« first day (2896 days earlier)      last day (2029 days later) » 

David Carlisle
David Carlisle
12:27 AM
@vlg \p@ is just pt it's equivalent to using 65536 as I did above. So you can use any other length in the same way. \p@` is handy though as it reverses teh dimen-to-integer conversion if you assing 7pt to a count register it is 7\times65536 as an integer so dividing by \p@ gets you back to 7
 
manooooh
manooooh
12:57 AM
 
marmot
marmot
@manooooh +1 but your if does not work, try
\ifnum\iangle=0
\else
\path (M.south) coordinate (Ms) pic[my angles=\SI\iangle\degree] {angle=H--Ms--G};
\fi
(I did not test, all I know is that you need to use `\ifnum` to test **integers**, for real numbers you can cheat by pretending it's a dimension).
@manooooh BTW, you may try to tell @Skillmon very politely that the angles library got superseded by angles.meta. But you need to be very polite, from his avatar he looks like a person who might consider eating ducks. ;-)
 
manooooh
manooooh
@marmot yes, many thanks! Btw I want to delete that label when \iangle=0, I express myself so bad :/
 
marmot
marmot
@manooooh Yes, but that's what happens here: there is an \else.
 
manooooh
manooooh
@marmot oh hahahaha, Overleaf says in \usetikzlibrary{arrows,decorations.markings,angles.meta} "I don't find the library angles.meta" O.o
 
marmot
marmot
@manooooh Yes. My bad (Sorry @Skillmon!) I meant to write arrows.meta.
 
manooooh
manooooh
1:07 AM
@marmot lol yes. I mean: I wanted to delete it no matters the \iangle value, so finally I delete the last path and the last draw hihi, I don't like it so much, but the output is very amazing! @Skillmon
 
marmot
marmot
@manooooh Yes, truly amazing: 53+20=90. ;-)
 
manooooh
manooooh
@marmot :D
I was wondering how to write the 53º label and the arc in the other side, next to ||F||cos53º and not next to the gray box. He probably will see my comment
 
marmot
marmot
This is all determined by the `my angles` style. Try:
,my angles/.style=
{
,draw=green!70!black
,->
,angle radius=14mm
,angle eccentricity=0.7
,pic text=#1
,font = \small
}
@manooooh the eccentricity determines whether the angle number is inside or outside the arc and the radius is, well, the radius. Seems like @Skillmon is more eccentric than you are. ;-)
@manooooh BTW, if you do
,my angles/.style=
{
,draw=green!70!black
,->
,angle radius=14mm
,angle eccentricity=0.7
,pic text=#1
,font = \small,opacity={ifthenelse(\iangle==0,0,1)}
}`
then the angle won't show if `\iangle` equals 0...
 
Alan Munn
Alan Munn
1:27 AM
Any github people around?
 
marmot
marmot
@AlanMunn No, only a duck and a marmot. ;-)
 
Alan Munn
Alan Munn
:)
 
manooooh
manooooh
2:00 AM
@AlanMunn I have git!! I'm a newbie, what do you want?
 
Alan Munn
Alan Munn
@manooooh I'm looking for opinions on whether it would be possible to use github to track typos in a textbook that I use. This is a published book, so I don't have sources for it, but it seems to me that I could use the bug tracking facility of github (actually probably gitlab) to allow people to submit typos.
 
manooooh
manooooh
@marmot wow, the Algebra's course said the same as you say! :)
@marmot although it's great to be able to put the number in, I was looking to something like this: vvvvvvvvvvvvvvvvvv
@AlanMunn that sounds amazing! How could I help? I don't have technical knowledge in LaTeX/github, but if it's enough for you, tell me what I have to do
Although I like to learn
 
Alan Munn
Alan Munn
@manooooh Thanks, but I don't really need help. I have all of my own LaTeX projects on github already, some public, some private, so I'm fairly familiar with how things work. I was mainly asking to see if anyone with lots of project experience might have an opinion on this unorthodox (mis)use of the bug tracking facility for typos.
@manooooh The idea would be that I would point my students to the bug tracker, and every time they find a problem with the textbook they could file a bug report. If it worked I'd advertise it more generally to other linguists who might be using the textbook and we could crowd-source an errata document, and ultimately pass it along to the authors.
@JosephWright Good morning! Might you have some comments on the above?
 
manooooh
manooooh
2:16 AM
@AlanMunn ok. I like that you insert the LaTeX and git (in general, technology) to the task of the students. How are the typographical errors that appear?
 
Alan Munn
Alan Munn
@manooooh Well in this case there would be no LaTeX involved, just bug reports. Although the errata document would be in LaTeX I guess. They really are a variety of things. Some things are things like an exercise that says "Draw trees for this sentence" when there's only one tree that could be drawn. Others are more substantive, like a + that should be a –. It's a real mixture.
 
manooooh
manooooh
@AlanMunn good
 
 
2 hours later…
manooooh
manooooh
4:26 AM
This has always been happening to me, but now I want to share this problem with you. Please watch this video
Apparently, when you try to quickly click two different buttons on the Stack Exchange site, the page is bug and the user and password magically disappear. This I only tried with Google Chrome. Any idea what is happening? Should I publish it elsewhere?
Is this a bug?
 
 
2 hours later…
Joseph Wright
Joseph Wright
6:02 AM
@AlanMunn Seems entirely reasonable: you can just have a README or whatever as the 'source'
 
 
2 hours later…
Skillmon
Skillmon
8:11 AM
@marmot I should know that arrows.meta is the correct library to use nowadays... I've made an edit.
 
AlexG
AlexG
Ducks are brave:
 
Skillmon
Skillmon
@AlexG looks more like it is afraid of water.
 
AlexG
AlexG
@Skillmon ;)
user image
6
 
Skillmon
Skillmon
8:27 AM
@marmot Any idea why I can't use a node anchor in the argument of angle (e.g. here I used this ugly \path (M.center) coordinate (Mc) pic [my angle=...] { angle=fr--F--Mc}; workaround instead of \pic[my angle=...]{angle=fr--F--M.center};)
 
Paulo Cereda
Paulo Cereda
9:23 AM
@AlexG ooh
 
daleif
daleif
9:41 AM
@DavidCarlisle Remember why we needed contact with Frank Küster? There was an error in a class due to a change in the core. I now have an active email on him. I'll forward it via private email (I think I have your address somewhere)
Hmm, I think is was labbook
 
 
1 hour later…
marmot
marmot
11:08 AM
@Skillmon I do not know for sure but think it comes from
\def\tikz@lib@angle@background#1--#2--#3\pgf@stop{%
\path [name prefix ..] [pic actions, draw=none] (#2.center)
-- ++(\tikz@start@angle@temp:\tikz@lib@angle@rad pt)
arc [start angle=\tikz@start@angle@temp, end
angle=\tikz@end@angle@temp, radius=\tikz@lib@angle@rad pt] -- cycle;
}
in tikzlibraryangles.code.tex . Note the `#2.center` in the second line. So if you have `M.south` for `#2` it will become `#2.south.center` which does not exist.
 
Paulo Cereda
Paulo Cereda
11:57 AM
@barbarabeeton one for your column? :) sansforgetica.rmit
 
David Carlisle
David Carlisle
12:34 PM
@daleif I got your mail saying you'd made contact but I couldn't remember the issue:-) Oh is this an \@ifundefined breakage?, I'll see if I can reconstruct...
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle regarding github.com/lualatex/luaotfload/issues/416: I pushed a first version here github.com/u-fischer/luaotfload/commit/…. Perhaps you would like to take a look (the build failed but only because the name of the xits math font changed ;-(.)
 
David Carlisle
David Carlisle
@UlrikeFischer completely forgotten about that! (will test yes)
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle well I thought I should check the issue list.
 
David Carlisle
David Carlisle
@PauloCereda it works ^^
 
Paulo Cereda
Paulo Cereda
@DavidCarlisle OH NO
 
David Carlisle
David Carlisle
12:39 PM
@UlrikeFischer I fear I'll be coming back to you re font.each and test2e.tex at some point (I can't get it to work in the test suite setting)
@PauloCereda and it's scientifically proven that you won't forget now.
 
Paulo Cereda
Paulo Cereda
@DavidCarlisle OH NO
 
daleif
daleif
@DavidCarlisle I do not remember either, but I think it was related to labbook and that it no longer worked with the updated kernel. Since I wrote him on 4 May, it is probably this one: tex.stackexchange.com/q/429955/3929
 
David Carlisle
David Carlisle
@daleif thanks looks like the \@ifundefined change. I suppose I'll have to ask him if I can get update rights to the package on ctan....
 
daleif
daleif
@DavidCarlisle isn't it more related to §?
 
David Carlisle
David Carlisle
@daleif yes it's the utf-8 change not the \@ifundefined, I was just reading that post now thanks
 
Ulrike Fischer
Ulrike Fischer
12:54 PM
@DavidCarlisle Sure. Can I get the problematic setting from github and try out here?
 
David Carlisle
David Carlisle
1:07 PM
@UlrikeFischer yes I can mail you later or do you have a copy of latex2e checked out?
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle I have a copy. Is in master?
 
David Carlisle
David Carlisle
@UlrikeFischer yes to make it quicker in latex2e/base then l3build check github-0049 used to pass and now it doesn't. support/fonttext.cfg undefines \Umathchar while the format is made so the lualatex used for testing has preloaded cmr10 and defaults to OT1 encoding, but then in test2e with or without that font callback and with or without require(luaotfload-main) I get a difference in reported font names.
I suspect i was getting tired and doing something silly last night, nothing seemed to make any difference. I'll try again at the weekend...
@JosephWright ^
 
 
2 hours later…
Skillmon
Skillmon
2:59 PM
@marmot seems like this is the issue. I wonder why one hard coded the .center into it though. Thanks for looking into it.
@DavidCarlisle I second this.
 
David Carlisle
David Carlisle
@Skillmon don't let @PauloCereda know
 
Paulo Cereda
Paulo Cereda
@DavidCarlisle oh no
 
Skillmon
Skillmon
@DavidCarlisle I'd never eat a duck which can code in TeX.
 
Paulo Cereda
Paulo Cereda
user image
3
 
Skillmon
Skillmon
@PauloCereda intel Pentium? How old is this machine? (or is it one of the new pentium processors)
 
Paulo Cereda
Paulo Cereda
3:06 PM
@Skillmon :)
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle you got mail ...
 
David Carlisle
David Carlisle
3:30 PM
@UlrikeFischer ah good job you told me, gmail had spam filtered it:-)
 
Paulo Cereda
Paulo Cereda
@DavidCarlisle ooh spam
 
Ulrike Fischer
Ulrike Fischer
3:56 PM
@DavidCarlisle new spam ...
 
Marijn
Marijn
4:07 PM
Maybe we should suggest to @thevesh-theva to re-ask tex.stackexchange.com/questions/453940/… at Academia.SE or The Workplace?
 
Paulo Cereda
Paulo Cereda
@Marijn Hi!
@Marijn arara -L nl features your translation. <3
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle and now that I know where to look: your font.each() no longer works as the new "safe" font.each in the new fontloader only wades throught the fonts loader with the fontloader. So the following gives only one hit, while previously you got over 50. @MarcelKrüger Do you have an idea for a better font.each()? 
\font\y =cmr10
\input luaotfload.sty
\font\z=cmr10

\directlua{for ii,vv in font.each() do texio.write_nl(ii) end}
blub

\bye
 
marmot
marmot
4:33 PM
@Skillmon Well, yess, that's a debatable choice. Perhaps the author was under the impression that many users do not know what an anchor is, so it is better to add one. Anyway, the way you use it is not too complicated. You could of course move the pic up into the line \draw[dashed,darkgray] (M.south) -- ++(2,0) coordinate (H); such that you do not have to type M.south twice, but then a significant number of users may no longer understand what's going on.
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle actually I just remembered that @MarcelKrüger suggested another definition. If I add the following to test2e and then use font.latexeach instead of font.each the test pass. 
  local function fonteach_next(max, f)
      repeat
        f = f + 1
        if f > max then return end
      until font.frozen(f) \string~= nil
      return f, font.getfont(f) or font.fonts[f]
    end
    font.latexeach = function() return fonteach_next, font.max(), 0 end
 
Skillmon
Skillmon
5:04 PM
@marmot I think often it is better to write not too dense code. A colleague of mine has the habit of writing really dense code and it is hard working with him/his code therefore (C that is though). After experiencing this myself I turned to write looser code for others sake.
 
David Carlisle
David Carlisle
@UlrikeFischer ooh thanks, let me try...
 
marmot
marmot
@Skillmon I fully agree. (You may nevertheless pull the coordinate (Ms) in the \draw above.) I find that one of the toughest challenges when writing answers here is precisely this: if the code is too dense, the OP may not understand it. If it is very easily understandable, sure enough there will be someone coming and tell you that you have missed that this can be written so much shorter (and make the code more "concise" ;-).
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle the first 200 tests of the suite worked for me already ;-).
 
Skillmon
Skillmon
@marmot Though I understand the advantage this might give I'll not do it. The effort doesn't seem worth it in this case, it is just one line less and doesn't improve the answer considerably.
 
marmot
marmot
@Skillmon I agree. (But don't complain if later someone comes, changes just this, claims his code is much more concise and gets the tick. OK, in this case it won't happen because @manooooh cannot be tricked that easily. In other cases, it does happen, and sometimes it is rather mean when you have gone through 100 iterations with the OP to find out what the question was etc...)
 
Marijn
Marijn
5:22 PM
@PauloCereda good to know the translation is in place :) if you need any additional messages translated, just let me know!
 
David Carlisle
David Carlisle
@UlrikeFischer @MarcelKrüger 33 tests passed so far:-) Thanks for the help!
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle so slow I have 320 ;-).
 
David Carlisle
David Carlisle
@UlrikeFischer I had to re-arrange few things to get to a state that might pass first:-)
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle you could name the function marceleach instead of latexeach :-)
@DavidCarlisle are you now loading luaotfload?
 
David Carlisle
David Carlisle
@UlrikeFischer I was thinking fonts.ifthisgoeswrongblameulrike
@UlrikeFischer er... I guess not I just made the changes that you suggested above, should I back out the fonttext.cfg weirdness with Umathchar? and.or load luaotfload-main directly in test2e?
 
Ulrike Fischer
Ulrike Fischer
5:30 PM
@DavidCarlisle fonts.ifthisgoeswrongblameulrikeandmarcel -- that's a joint work - he wrote the function and I suggested it.
 
David Carlisle
David Carlisle
@JosephWright test suite is looking better ^^
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle I'm not sure that I really understand your setup. I just managed to open one real log of lualatex and this loads the fontloader already. Are you really suppressing it completly?
 
David Carlisle
David Carlisle
@UlrikeFischer actually I'll check (It went through so may iterations) maybe I just prevent the TU codepath so the format defaults to OT1 but leave the everyjob loading, I'll look again, sorry if I mislead you
@UlrikeFischer well as you say it's certainly loading luaotfload now:-)
@UlrikeFischer it's a good job someone understands this stuff:-)
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle ;-). Another question: can you close issues at lualatex (from other people)? I seem not be able to do it.
 
David Carlisle
David Carlisle
5:46 PM
@UlrikeFischer I can try, which one?
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle e.g. this one github.com/lualatex/luaotfload/issues/419
 
David Carlisle
David Carlisle
@UlrikeFischer let's see, do you think I should also add a note to the luaotfload README saying development is at your fork?
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle Hm. Yes would be probably a good idea.
 
David Carlisle
David Carlisle
@UlrikeFischer on the issue I only see a "comment" button not close. I suppose we could go back to Elie and ask to bump out access rights in the lualatex github
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle ok, I will write. The test suite failed on tu-tl2e7 but the diff says that the new run hasn't the factor error ;-)
 
David Carlisle
David Carlisle
5:55 PM
@UlrikeFischer I can't edit the luotfload README either
@UlrikeFischer is that the one with kern differences in lmodern?
@UlrikeFischer I was going to ask about that, has the font changed or is the new font loader injecting some missing inter-letter kerns?
@UlrikeFischer I'm still at 252/358
 
Ulrike Fischer
Ulrike Fischer
6:22 PM
@DavidCarlisle difficult to say definitively but as e.g. there is now an expected kern between T and e I would say the tlg-output was simply wrong due to the font.each() problem which affected the font parameters.
 
David Carlisle
David Carlisle
@UlrikeFischer ah.
 
David Carlisle
David Carlisle
6:39 PM
@UlrikeFischer pushed to Travis, let's see what happens there:-)
 
samcarter
samcarter
7:16 PM
Thanks! Notice that visible on etc. is part of the very useful library overlay-beamer-styles, whose only disadvantage is that nobody can recall its name. ;-) — marmot 2 hours ago
@marmot I love this description of the aobs package!
 
marmot
marmot
@samcarter Yes. Maybe Claudio wants to rename it to "the-one-you-know-who"? ;-)
@samcarter It actually took me a while till I understood why the soul package is called soul, but after I understood I love the name. Most likely this process of understanding just takes a bit longer in the case of overlay-beamer-styles... ;-)
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle Failed on amsmath. But that's not my fault. xetex fails too!
 
samcarter
samcarter
@marmot Good suggestion for a new name! If he doesn't want to rename the package, maybe texdoc could use it as alias :)
@marmot The name of the soul package is great. It is short and an already existing word. Both make it very easy to remember.
 
marmot
marmot
7:35 PM
@samcarter Unfortunately, I do not know how to complete beamer overlay duck ... to "body", do you have a suggestion for y? ;-)
 
David Carlisle
David Carlisle
@UlrikeFischer a bit odd actually it's picking up TU encoded roman not OT1 everywhere (so then box differences all over) eg 
! ....\OT1/cmr/m/n/10 C
! ....\OT1/cmr/m/n/10 D

! ....\TU/lmr/m/n/10 CD
as you say it's xetex as well, but something has changed....
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle did you remove the fonttext.cfg?
 
Skillmon
Skillmon
@marmot how about beamer overlay duck, yay!
3
 
David Carlisle
David Carlisle
@UlrikeFischer oops I probably did (or at least I was moving it around locally, git is rather too keen to delete stuff, let me check...)
@UlrikeFischer no it's still there as it was
 
marmot
marmot
@Skillmon Now you only need to convince Claudio.... ;-)
 
Ulrike Fischer
Ulrike Fischer
7:45 PM
@DavidCarlisle well actually you can't, as I got the same error too and I haven't pulled anew. So whatever changed changed already earlier.
 
Skillmon
Skillmon
@marmot I think first we'd have to smuggle some ducky content into it to justify the duck in the acronym.
 
David Carlisle
David Carlisle
@UlrikeFischer travis history shows it passing 4 days ago so can't be that hard to track down
 
marmot
marmot
@Skillmon I guess if you really would start justifying ducks, you will immediately be in trouble.
 
Skillmon
Skillmon
@marmot What does the a stand for in the current package name? Please don't say it's supposed to mean auxiliary.
 
marmot
marmot
@Skillmon It stands for "arbitrary". ;-)
 
David Carlisle
David Carlisle
7:58 PM
@UlrikeFischer I wonder if we can blame @JosephWright. after running check in amsfonts latex2e/build/test has a "clean" copy of fontext.cfg as generated from base, not the one from latex2e/support that forces OT1 as default encoding, so the text outside math eg the CD before the \begin{CD} test is in TU/lm not OT1/cm
 
Ulrike Fischer
Ulrike Fischer
@JosephWright can you make this a community question? tex.stackexchange.com/q/454031/2388
@DavidCarlisle well there was a l3build update yesterday I think ;-).
@DavidCarlisle I can now close issues ... ;-)
 
David Carlisle
David Carlisle
@UlrikeFischer me too by look of it:-)
 
Joseph Wright
Joseph Wright
8:38 PM
@UlrikeFischer Done
 
David Carlisle
David Carlisle
@UlrikeFischer I get safer option set, function "io.open" is limited as you show for that lualatex -safer test using the luaotfload in texlive, did you want me to test the one from github?
@JosephWright is the fonttext.cfg issue in amsmath a l3build change? I lost the trail a bit?
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle the change I made is in the "issue12" branch github.com/u-fischer/luaotfload/tree/issue12 (changed are only luaotfload-main.lua and the sty (for plain loading).
 
David Carlisle
David Carlisle
@UlrikeFischer ah OK let me sync my local copy....
 
Joseph Wright
Joseph Wright
@DavidCarlisle I'll take a look: nothing should have changed ...
@DavidCarlisle On the train at the moment: may be tomorrow before I can look carefully
@DavidCarlisle I should be checking in amsmath, right?
 
David Carlisle
David Carlisle
@JosephWright no rush I thought it was going to b etomorrow before any pf this worked but @UlrikeFischer fixed all teh failures in base so things moved on:-)
@JosephWright yes, the non math text is being set in TU/lm instead of OT1/cm for xetex and luatex, it appears that the build area has an unpacked version of fonttext.cfg from base rather than than the one from latex2e/support that re-instates OT1 to cut down on diffs. But travis said it all passed four days ago
 
Joseph Wright
Joseph Wright
8:45 PM
@UlrikeFischer Could you up me to 'owner' too? Seems sensible ...
@DavidCarlisle OK, I'll do a bisect on l3build to check
 
Ulrike Fischer
Ulrike Fischer
@JosephWright Oh I could do it ;-)
 
David Carlisle
David Carlisle
8:59 PM
@UlrikeFischer would you expect something differentto happen with teh issue2 branch? I get: 
$ kpsewhich  luaotfload-main.lua
/home/luaotfload/texmf/tex/luatex/luaotfload/luaotfload-main.lua

davidc@dc-bantham /home/luaotfload
$ git branch
  dev-2.9
* issue12
  master

davidc@dc-bantham /home/luaotfload
$ lualatex --safer sample2e
This is LuaTeX, Version 1.07.0 (TeX Live 2018)
 restricted system commands enabled.
(/home/davidc/texmf/tex/latex/base/sample2e.tex
LaTeX2e <2018-12-01> pre-release-6

safer option set, function "io.open" is limited
 
Ulrike Fischer
Ulrike Fischer
@DavidCarlisle With 
\documentclass{article}
\begin{document}
a
\end{document}
and option --safer you should see something like 
LaTeX2e <2018-04-01> patch level 5
luaotfload can't run with option --safer. Aborting
...aotfload/texmf/tex/luatex/luaotfload/luaotfload-main.lua:47: safer_option use
d
Error in luaotfload: reverting to OT1
 
David Carlisle
David Carlisle
oh hang on, git error I think
 
boycott.se - yo'
boycott.se - yo'
@DavidCarlisle git error? impossible, git is made to perfection :D
 
David Carlisle
David Carlisle
@boycott.se-yo' git user error
@UlrikeFischer vvv 
This is LuaTeX, Version 1.07.0 (TeX Live 2018)
 restricted system commands enabled.
(/home/davidc/texmf/tex/latex/base/sample2e.tex
LaTeX2e <2018-12-01> pre-release-6
luaotfload can't run with option --safer. Aborting
...aotfload/texmf/tex/luatex/luaotfload/luaotfload-main.lua:47: safer_option us
ed
Error in luaotfload: reverting to OT1
@UlrikeFischer @JosephWright I wonder if that latex format Error in luaotfload: reverting to OT1 ought to be an actual error not a message so it doesn't just scroll past....
 
Ulrike Fischer
Ulrike Fischer
9:25 PM
@DavidCarlisle I thought about this too, but imho there can be valid setups without luaotfload. For testing it is actually quite neat ;-) But one could add more !!!! and so to make the message more visible.
 
boycott.se - yo'
boycott.se - yo'
9:39 PM
@DavidCarlisle ooh git user feature :-)
 
 
2 hours later…
vlg
vlg
11:38 PM
Reading the pythontex package doc I encounter the following: " Additional notes: matplotlib added a pgf backend in version 1.2. You
will probably want to use this for creating most plots. However, this
is not currently configured automatically because many users will want
to customize font, T E X engine, and other settings. Using T E X to create
plots also introduces a performance penalty." - somebody here was chastizing me for using tikz, pgf and windoos. Now you this says that there be a perf penalty for using tex?
Should I use the pgf backend to create .pgf or .pdf figures then embed these, or do it from within the document, what is recommended?
Or use ms paint?
 
Alan Munn
Alan Munn
@vlg Sure, because TiKZ, although powerful, is pretty slow at many things. But that doesn't mean you shouldn't use it if it's useful for your workflow. (That's what they mean about "using TeX" in this context.)
 
marmot
marmot
@vlg In addition to what @AlanMunn is saying, there is of course the option of externalizing these plots, so you only compile them once (unless you change them, in which case you need to recompile).
 
Alan Munn
Alan Munn
@vlg If you are doing complex data analysis and plotting, then pgfplots may not be the best solution for you, and pythontex allows you to offload that work to something more suited for it, like one of the many data visualization libraries available for python (or R or Julia, etc).
 

« first day (2896 days earlier)      last day (2029 days later) »