« first day (3767 days earlier)      last day (1164 days later) » 

8:21 AM
@HaraldHanche-Olsen oh my!
quack
 
8:36 AM
@DavidCarlisle and @egreg Kannappan sends his best regards!
 
9:08 AM
@PauloCereda Best regards from me!
 
@DavidCarlisle should we say "don't use curious dots over the letters in your titles"?
 
@UlrikeFischer yes
 
@DavidCarlisle ööh
 
9:41 AM
@PauloCereda that doesn't count
 
@DavidCarlisle oh
 
@PauloCereda We've grown up knowing about Ööç
 
@egreg ooh galaxies :)
 
@DavidCarlisle ooh
@DavidCarlisle EPOCH
 
10:28 AM
Good morning, could I ask you for your feedback, please? I have here a minimum example for a CV. The content for the title is two lines long. But it seems this command does not keep the distance between the two lines. Or do I see it wrongly, please? Could you have a look, please?
My example is
\documentclass[11pt,a4paper,sans]{moderncv}
\usepackage{moderncvcompatibility}
\usepackage{moderncvstyleclassic}

\usepackage[left=1.7cm, right=1.7cm, top=1cm, bottom=1cm]{geometry}


\firstname{Max}
\familyname{Mustermann}

\title{Chemist\protect\\[0.2\baselineskip]
Inorganic Chemistry, Organic Chemistry, Software development, Instrumentation, R\&D}

\address{Musterhausen, Musterland}{}{}
\email{max.muster@gmail.com}
\phone{+1 123456789}

\begin{document}

\makecvtitle

\section{Work Experience}
 
 
2 hours later…
12:02 PM
@PauloCereda github.com/latex3
2
 
@JosephWright Thank you, looks way better now :)
 
12:14 PM
@PauloCereda so if on a new machine I installed `open jdk version 15.0.2 2021-01-19 and the old machine has java version "1.8.0_131" / Java(TM) SE Runtime Environment (build 1.8.0_131-b11) what's the difference?:-) (I know I switched to openjdk but how does the numbering work these days?
installing something called "texlive" ....
 
@DavidCarlisle Remove the 1.* prefix and start counting again. :)
@DavidCarlisle Java 9 fixes this. :)
 
12:29 PM
@PauloCereda er so 9 is bigger than 15.0?
 
@DavidCarlisle Hm? No, 9 < 15. :)
@DavidCarlisle 1.8 is Java 8, then 9 is Java 9, ... 15 is Java 15. :)
 
@PauloCereda ah so I think what I thought was right that 1.8 == 8 and then I seemed to have jumped a few releases to 15?
 
@DavidCarlisle Some distros features version 11 as well, but you are right. :)
 
version numbering is difficult
 
@DavidCarlisle but java version numbering is probably undecidable – or at least NP-hard
 
12:32 PM
@DavidCarlisle ? ah the java question
 
@UlrikeFischer yes obviously java because version numbers for latex, or pdf or windows or tex are all so obviously monotonic
 
$ sdk list java
================================================================================
Available Java Versions
================================================================================
 Vendor        | Use | Version      | Dist    | Status     | Identifier
--------------------------------------------------------------------------------
 AdoptOpenJDK  |     | 15.0.2.j9    | adpt    |            | 15.0.2.j9-adpt
               | >>> | 15.0.2.hs    | adpt    | installed  | 15.0.2.hs-adpt
@DavidCarlisle ^^
 
@DavidCarlisle there is someting in the idea to use arbitrary town names, then at least everyone knows it has no meaning and you should better look up the order.
 
texlive install got as far as jadetex, so I have all the important stuff
 
@DavidCarlisle ooh
 
12:39 PM
@PauloCereda claims it'll take 55 minutes for a clean install, we'll see...
 
@DavidCarlisle it's a trap ooh
 
@DavidCarlisle oh did you upload it? I didn't see in the updates.
 
@UlrikeFischer not yet, but I am installing tex on a new machine and it just happened to be passing j... at that point
 
yo'
@DavidCarlisle xii is missing though then...
 
12:41 PM
@DavidCarlisle ah ;-).
 
@UlrikeFischer luahbtex now, do I need that?
 
@DavidCarlisle if you want to make the next hyperref upload.
 
@UlrikeFischer lt 4.14 next...
 
quack
 
12:57 PM
@UlrikeFischer I'm thinking about ICC conversion: at the moment I'm minded just to have a fixed mapping based on number of components (to std gray/rgb/cmyk) - if we need more control later, we can add it
@DavidCarlisle, @UlrikeFischer I see SVG mentions CIELAB: guess somehow I need to add support
 
1:16 PM
@JosephWright for me the iccbased init code is not executed. And this here looks rather odd:
\cs_new_protected:Npn \@@_model_iccbased_aux:nnn #1#2#3
  {
    \@@_model_init:nnn {#2} {#1} { iccbased }
    \use:c { @@_model_devicen_parse_ #1 :nn } {#2} {#1}
    \exp_args:Nx \@@_backend_devicen_init:nnn
      { \file_full_name:n {#3} } {#1} {#2}
  }
why devicen here?
I think it should be \@@_backend_iccbased_init:nnn
 
@UlrikeFischer Ah, yes, so it should
@UlrikeFischer The \use:c line is right
@UlrikeFischer Until we integerate PDF management, I am a bit limited testing here :)
 
@JosephWright ;-). Something is wrong with the color space name. When I do this
\color_model_new:nnn{icctest}{ICCbased}{file=sRGB.icc,components=3}
\color_set:nnn{blub}{icctest}{1,0.3,0.2}
then the colorspace dictionary says << /icctest 6 0 R >>. But in the stream you use /color1 CS 1.0 0.3 0.2 SCN.
 
 
1 hour later…
yo'
2:35 PM
note to self: it's not TeXas
4
 
 
1 hour later…
3:40 PM
1
Q: How do I use such multiple character ligature signs for Musical Notation?

Always ConfusedI found a portion of music Ebook The Basics of North Indian Classical Music by Abha Saxena (Google books) (Google play bookstore) (free preview), and this is so far the only ebook I have seen that has successfully implemented the North Indian classical music notation in a form which is selectable...

 
3:53 PM
Does anybody know if that's the proper way in LaTeX? I suspect something may be rotten here:
\documentclass{article}
\makeatletter
\newcommand\fibonacci[2][0]%
	{\ifx#10\@dofibo{#2}\else\@dofibos{#1}{#2}\fi}
\def\@dofibo#1{\@dodofibo{1}{0}{#1}}
\def\@dodofibo#1#2#3%
	{\ifnum#3=0 #2%
	\else
	\@dodofibo{\the\numexpr#2+#1\relax}
	{#1}
	{\the\numexpr#3-1\relax}%
	\fi}
\def\@dofibos#1#2%
	{\ifnum#1=#2 \@dofibo{#1}\else%
	\@dofibo{#1}, \@dofibos{\the\numexpr#1+1\relax}{#2}%
	\fi}
\makeatother
\begin{document}
\fibonacci{10}
\fibonacci[8]{20}
\end{document}
 
@JairoA.delRio \ifx#10 probably does wrong thing if #1 is 11
2
 
@DavidCarlisle Oh, so, should I use an \ifnum check instead?
 
@JairoA.delRio probably:-)
 
@DavidCarlisle I'd use Lua instead, but it seems not many people here love LuaTeX :)
 
4:05 PM
@DavidCarlisle Btw I realized the OP is not using e-TeX extensions. I give up :D
 
@PauloCereda Can't they just hide their hair under the black cloth-thingy?
 
@samcarter_waiting_for_summer ooh
 
@JairoA.delRio yes the code in the question is e-tex (I posted an answer)
@JairoA.delRio (s)he is using \unless and \numexpr so definitely this requires etex
 
@DavidCarlisle shklee
 
@JairoA.delRio What's wrong with eTeX?
 
4:16 PM
@JosephWright it's not ConTeXt?
 
@JosephWright I was suggesting the opposite: doing calculations without e-TeX is :) I wanted to answer a LaTeX question, but @DavidCarlisle was faster.
@PauloCereda lol
 
I don't know if it was shared in chat before but finl.xyz is an interesting write-up about what some people think could solve problems LaTeX has…
3
 
@TeXnician ooh
 
@TeXnician I think some of the points discussed are already solved with LuaTeX, right?
 
@JairoA.delRio Many of them, yes. And abandoning active characters etc. does not seem like the best deal to make. Apart from writing this alternative in a memory-unsafe language :D
@JairoA.delRio Ah, and notice the points about exchanging the document class and keeping the markup. ConTeXt solves that part in one way ;)
 
4:32 PM
@TeXnician But, for many TeX users, it's like ConTeXt doesn't even exist as an alternative. Interesting attitude :)
 
@JairoA.delRio ConTeXt is not an alternative, it's the only way to solve problems ooh a conundrum
 
@PauloCereda Jokes aside, it's a matter of considering options. I remember having used LuaLaTeX because ConTeXt doesn't support Harfbuzz (and Hans is very reluctant to include external libraries into the engine). If something works, use it. If it doesn't, fix it or change it
 
@JairoA.delRio we could also break existing stuff too. :)
 
@PauloCereda That's the best part. <3
 
@JairoA.delRio quack :)
 
4:39 PM
@PauloCereda Btw I think both LaTeX and ConTeXt concide here :D
 
@JairoA.delRio ooh
 
@TeXnician ah that's why Don H has re-appeared after all this time:-)
 
@TeXnician Some of this is ... quite often seen but never leads anywhere
 
@JosephWright this sounds suspiciously familiar PR PR PR lovely PR wonderful PR
 
4:42 PM
@TeXnician The 7-bit business is quite funny, when LaTeX moved to default UTF-8 a few years ago and when that's the default for XeTeX and LuaTeX since forever
2
 
> Unicode needs to be a first-class citizen. There’s no reason in 2020 for a document writer to have to type \’a instead of á in a document. UTF-8 is the new 7-bit ASCII.
ConTeXt already did it, first partially (bibliographies) and now totally in LMTX
 
@TeXnician Also on the LaTeX team agenda, and also hard
@JairoA.delRio Yes, a decision I really don't follow: there's a big difference between accepting UTF-8 as the standard input and Dropping useful ASCII input approaches: for the odd non-English word, for example, I'd much rather use backlash-accents than have to trawl around some character map GUI
 
> So what can be done? I believe a complete rethink of LaTeX is necessary. The macro-based language of LaTeX is intellectually satisfying in many of the same ways that solving a sudoku puzzle or playing chess can be. It is not, however, a straightforward programming model and does not conform to any popular paradigm (I don’t think any other programming language is of similar structure).
LuaTeX partially solves this, doesn't it?
 
yo'
How can you search for packages with hyphens in titles in CTAN?
 
@JosephWright Yeah, exactly. And I have the feeling, his thingy will go the SILE way…
@JairoA.delRio The discussion on the list showed that many consider those shorthands useful, though.
 
4:52 PM
@JosephWright Oh, matter of taste. Hans is moving those shortcuts to a module so they are optional, but they certainly won't be the default
@TeXnician Yep, I noticed that.
 
@JairoA.delRio Yes, although really ConTeXt shows what it means because you have a high-level programming concept (e.g. tables) and do not need to manipulate tokens or insert something into the input stream.
 
@JairoA.delRio This has come up time and again: it's of course true that one can write a document processor in lots of language, and there are lots of those. The challenge isn't in that sense technical, it's about getting a critical mass. Plus I still wonder about the edge cases: the odd place you need to do some 'programming' in the middle of a document
@TeXnician But it still uses macros, even with LTMX
@JairoA.delRio Like I say, this is what I don't get, but I suppose partly as to me this is a clear place it's breaking back-compatibly for ... what exactly?
 
Of course. But it makes programming them easier. Lua interfaces to TeX environments and commands is something LuaLaTeX is really missing. Generating tables is one of those things where using some string magic to insert the right amount of `&` is far inferior to something like `context.startxcell() context("content here") context.stopxcell()` in some loops with rows etc.
It is an alternative to what you actually want when programming many easy parts in TeX, simple repetitive loops and such. Of course, macro expansion is really useful, so it would be a shame to abandon it.
 
@JairoA.delRio I think it's worth bearing in mind that the LaTeX Project plan was for a long time to re-write a new format, but the reason we've moved away from that is that we'd get (to a first approximation) no users
@TeXnician On tables specifically, I've not looked a the detail in ConTeXt: is Hans doing everything himself now or still using engine code at all (\halign or some Lua interface to it)?
 
@JosephWright Oh, I see and it makes more sense now. Certainly ConTeXt has been more audacious in that sense.
 
4:58 PM
@TeXnician Programmatically generating tables is 'fun' but really misses that most of the time a good table for publication is short, focussed, and should stay far away from any auto-code!
 
@JosephWright I don't know the details. But I know that I love xtables, not only when using ConTeXt to typeset XML data but when generating things from databases as well.
 
@JairoA.delRio Not really: Hans has to a first approximation no users anyway :)
 
@JosephWright lmao
@JosephWright It depends. AFAIK natural tables (\bTABLE ... \eTABLE) rely on MetaPost for color, backgrounds and some other complex details
 
@JosephWright I disagree. Typesetting catalogs and such needs automated typesetting. Getting things from a database (SQL, …) and then doing things in the Lua end is great. Simple example: Here in Germany, module descriptions for university courses tend to vary in length and attributes (depending on the type of module). When trying to typeset these things, doing "null"-checks in Lua and typesetting conditionally in the language where you retrieved the data, it is very important to have automation.
 
@JairoA.delRio Well that partly comes down to a question of how much you encourage third-party code. There are external modules for ConTeXt, but very few compared to the scope of the core. Fine if they work for you, but ... more fun if they don't. For example, I'm not sure if TikZ is working properly on LuaMetaTeX (quite apart from all the specialist stuff one sees for academic subject areas)
 
5:02 PM
@JosephWright Unfortunately, TikZ currently has some rough edges, although it still works. pgfplots is more interesting, last time I checked :D
 
@TeXnician The usual question of what one thinks of as a 'typical' use case: for me, TeX is mainly about theses, academic papers, etc., certainly in terms of numbers of end users, and in almost all cases long(er) tables there are data dumps rather then really useful acts of typesetting
 
@JosephWright Shading is broken :D Henri has informed to me LMTX is currently unsupported, so to avoid nuisances one should go back to MkIV
 
@TeXnician I'm not saying it's not easier in Lua, I'm more saying that this is a specialised area in the total set of .tex files in the world
@JairoA.delRio Ah, I've not followed all of the detail
@JairoA.delRio I've got to say I've never understood the MetaPost business: TikZ/pgf shows one can happily sort stuff in TeX + backend specials
 
@JosephWright Yes, of course. But a few years ago, I used TeX to typeset a catalog of abstracts of student theses. These were data dumps, insofar that they dumped a database into a document. But really, having TeX for that was great to automatically ensure that it is in a proper format for double-sided printing, have the abstracts be nicely typeset etc. There are cases where the best of both worlds is just what would be nice ;)
 
@JosephWright I don't use TikZ and I don't like it, but some days ago an user asked why didn't her/his example work. I provided a workaround with MetaPost, but that's not a real answer. The why is: LuaMetaTeX has dropped primitives necessary for TikZ in LuaTeX
 
5:07 PM
@TeXnician Sure, I've done similar: I suppose I'm saying that when you get to doing that, you are a programmer not a document author, and that should be a small subset of TeX users (the point of both ConTeXt and LaTeX; plain TeX users are naturally programmers)
 
@JosephWright Performance is in a different order of magnitude :D
 
@JairoA.delRio I know Hans has dropped loads of stuff; I'm not sure why you don't like abstratcting out PDF/PostScript/SVG stuff, but that's your call :)
 
@JosephWright well, it's a rather thin line. At the end of the day, each and every TeX user is a programmer, even if unbeknownst to them...
 
@JosephWright TikZ wasn't born when ConTeXt appeared for the first time
 
@JairoA.delRio Sure, but at that time I'd have expected people to be sticking strictly to PostScript graphics
 
5:09 PM
@JosephWright Yes, of course. Nevertheless, I see that if we already got LuaLaTeX, Lua interfaces to the available constructs would be a logical step in the future.
 
@JosephWright Also, as @TeXnician says, it's faster, That's noticeable in larger documents
 
@JairoA.delRio That I can believe; I guess we are back with the question about 'typical' users (for most end users, performance is not really an issue in that sense, as we are talking at worse a few minutes per run of a file that's likely the only one they are processing)
 
@JosephWright Oh, there are cases when performance makes a difference. I was doing some stuff with auto-generated mazes for kids last year using a Lua library. Drawing with TikZ took hours. With MetaPost and ConTeXt it takes minutes. So I dropped TikZ. But I'm aware that's not a typical case
 
@JosephWright I'd say the difference is really noticeable. I regularly typeset documents with a large number of sketches. The only point that makes debugging these things not too painful in terms of typesetting delay is TikZ' externalization library (editing something in external standalone files is a bit clumsy in documents with large preamble). With ConTeXt and MetaPost, I never had problems with delays.
 
@TeXnician we really should provide a proper metapost interface in lualatex, context itself wouldn't have been the speed difference I assume?
 
5:17 PM
 
@DavidCarlisle No, the speed difference of ConTeXt and LuaLaTeX is not that large. But TikZ vs. Metapost in this case is.
 
@JosephWright To be honest, Postscript looks cryptical to me.
@JosephWright MetaPost is also more intuitive imho. It evokes the feeling of using a rule and a compass when you have enough time drawing stuff
 
@DavidCarlisle And I know there is luamplib but it's GPLv2 (iirc) and misses some of the neat ConTeXt shorthands (unique MP graphics etc.).
 
@JairoA.delRio latex can solve those easily:
 
@JairoA.delRio That's probably a question of background. I know some people who do not like to think in linear equations and prefer TikZ (have worked for projects requiring MetaPost, so there were complaints about usability compared to TikZ) :D
 
5:22 PM
@TeXnician Linear equations are no-brainers. I'd worry for those who complain :)
 
@JairoA.delRio In maths, they are. But there are fields (humanities) where you should not make such assumptions :D
 
@TeXnician luamplib is GPLv2 since it uses ConTeXt code. That could be avoided (e.g. I have some independent LuaLaTeX mplib backend stored ... somewhere) but then there is no access to MetaFun etc., so we would have to build up a completely independent ecosystem for LaTeX mplib which probably wouldn't happen with TikZ as a more established alternative.
 
@MarcelKrüger Ah, right, good point. So a necessary evil then.
 
Since we are talking about GPL and ConTeXt: Does anyone have any news about the luametatex sources?
 
@MarcelKrüger Are they available? I thought they were to be released this year or the next one, according to what Hans said in reports/manuals
 
5:33 PM
@JairoA.delRio That's what I was trying to ask.
 
@MarcelKrüger Ah, no news in that case. In the last Luametafun manual it's reported that some MetaPost code has been removed, but nothing else worth mentioning.
 
GPL is mean
 
Sorry for being a killjoy, but what's wrong with GPL? Genuinely asking
 
@JairoA.delRio Oh, it is: I did say abstracting from it
@JairoA.delRio I've never used it, mainly as it's not automatically cross-engine I suppose
 
@JosephWright gmp works pretty well in LaTeX.
@JosephWright Ah, sorry :) MetaPost does too, but TikZ abstracts too much,
 
5:53 PM
@JairoA.delRio IMO: First it's simply a bad fit for TeX related stuff since it is unclear if documents are derived works. So maybe (or maybe not) documents must be published under GPL rules. (If not, where is the separation? Does it make a difference if I move code from my document text into a separate package?)
That's not intended by most package authors. (e.g. ConTeXt's documentation uses non GPL compatible licenses itself, so if it's considered a derived work then it's a license violation...)
 
@MarcelKrüger I think documents are clear in separation from the engine; I agree for macro stuff life is a lot less clear
 
@JosephWright Yes, but I was taking about format and package stuff. (We started with luamplib which is a LaTeX package)
 
@MarcelKrüger Oh, sure: I'd go for MIT or LPPL, depending on the history of the code/target collaborations (so for generic stuff probably MIT)
 
@MarcelKrüger Is there a better alternative then? LaTeX has a license, but I'm not sure the ConTeXt crowd is willing to ever consider it
 
@JairoA.delRio Related to that it's just a drama license. It's often used by people who don't want to deal with licensing (e.g. ConTeX's " . . and don’t bother discussing licence issues and related things with us for the mere sake of discussing licence stuff.") while on the same hand triggering more licensing issues that almost any other license I know.
3
 
5:59 PM
@MarcelKrüger I have to admit I find Hans and Taco's statements fun, but I don't really understand their consequences
 
@JairoA.delRio I use MIT or BSD outside of LaTeX, but IIRC Apache is a copyleft license which only applies Copyleft to the file itself instead of the entire project. That might be an option.
 
@JairoA.delRio MIT: with hindsight, that would have covered many things the LPPL does (as the latter started out wanting to do things that are not really best handled by a license)
 
@JairoA.delRio Replace Apache with Mozilla Public License in my previous comment.
 
@MarcelKrüger IoT too (BSD, MIT and I think we once had a Apache suggestion)
@MarcelKrüger ooh :)
@JosephWright Debian LPPL woes II: woe harder
@MarcelKrüger @Phelype, @samcarter and I encourage this one: github.com/supertunaman/cdl
4
 
Why are too many licenses? Before knowing any TeX, I only knew Creative Commons. In fact, I began to use this chat because I (still) don't understand LPPL
 
6:07 PM
@JairoA.delRio IMO that's true for many statements in mreadme.pdf, especially the part about distributors "choosing" the LPPL instead.
 
@JairoA.delRio Creative Commons is not (in the main) good as a choice for code licensing
 
@PauloCereda That's a good one, but I also have some code where I was tempted to use github.com/me-shaon/GLWTPL
3
 
@JosephWright What would be the eventual issues with TeX related stuff if I used CC?
 
@MarcelKrüger OMG YES <3
@MarcelKrüger Specially this line: You wrote some code but not quite proud of it, yet want to release it. :D
Mr. Squirrel submitted an update to the CircuiTikZ package.
@Rmano ^^ ooh
 
6:11 PM
@JairoA.delRio For packages, the LPPL is nice because although it's a free licence it has a renaming convention built in. So this means anyone can use your code, but they have to release a modified version under a different name. This makes everyone's lives easier.
 
@JairoA.delRio Probably none, as in the end license issues are only significant if there is a court case. However, Creative Commons themselves don't recommend using their stuff for code
@AlanMunn Er, not has to, more is encouraged to
 
Is a license applicable to itself?
 
@JosephWright So I guess strongly encouraged to.
 
> The author has absolutely no clue what the code in this project does.
It might just work or not, there is no third option.
Nice
 
@PauloCereda And you wonder why you end up as dinner.
2
 
6:13 PM
@AlanMunn oh no
/quacks in despair
 
@AlanMunn This is the whole 1600 emails to Debian legal business; like I said, with hindsight, MIT would probably have covered a lot of it (not that anything will every change now)
 
@JosephWright Wait, won't L4 be MIT-based?
 
@PauloCereda Yeah!
 
@JosephWright Having only browsed those emails, I thought Debian had caved, but I guess they didn't.
 
6:15 PM
@AlanMunn Older LPPL requires a name change, current one simply says changed versions should make this clear
 
@AlanMunn Licenses should be explained that way as a preamble
> The Bugs License is in the public domain, because nothing else seems suitable.
 
> By entering into use of the materials licensed under this license ("The Mephistopheles License"), you agree to forfeit your own soul rather than negotiate with the authors of the materials.
2
 
@PauloCereda ^^ Licenses don't apply to themselves :)
 
@JairoA.delRio I beg to differ: :) github.com/benlk/misc-licenses/blob/master/…
> The Schrödinger License is licensed under the terms of The Schrödinger License.
3
 
6:19 PM
@PauloCereda Oh, God
 
@JosephWright I see, so that was the eventual compromise. But CTAN won't accept duplicate names, so in practical terms there's a renaming convention which (partially?) satisfies the "make this clear" clause.
 
@PauloCereda We should write a duck license :D
 
@JairoA.delRio ooh
 
@PauloCereda lol
 
@AlanMunn Yes; that's what I meant about this being something solved not by the license but by other things. Renaming (source) files, picking package names, etc., is a social contract
 
6:22 PM
@JosephWright Which is why the MIT licence would have sufficed.
 
@AlanMunn Yes, exactly: @DavidCarlisle and I have chatted about this a few times
@AlanMunn At this stage, this is all academic: LPPL works for us
 
@JairoA.delRio -- ConTeXt for a very long time was a moving target. For a production environment (a publisher of books and journals, for example), it wasn't an option. And once such an environment is stabilized, and authors are reliably submitting files that meet the stated requirements, it's awfully difficult (and even more expensive) to change.
 
@barbarabeeton Is ConTeXt stable now?
 
@barbarabeeton And still is. That's why it's recommended to us (ConTeXt users) that we make a backup of our current distribution if an update changes something too radically
@FaheemMitha ConTeXt MkIV is going to be, sorta, as LMTX is being developed.
@barbarabeeton I see LaTeX is safer here, even when I use and prefer ConTeXt
 
@JairoA.delRio I know there is a lot of churn, some of it Lua-related to ConTeXt, but shouldn't the user interface (or whatever you call it) be stable?
 
6:35 PM
@FaheemMitha It is. Only more low level or obscure complexities are unstable. To avoid problems, stick to higher level commands.
 
@JairoA.delRio I'm not a ConTeXt user. Though some day I might be.
 
@FaheemMitha Lua interfaces are rather stable. I didn't notice anything breaking, as it happens more on the TeX and to a lesser extent in the MetaPost side
 
@JairoA.delRio OK.
End user usage of Lua is still not common for LaTeX users. Is ConTeXt any different?
 
@FaheemMitha You could use ConTeXt knowing (nearly) nothing about Lua. You'd use Lua for graphs, automatic tasks or calculations
 
6:45 PM
@FaheemMitha However, some fancy things only can be done via Lua (notice this is posible in LuaLaTeX, too): tex.stackexchange.com/questions/582551/…
 
@JairoA.delRio I was just asking whether end user usage of Lua was common in ConTeXt or not. I'm aware you can use it without knowing Lua, since it's mostly on a the backend.
BTW, what are my options, if any, for profiling LuaLaTeX?
The compilation hangs at times. I don't know why.
 
@FaheemMitha l3benchmark is one (LaTeX-specific but can be used with a little effort with other formats)
 
@JosephWright I need to use it for LaTeX. but how does it cope with LuaTeX?
 
@FaheemMitha Of course
 
I need to try to pin down where the compilation is spending its time. Profiling usually involves a list of functions, and how much time (possibly relative time) is spent in each.
 
7:04 PM
@FaheemMitha this does not sound much of a TeX bottleneck, does it?
 
@PauloCereda I'm not sure what you mean
 
@FaheemMitha you know what a bottleneck is?
 
@PauloCereda Yes.
 
@FaheemMitha good, so you want to profile Lua or TeX code?
 
@PauloCereda Both, I suppose. Though the problem is more likely to be in Lua.
 
7:07 PM
@FaheemMitha hence my comment, "this does not sound much of a TeX bottleneck" in your code.
 
 
1 hour later…
8:08 PM
@PauloCereda the thin bit at the top of a bottle.
 
8:19 PM
@JairoA.delRio which one? (there are many CC licences)
 
8:36 PM
@DavidCarlisle The thing I love about this room is how cooperative we all are.
 
@AlanMunn we do try to answer all questions asked.
 
 
2 hours later…
10:43 PM
@DavidCarlisle quack
 
11:14 PM
@PauloCereda supper
 

« first day (3767 days earlier)      last day (1164 days later) »