

@EnthusiasticEngineer did I kill you already? (Hard to remember these details:-)

@DavidCarlisle :)) yes, and now I am a dead who never uses \bf instead of \textbf with extra brackets!

@wilx margins (as a place to put your thumb while holding a book) don't really have a place but reduce the margins by reducing the page widths, not by making the text block wider (as a general rule)
@cis there may (possibly) be a reason to use the pgfplotstable typesetting command, but there can be no reason to write out data to a text file to read back using pgfplotstable file reading, as that just defines a macro that you could define directly in tex but does it orders of magnitude more slowly.

11:32 AM
I almost dropped the idea with paged text files.

I'm trying it now in just pgfplotstable.

1 hour later…
12:48 PM
@DavidCarlisle OK.

1:14 PM
On Mathematics Meta, somebody has asked this question: Why the bracket(s) disappear? Basically, it seems that their after converting formulas to images. (Although the question is about specific website.)
I have suggested using Ghostscript as a possible way to do that. Is that a reasonable solution?
Is there some post on TeX - LaTeX to which I could direct this user?
(The OP asked about converting "equations written in MathJax" - but I suppose as far as we're talking only about stuff used in math mode, there aren't big differences between the basic usage of MathJax and LaTeX.)

209

I need to convert my LaTeX documents into PNG. The problem is, I also need the resulting image to be as short as possible (height-wise). I've tried latex followed by dvipng, but the result is always the size of a page. For instance, take a .tex file containing: \documentclass{article} \begin{d...

Thanks! I was about to write that I linked to the same post.
in MathJax, 4 mins ago, by Martin Sleziak
Possibly answer to this question might contain some suggestions: Compile a LaTeX document into a PNG image that's as short as possible
I have encouraged the user who asked the question to come to this room if they have questions - I hope that's ok.

1:39 PM
ugh, I can't get smallcaps working with a otf font in context, and I also can't find any relevant information despite scouring the context wiki & mailing list... i know the font has the smallcaps feature, but either i am not enabling it correctly or enabling it does nothing

1:54 PM
I think this is something with let, edef, exandafter a.s.o. :(
0

I get only the last value (that is '0') expanded in the following pgfplotsinvokeforeach-loop. What have I to do? BTW: I would like to keep create col/set list. I hope it is possible. \documentclass[a4paper, landscape]{article} \usepackage{pgfplotstable} \pgfplotsset{compat=1.16} \begin{docu...

2:37 PM
My first palindrome:
2

3:13 PM
Hi all, I have a font which has \prime already a "superscript", in the sense that the regular definition of active $'$ in latex leads to an optical effect of a double superscript
What are my options? is this the point where I redefine prime from the latex kernel? it feels dirty

4:02 PM
@1010011010 almost every non math font has that problem:-) but you should almost never have \prime in any document, all you need to is make ' a normal mathcode so it acts as a normal character then x' will make x' not x^\prime
@MartinSleziak the original tex expression is in error, so most tex emulations on web pages will just do some arbitrary thing.
@cis it's still vastly more complicated than needed.

4:28 PM
@DavidCarlisle I'm considering to remove pdfencoding=auto and pdfdoc from hyperref ... Doing everything in unicode seems so much easier.

@DavidCarlisle With the asterisk that one does not want to break existing primes in the document of course, but yes much of the definition can be copied and ..simplified in view of the "multiprime" functionality not being required

@1010011010 the document markup shouldn't need to change
@UlrikeFischer yes

\def\'{\text{\XeTeXGlyph1158\relax}} ?

4:43 PM
@1010011010 no that is defining the acute accent command, \' not ' something like {\catcode\'\active \xdef'{\text{\XeTeXGlyph1158\relax}}}

Ah fair enough

Hello, some of the users of TeXLive (Updated) in windows can check the output of tlmgr check all

4:59 PM
@1010011010 gdef not xdef

@DavidCarlisle This appears to lead to "doesn't match its definition" errors... (?)
Both xdef and gdef
Never mind
All it needed was an extra pair of braces

@1010011010 no gdef should work (xdef less so)

I copied your first try without the braces that was the culprit

@1010011010 ah yes copying untested code off random internet sites has its drawbacks

@DavidCarlisle It even has a name: spaghetti code

5:05 PM
@DavidCarlisle Only later we noticed that there is unpaired \left..\right. But eventually we figured it out.
In any case, thanks for response!

@DavidCarlisle I think oberdiek bundle should now be called oberdiek b ... after so many cuts to become a little smaller :-). That sword will need to be sharpened again

5:21 PM
@PabloGonzálezL blame @UlrikeFischer it's not safe if she is waving sharp blades about

@DavidCarlisle jeje, a consultation, will you divide it completely?

@PabloGonzálezL no probably stop after this next batch, should be around 50 left, of less used packages, and I'll just pull them out if they need updating

5:37 PM
Great, the listingsutf8 package is on your list?

@PabloGonzálezL possibly, do you use it?

@DavidCarlisle Yes, in conjunction with tcolorbox
And I would bet that many users (who don't read the listingsutf8 documentation) load it without knowing what its purpose is :)

5:51 PM
@DavidCarlisle Thanks to the current division, I no longer charge oberdiek in one of my configurations in travis :)

@DavidCarlisle you are fast ;-)

@UlrikeFischer Really fast, I'll wait for the update and remove oberdiek...every second counts in travis!!!!

@PabloGonzálezL it's still needed if you use hyperref, I wondered about pulling out all the hyperref dependencies but it would be better to fix hyperref not to use them.

6:08 PM
@DavidCarlisle Mmmm....mmmm....as an idea, if some packages are only working with hyperref, maybe it's better to take them out (for now), then, when hyperref is set, you just declare them as inactive and they won't need updating anymore (apart from hyperref no other package uses hypcolor for example).

@DavidCarlisle you mean the bundle? Which packages are still needed in hyperref?

@UlrikeFischer vvv
$pdflatex hy1 | grep /oberdiek (/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/ltxcmds.sty) (/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/kvsetkeys.sty (/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/etexcmds.sty (/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/kvdefinekeys.sty) (/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/pdfescape.sty) (/usr/local/texlive/2019/texmf-dist/tex/latex/oberdiek/hycolor.sty (/usr/local/texlive/2019/texmf-dist/tex/latex/oberdiek/xcolor-patch.sty @UlrikeFischer ltxcmds and etexcmds seem completely pointless in hyperref, it only works in latex anyway so why \ltx@ifundefined rather than \@ifundefined.... @DavidCarlisle yes. @UlrikeFischer I'm not keen on highlighting kvdefinekeys either, it's a tweak to keyval but given where we are now things should converge on keyval or l3keys (or pgfkeys if living in a pgf world) @UlrikeFischer I have a version of hyperref more or less done that writes out those two. @DavidCarlisle I'm not too keen about pgfkeys - I'm not getting along with its system to define and set keys with one command - so l3keys ;-) @DavidCarlisle I will try to find what this xcolor-patch is doing. 6:17 PM @UlrikeFischer why anyone ever deviates from the one true keyval is beyond me. @DavidCarlisle ;-) I wanted to give an update miktex answer but i see you had commented already. 2 hours later… 8:12 PM @JosephWright did you see the hyperref discussion about how "text-purify" should handle ligatures? @UlrikeFischer Yes :) @UlrikeFischer They are a bit 'interesting' as they are already ASCII text @UlrikeFischer They fit really with the accents question: I could arrange that \text_expand:n simply turns them into Unicode: perhaps we need \text_unicodify:n @JosephWright well the main question is how configurable this should/can be. What to do if some people want -- -> -- and other -- -> endash? And what if both is wanted in one document? 8:35 PM @UlrikeFischer I've just tested \text_purify:n + \str_set_convert:Nnnn against \Hy@pdfstringdef using l3benchmark: hyperref 4.33e-4 seconds (1.58e3 ops) expl3 0.001 seconds (3.2e3 ops) @UlrikeFischer We can add mechanisms ... if we want to @UlrikeFischer Now, just need some idea of a spec for math mode @JosephWright I wished l3benchmark would use the same unit for all numbers ;-(. How much of the various replacements, redefinitions etc of pdfstringdef is in \text_purify? @UlrikeFischer A small subset, but done using a hash table lookup so should not be too bad in performance terms @JosephWright From the math that I would expect in this place the most problematic are probably limits and sub/superscripts. 8:50 PM @JosephWright a similar issue is @Sebastiano's question about math as unicode strings, as he wants to go further, using superscript characters etc @DavidCarlisle Yes: and those tools all skip operators! @DavidCarlisle Try \sin for example :) @UlrikeFischer Needs a spec @JosephWright at first it needs an idea how to write this in the bookmarks ;-). I don't think e.g. \lim_{x\to\infty} has a sensible translation. @UlrikeFischer at times at work we have to generate matlab documentation, and in the GUI there the document is basically html but the built in table of contents system is similar to pdf bookmarks, and restricted to plain text. Have various heuristics (including re-writing the section headings to avoid issues) but the automatic conversion if necessary falls back on essentially an asciimath linearisation, so tex-like but without \ , so lim_{x → ∞} in that case 9:06 PM @DavidCarlisle that sounds like a plan. @UlrikeFischer we don't actually generate legal asciimath, but if looking for a spec, asciimath.org @UlrikeFischer -- There are some really strict limitations on what Unicode can (or will be "allowed to") do; I had this discussion with the UTC. We're headed out to dinner (it's Gordon's birthday), but I'll address this mess when I get back. @DavidCarlisle @JosephWright ^^^ @barbarabeeton I'm at my Mum's because it's her Birthday. Happy birthday to Gordon:-) 2 hours later… 11:00 PM @PabloGonzálezL I was distracted ^^^ :-) 11:27 PM @UlrikeFischer I'm minded to go with Chris's idea: $\lim_{x \to\infty}$ is 'it' @JosephWright you mean the verbatim tex code? @UlrikeFischer Yes @UlrikeFischer Nice and easy to do :) @JosephWright what about simple stuff like x=y or 1+2=3? @JosephWright I'm with Chris on this one! 11:56 PM @UlrikeFischer Yes, I see that: something like 'Properties of$x\$`' seems like something we could handle
@yo' I see that argument too: one might make 'linearisation' of math mode optional, with big caveats about the outcomes