@user2692669 Look in the log file for minted.sty and the version number should be available: what number is reported?

@egreg Package: minted 2015/01/31 v2.0 Yet another Pygments shim for LaTeX

ouch

@user2692669 Yes, you need to update. What TeX distribution are you using?

@egreg Texlive, gui crashed, I just updated from command the tlmgr and now I run update all

@user2692669 It should be OK, then.

@egreg I pressed alt+tab to check the window and I got unexpected stdin error and it crashed xD yay!!

@egreg skipping forcibly removed package dvips

what in the?

@user2692669 sudo tlmgr reinstall dvips

windows :P

oh

i huees no sudo

guess*

@user2692669 Sorry, can't help. Usually machines running windoze stop working when I'm near them. But I think they don't work when I'm far from them either. ;-)

@egreg You've been helpful enough, I'm just waiting for update to finish to check but I think you nailed the problem

@egreg I tried 3 different Windoze machines to get python working in one properly with minted and still I'm not sure it's fully functional

@egreg The weird thing is that the 2 previous won't compile and can't kill the process (probably get stuck in pygments) but when I compile form the third machine then the two previous ones can compile it (I think this is due to the fact that minted has already created something like directives?)

@user2692669 Possibly. All I know is that there used to be issues with minted in Windows. Good night.

@egreg Goodnight, thank you for the tips!

@DavidCarlisle I see I was right about \write18!

@DavidCarlisle Other than issues with \write128 we should not be OK testing LuaTeX v0.85 (and older versions)

@DavidCarlisle I'm locally using \write127 until things are resolved

@JosephWright Where do I submit feature requests for expl3?

@HenriMenke latex-l list or as a github issue at github.com/latex3/latex3/issues

@JosephWright yes, I think it should be OK, I'll do some tests later.

@DavidCarlisle Thanks. github.com/latex3/latex3/issues/288

@HenriMenke interesting time to ask that luatex looks like it's going to drop write18 support (as of last night:-)

@DavidCarlisle Oh, do you know the reason?

@HenriMenke it's a bit of a hack added to keep classic tex syntax, but luatex has os.execute and other lua access to the system so it doesn't add any features and complicates some aspects of \write internals so....

@DavidCarlisle True. I couldn't find the thread on the LuaTeX list. Did they discuss it elsewhere?

Nice hidden feature in TikZ arrows. :(

\documentclass{article}
\usepackage{tikz}
\usetikzlibrary{arrows.meta}
\pagecolor{blue}
\begin{document}
\tikz \draw[double equal sign distance,-Implies] (0,0) -- (1,0);
\end{document}

I can understand that bending two lines keeping them parallel is difficult, but the solution they found is disputable.

@egreg "They" being the TikZ people?

@HenriMenke Direct mails between me, David, Hans Hagen, Luigi Scarso and Akira Kakuto

@FaheemMitha Whoever wrote the arrow libraries.

@FaheemMitha And “disputable” means “unacceptable”, of course.

@egreg I'm surprised you didn't blame whoever wrote \pagecolor

@UlrikeFischer @UlrikeFischer But this is the point: what is »embedding«? Can you produce a PDF with text without embedding the used glyphs? I don't think »embedding« the glyphs is the crucial point. Linotype disallows embedding of the whole set of the fonts in a way that the fonts can be recovered from the PDF. The other company disallows to distribute a PDF to more than a hundred people, a ridiculous clause in their license contract.

@DavidCarlisle Once in a while it's not your fault.

@KeksDose Sure you can avoid embedding the glyphs. That's easy with maps and normal type1 fonts. The pdf then contain only a font name and it is the pdf reader which has to find a suitable font. Not embedding was the default for the "standard" fonts like times for a long time to get small file sizes. See the option pdftexDownloadBase14 in updmap.cfg. Legal questions are certainly important too but imho the OP wanted to know if the glyphs are embedded.

@UlrikeFischer Aha! The default behaviour of pdftex is to embedd the glyphs? So the PDF will have the same look, more or less, on every screen. But you can produce a PDF deliberately without!

@egreg Ok. Are you going to tell them? :-)

@FaheemMitha If I find an issue tracker

@FaheemMitha Well, there is one, but requires registering to SourceForge. :(

@DavidCarlisle I've got down to 56 LuaTeX diffs for the 2e test suite (OK, I get 60 but at least 4 are the \write16 business on Windows). Of the remainder, a few might be sorted if we show all box lines for all tests: they are caused by etc. when pdfTeX and LuaTeX have different lines so truncate at different places. I suspect if I alter that I can get down to about 50: around the level we might add LuaTeX to the suite?

@UlrikeFischer I see you're authorized to open tickets for TikZ/PGF: can you look at chat.stackexchange.com/transcript/message/25497486#25497486

@KeksDose The "default" behaviour depends on the settings in the map files/updmap.cfg, but yes you can decide if a type1 font is not embedded, in part or fully. Try out the following and you will see that cmr10 will not be embedded anymore:

\documentclass{report}
\usepackage[utf8]{inputenc}
\pdfmapline{=cmr10 CMR10}
\begin{document}
ABCDED äöüß
\end{document}

@egreg Yes, Sourceforge sucks.

@JosephWright yes I was thinking same other day it's getting to the point now where we understand the differences and we just need to live with them rather than normalise them away. In some cases (and for new tests) we might be able to adjust the test files to test what is needed in a way that doesn't show engine diffs, but for teh rest just check in luatex.tlg

@UlrikeFischer I get:

Does anyone here have contact with the TikZ/PGF people, whoever they are?

@egreg Imho the behaviour is to be expected: double lines are made by first drawing a thick (black) line and then a thinner (white) line on top. You would have to use a node to get a good filled arrow.

\documentclass{article}
\usepackage{tikz}
\usetikzlibrary{shapes.arrows}
\pagecolor{blue}
\begin{document}
\tikz\node[fill=white, draw=black,line width=0.8pt,single arrow] (arrow) {\hspace*{2cm}};
\end{document}

@KeksDose Yes, pdftex is warning you, that it is not a good idea not to embed cmr10. And to insert code: I'm currently sending first the text. Then as fast as possible insert the code, click on the "fixed font" button (which appears behind the upload button" and then send it.

@egreg And I knew I had seen an answer somewhere: tex.stackexchange.com/questions/72784/…

@UlrikeFischer Thank you for educating me. Is this senctence true: the default behaviour of texlive and MiKTeX is to embedd the glyphs of the fonts, which are (= the glyphs) used, into the PDF?

@KeksDose for many many years not embedding was the default (and an original "selling point" of acrobat that the files were small precisely because the fonts were not embedded. In recent years move more towards embedding as file size less of an issue and gives greater fidelity, but since the question is about commercial fonts, which often don't allow embedding, it's more likely to be an issue than with free fonts.

@KeksDose it's specified per font in the map file. (more normal to talk about embedding fonts rather than embedding glyphs)

@KeksDose Miktex still have "dvipsDownloadBase35 false" in the default updmap.cfg (no idea why ...). For fonts outside the "base" set, you can normally expect that the map-file embeds at least the used glyphs, sometimes the full font. \setmainfont should normally embed the font partially.

@UlrikeFischer @DavidCarlisle This explains why a printer company had difficulties with a PDF I delivered, they obviously expected only the outlines in the PDF, not the fonts. At least Linotype nowadays allows embedding of the fonts, as long as it can't be recovered from the PDF as fonts. OK, offline now...

@UlrikeFischer Well, this is what's done. But it's wrong.

@egreg If you know a better solution I will be happy to create a ticket on the tikz/pgf site. But imho it is not easy. E.g. the arrow tip don't need to touch the straight lines.

@UlrikeFischer well, the solution is (internaly) to apply crop on the thick line, to achieve the double line effect. But that's quite a complicated thing; moreover, not what people usually expect with crossing lines...

If I want to have line breaks inside a quoting environment, should I use \\, or is there a better way? From conversation in this room, I got the feeling that every time I use \\, God kills a kitten.

@JasperLoy Around 50% of mankind are women and you didn't consider the possibility?

@DavidCarlisle Found another LuaTeX bug :-)

@yo' @egreg I can create graphics, but I don't like it very much and I'm not really good at it. This is not my favourite playground ;-). So I won't start some complicated discussion about new features and new implementation with the tikz maintainers, I can only loose. But I can pass messages around if wanted.

@UlrikeFischer I just don't want to register to SourceForge just for this. Maybe @percusse could help.

@UlrikeFischer Not on TeX.sx. :-)

Ooh, TUG interview. Celebrity alert.

@barbarabeeton: On d.c.t.t, the german usenet group, there was a question regarding align: How to suppress the "empty subformula" after & to align the arrow with the A in the following code, but still get the correct spacing after the arrow? \mathopen works, but feels like a bad trick ;-(

\documentclass{article}
\usepackage{amsmath}
\begin{document}
\begin{align*}
&A\\
&\longrightarrow x\\
&\mathopen{}\longrightarrow x\\
\end{align*}
\end{document}

So, supposing I want to quote an address in a letter. I'm curently doing:

\begin{quoting}
\noindent
Address
\end{quoting}

Which works, but is ugly. Is there a better way?

@UlrikeFischer There must be no trailing \\ in align

@egreg Yes, remained from other examples, but it doesn't affect the spacing before the arrow.

@UlrikeFischer align inserts {} at the start of even numbered columns. So it provides the spacing for a relation or operation symbol. \mathopen{} is the workaround

@UlrikeFischer put the whole thing in braces: &{\longrightarrow x}\\

(not tested, no LaTeX here...)

@yo' Good idea. That works and looks much better than the \mathopen. (I actually had an example with braces but didn't got the idea to put the x inside them too ...).

@UlrikeFischer :)

Now, I gotta go teach improper integrals to my chem students :)

What is the CLI converter called to take markdown syntax and compile to pdf?

Pandoc must be it.

@Johannes_B Yes, pandoc it is. You might be interested in the interpreter package though.

@HenriMenke tex.stackexchange.com/questions/278852/…

@HenriMenke That sound interesting.

@DavidCarlisle 14 hours without electricity. :)

@clemens I redid the wikibook page on title creation this summer. Now i am looking at the german counterpart which is a translation of the former english page. And the discussion page is also weird.

@PauloCereda you blame that on ABNT ?

@DavidCarlisle of course, ABNT is evil. :)

@PauloCereda thoughtful looks like it was an etex.sty/sharelatex problem:

@DavidCarlisle I'm going to start adding LuaTeX .tlg files for 2e. We get a few fewer with LuaTeX v0.85 than v0.80 due to a change in handling of \local... (no \whatsit any more). I think we'll end up needing roughly 50, but want to do them in batches with a record of the reasons for them.

@JosephWright ah. I got distracted by real work so done nothing.

@DavidCarlisle interesting.

@DavidCarlisle Try $$\sin\pi$$\bye with v0.85 :-)

@DavidCarlisle ^^

@JosephWright before that I was trying to get minted to work without using \write18 but takes quite a lot of fiddling to get it to load at all as all the ifluatex.sty ifpdf.sty etc tests all go the wrong way in .85 (not unexpected but fiddly to get right again) ...

@DavidCarlisle Aren't we just going to redefine \write using the token library or \afterassignment or whatever?

@JosephWright that was the plan, but first I have to get minted to actually load without telling me that "someone has redefined pdfoutput" etc etc etc...

@JosephWright I get pi but no sin

@DavidCarlisle Indeed

@DavidCarlisle \let\pdfoutput\outputmode doesn't do it?

@DavidCarlisle I get the error with a TL14. But I couldn't get from all the comments if you already know the source or if someone (me) should investigate ...

@JosephWright that gets further then it complained that pdftex option had been used when not pdftex, as I say it's probably not that hard to fix everything but It'll have to be this evening, just \let'ing alias for the old primitives isn't enough as Heiko's package explictly use enableprimitives using the old names, and get upset if they don't then work

@UlrikeFischer I think it's one of the bugs I rememeber if etex gets loaded at the wrong time, so perhaps just forcing it to load early is enough, but as I note to Joseph I had to drop it and do some work:-)

@DavidCarlisle Ah

@DavidCarlisle WE need to get Heiko to update his stuff

@DavidCarlisle Hans does too, it seems

@DavidCarlisle I am a bit lost right now with the abnt etex business.

@Johannes_B sorry I'm out of time, will look later if you or @UlrikeFischer hasn't fixed it byt then:-)

@DavidCarlisle :-)

@Johannes_B I can't even find the entry in the wikibook… :(

@clemens en.wikibooks.org/w/index.php?title=LaTeX/Title_Creation

@UlrikeFischer -- i see this has been sorted out. i was about to suggest putting both the arrow and the x in braces, but decided to keep reading before i answered. i agree it looks better than \mathopen, and probably also easier to remember.

@Johannes_B yes, but where is the German page?

@clemens de.wikibooks.org/wiki/LaTeX/_Eine_Titelseite_erstellen

@Johannes_B Oh, there are pages without entry in the toc? How many might that be…?

@clemens Oh really? That isn't even accessible?

Oh maaaaaaann.

@Johannes_B Hm… this gets weirder and weirder: de.wikibooks.org/wiki/LaTeX:_Neues_Konzept vs de.wikibooks.org/wiki/LaTeX-Kompendium

@clemens So, how many different similar projects are there? If Jürgen knows something?

@Johannes_B This is best english i ever writtten down.

@Johannes_B is Jürgen @tampis (tex.stackexchange.com/users/31701/tampis)?

@DavidCarlisle @Johannes_B: I can avoid the error by loading calc. (And I checked in tl13 the class loaded calc, but has dropped this in tl14.)

@clemens I think so.

@UlrikeFischer Not sure if you can open that project or not, but you can try: sharelatex.com/project/564c569137ee5a733fea989d

@UlrikeFischer Loading package calc before anything else avoids the error mentioned. Same with the normal preamble.

@clemens Is there a way to automatically see all changes in the LaTeX relevant parts of the wikibooks? Also including discussions? I know you are doing something, but i never know where. Or if Jürgen has some helpful links.

@Johannes_B I don't know either. I haven't done anything for a few days but I'm either working on de.wikibooks.org/wiki/LaTeX-Kompendium or on entries in de.wikibooks.org/wiki/LaTeX-Kompendium:_W%C3%B6rterbuch

@clemens So we have Kompendium, Schnellstart, Wörtebuch and Neues Konzept which you just dug up. Oh my.

@Johannes_B de.wikibooks.org/wiki/LaTeX:_Neues_Konzept should be deleted altogether IMHO or integrated in the Kompendium (or the other way around). Having two books is less than helpful

@Johannes_B Maybe Kompendium, Schnellstart and Neues Konzept should all be merged into one

@clemens HEavily agreed

@Johannes_B the Wörterbuch isn't too bad, once the faulty entries are fixed

@Johannes_B we should get Jürgen in on the discussion…

@UlrikeFischer -- i know nothing about miktex, in particular, where to look for a log file. can you help with this, please: tex.stackexchange.com/q/278597

@clemens Agreed. Just trying to figure out how to geht to the main page of that wikibook.

@clemens de.wikibooks.org/wiki/…

@UlrikeFischer -- thanks much.

@DavidCarlisle -- i think you'll enjoy Henri Menke's second comment on this question: tex.stackexchange.com/q/278840

@Johannes_B commented

@egreg -- is this related to the "fuzzy" end caps on the double line in this answer: tex.stackexchange.com/a/278791 ?

@barbarabeeton I'm so tempted to post xii as an example of what tex would look like if it didn't use \

@DavidCarlisle -- you don't actually have to post it. a simple reference would suffice. (i was tempted to say that the first japanese implementation of tex that i ever saw used the yen symbol instead of \ . maybe i should ...)

@barbarabeeton I think the yen is just an encoding thing isn't it, it's the same slot as backslash in one of the legacy encodings so they just went with it.

@UlrikeFischer oh so looking at the chat thread @PauloCereda mentioned, seems I was right to half remember the error message, I just remembered the wrong cause:-) Thanks for sorting that out.

@DavidCarlisle We are getting old, as we forgot that important bit. :)

@PauloCereda not as old as you-know-who

@DavidCarlisle -- i didn't explore that possibility. i just assumed that there was no backslash on a typical japanese keyboard. but the encoding concept makes more sense. (to be sure, an input file for use with japanese tex looks a bit odd to my eyes, even when everything else is in what looks like plain ascii.)

@DavidCarlisle Voldemort?

@PauloCereda I got mail. :) <3

@DavidCarlisle ooh :)

@egreg yaaaaay <3

@barbarabeeton Yes, that's what I investigated on: the strange artifacts made me suspect about how the double line was obtained.

@JosephWright it could be that he goes by another name in these parts

@DavidCarlisle -- i think back to darmstadt and that lovely mastodon skeleton, and am grateful that there are creatures older than i that have survived in such good condition.

@barbarabeeton I'm older than the mastodon, according to @DavidCarlisle

@barbarabeeton At that time, typesetting was a quite hard job, but we had StoneTeX.

@egreg: You're probably the only one of us who watched even the Big Bang ;-)

@ChristianHupfer Nice show indeed!

@egreg What was the supporting band?

@ChristianHupfer But that makes him in some sense infinitely OLD. Oh no, TeX capacity exceeded!

@egreg -- though somehow you managed to avoid the later expansion; you have remained rather compact.

@Johannes_B The Big Band, of course

@egreg :-D

@egreg Did you know the Flintstones socially?

@FaheemMitha There was an email exchange between us.

@egreg Ah.

@FaheemMitha The mail forwarding application was called StoneMail, of course. One had to be careful when the forwarder was nearby, because it dropped the messages from the air.

@egreg Naturally.

@egreg -- hmmm. i would have thought that airmail delivery by pteranodon would have been the norm.

@barbarabeeton Similar, but StoneMail was much faster.

@egreg and painful. :)

@barbarabeeton The problem was that the Dimorphodon flied higher, so even smaller messages had a big impact, if they fell on one's forehead.

@PauloCereda LOL

Two computers in front of me. I have to remember that the inner keyboard and the inner mouse belong to the inner screen :)

@yo' Never mind, I found it.

S15.3.4

@egreg ok

@yo' I can understand that for complicated paths this is the sanest thing to do, but for straight lines?

@egreg well, for non-intersecting lines, you can always use the crop trick.

but note that it need not be the desired effect.

@JosephWright 127 failures in base with my 0.85, is that expected?

@DavidCarlisle No

@JosephWright thought not. It just finished, I'll look at the diffs..

@DavidCarlisle Do you have the very latest l3build installed?

@JosephWright that thought did just occur to me...

U    l3kernel/l3keys.dtx
G    l3build/l3build.dtx
U    l3build/l3build.lua
Updated to revision 6299.

@JosephWright ^^ I'll try again:-)

@DavidCarlisle With all of the checkins today you should be around 30 outstanding diffs

@DavidCarlisle The change of \local... from a whatsit to a special kind of node means that some of the test results vary between v0.85 and v0.80. I suspect we can't turn on routine testing until TL'16 as a result (though we add the test results for v0.85 now)

@DavidCarlisle I understand most of the outstanding failures but I'm trying to make sure I only add .tlg files when I'm certain of the reasons and that we can't deal with it some other way

@DavidCarlisle I'm not sure what to do about issues caused by engine bugs that are 'long-standing': you'll get a couple of weird 'there's no XXX in cmr' errors where the codepoint is random.

@JosephWright sounds familiar...

@DavidCarlisle One of the failures is that $$\sin$$ business, which clearly needs fixing at the engine end!

@DavidCarlisle The change of hyphenation approach shows up in several tests, which currently the expl3 ones don't pick up (as we don't typeset much there!)

@JosephWright If I do a new \newwrite allocator for latex, I wonder if we should do same for plain, currently ltluatex.tex uses etex.src for the classic allocations but with input and output streams now added to the list of things that that does wrong for luatex, I wonder if it ought to use an \e@alloc based allocation for everything..

@JosephWright what exactly goes, all math roman?

@DavidCarlisle We really do need to get etex.src updated I think

@DavidCarlisle I didn't pursue it beyond 'this is very odd'

@JosephWright that would be the other way, true.

@DavidCarlisle Nope

@DavidCarlisle $\sin a{\rm b}$\bye has a b but no sin

@JosephWright oh that's probably even more weird

@DavidCarlisle It's \nolimits that does it :-)

@JosephWright \nobase :-)

${\rm a} \mathop{\rm b}\mathop{\rm c}\nolimits$\bye prints a and b but not c :-(

@DavidCarlisle I am left wondering if there any unit tests for LuaTeX

@JosephWright I doubt it

@DavidCarlisle Same here

@DavidCarlisle Speaking of which, PLK is working on using l3build for biblatex: I've got a feature request from him to allow Biber to be 'slotted in' to test runs

@JosephWright re-implement arara in lua then put directives for the processing pipeline in each test..

@DavidCarlisle That's not far from what I suggested some time ago for engine-specific tests :-)

@DavidCarlisle We could easily add first-line parsing

@JosephWright I know:-)

@PauloCereda ^^^

@JosephWright Can you reopen tex.stackexchange.com/questions/278719/accents-not-in-math-mode ? It's not a duplicate, because of amstex. Reading the manual would help, of course.

@egreg Was just reading the same question :-)

$ ls ../build/test/*diff | wc -l
26

@JosephWright ^^

@JosephWright ooh :)

@DavidCarlisle Sounds about right, plus the .luatex.tlg files I've added means we are at around 50 variations, which is roughly what I think we needed to get down to before tackling them one at a time

@clemens Well, this will be helpful for non-english speakers: de.wikibooks.org/wiki/LaTeX/_Aufbau_eines_Dokuments

@Johannes_B reminds me of this: de.wikibooks.org/wiki/LaTeX-Kompendium:_Aufbau_der_Quelldatei

@egreg -- thanks for fielding this. i am pretty sure i voted for keeping it open because of the amstex (not latex), but nobody else was paying attention.

@clemens There needs to be a change.

@clemens Any response from Jürgen yet?

@clemens Did you use any templates so far in the german wiki?

de.wikibooks.org/wiki/…

@Johannes_B what I really miss is a usable table of contents in the Kompendium… I never find what I am looking for

@Johannes_B no

@clemens TOC missing, semantic markup missing, a reasonable markup missing (i am not used to wikitext)

@clemens Just checked, had to do a click orgy to find the page.

@Johannes_B I mostly just edited articles so far… not all pages seem to follow the same structure. The one I just linked is missing back links to the main page (I guess because no template has been used)

@clemens {{LaTeX/Inline|code=latex code}} is the alternate template for inline code. All that fuzz for those two words latex code. The original would be {{LaTeX/LaTeX|code=latex code}}

@barbarabeeton Does the AMS have a spare copy of “The Joy of TeX”? /me hides

There is also {{LaTeX/Usage|code=latex code}} and {{LaTeX/Example||code=latex code}}

@Johannes_B are those explained somewhere?

@egreg I'd like to here your thoughts. en.wikibooks.org/wiki/LaTeX/Page_Layout

@clemens en.wikibooks.org/wiki/Template:LaTeX/Example

@egreg -- let me see what i can do. i'll report via e-mail.

@clemens They are created on the very same pages they are explained as well.

@Johannes_B I was happy after drinking my glass of wine. ;-) I'll look at it later.

@egreg All right :-)

@Johannes_B do you know if they're meant to be used on the German site as well?

@clemens No freaking idea. To be honest, i would be more happy if we could define aliases and use <package>mathtools</package> instead of {{LaTeX/Package|mathtools}}.

@egreg For the novice, exercises sprinkled generously throughout each chapter encourage the reader to sit down at a terminal and learn through experimentation.

:)

@David: it seems I need to work on a l3arara.

@PauloCereda 1) finish/release the current version 2) finish/submit thesis 3) think about what comes next :-)

@DavidCarlisle Oh no! :)

@David: you can team up with @barbara for #2. :)

@PauloCereda I've already done (2)

@DavidCarlisle I mean, prod me to finish it. :)

@PauloCereda -- consider yourself prodded! nagged! encouraged! (i shall continue ... inexorably.)

@barbarabeeton ooh :)

@PauloCereda how is the thesis coming along?

@DavidCarlisle pretty good, actually. :) I need to write three more chapters, but I am getting there.

@David: sadly, soon I will need to go with ABNT. :)

@PauloCereda insist on DIN

@PauloCereda SHIELDS UP!

@DavidCarlisle The German equivalent? :)

@PauloCereda exactly

@JosephWright I am really scared! :)

@PauloCereda I find the whole business of these weird rules for theses very strange: my own one was basically 'probably print it on paper' :-)

@David, @Joseph: thankfully, I know how to tackle ABNT, but the result will surely be typographically questionable. :)

@JosephWright LOL

@PauloCereda I looked really hard in case there were any rules: there were not other than 'final copy has to be single sided'

@JosephWright print on Möbius band?

@DavidCarlisle :-)

@JosephWright: you got mail, have fun. :)

@JosephWright I just remembered that I modified my local lualatex.ini to add some alias for 0.85 do we have a plan for what's the best thing to do, and where.

@DavidCarlisle Well I've got a plan, but I need to see what everyone else thinks :-)

@DavidCarlisle Now i know :-( en.wikibooks.org/wiki/LaTeX/Page_Layout#Odd_and_even_margins

@DavidCarlisle I'll drop something to the team list in a bit

@Joseph: how many Teensies saved? :)

@PauloCereda 700

@JosephWright :)

@JosephWright do we support 0.80 in 2016/01/01 (do I have to guard the \newwrite with a luatex version test)

@JosephWright oh I added one anyway

@DavidCarlisle Presumably as all that will happen is the engine will complain we don't really need a test here: if it's LuaTeX then (probably) it will be >0.85 by next year so ...

@JosephWright well that's what I was wondering, similar thoughts made me wonder if we really needed to guard it with includeinrelease most changes (although not this) make no sense to back out anyway as they just won't work on the new engine if you revert the macros. Currently not checked in I have the \newwrite in includeinrelease guards for 0000 2015 and 2016 and the last in a luatex and luatexversion test, but it all seems a bit over-engineered

@DavidCarlisle Yup

@DavidCarlisle I'd say we can make a policy decision here: LuaTeX is not stable and given the changes going on it's not realistic to worry about backing stuff out

@JosephWright I think I'll put that to one side, check it in without and raise on team list.

@DavidCarlisle For the same reason I'll probably remove some stuff from l3build once TL'16 ships

@DavidCarlisle Sounds fine

@DavidCarlisle I'm working through the diffs for the test suite: I'll probably add .luatex.tlg files one at a time

@JosephWright "I don't do 2e" :-)

@DavidCarlisle Broadly true: I'm looking for things I need to know for L3

@DavidCarlisle Looking for engine variations, basically

@DavidCarlisle Need to pursue Hans again about various things

@JosephWright your usual excuse, but gradually you get sucked in to the void 2e support

@DavidCarlisle I'll leave you and Frank to mend multicol and the like ;-)

@DavidCarlisle I'm going to hyphenation for L3 probably tomorrow :-)

@DavidCarlisle Speaking of 2e support, we really need to get hold of Heiko

@DavidCarlisle He's not seen the engine stuff I think

@JosephWright with individual penalties in each discretionary?

@DavidCarlisle Step 1: actually load some patterns

@DavidCarlisle Perhaps Johannes might take an interest when I start messing about with this

@DavidCarlisle Speaking of which, do you like my 'only EC support for pdfTeX' plan?

@JosephWright yep with his iffoo packages doing the wrong thing (and so graphics) it's hard to test many real documents with .85

@DavidCarlisle Yup

@DavidCarlisle Like I say, I'll mail the team list shortly (need to do some cleaning first)

@JosephWright i must look in detail at that again...

@DavidCarlisle OK: recent checkins should make it clear what I'm up to (only Bruno seems to read all those automatic mails ...)

@DavidCarlisle Also need to do \write18 support (particularly for LuaTeX as it's no longer \write18!)

We probably should tell people about that ...

@JosephWright as Imentioned I started to do a luatex version that today but got busy, wil try again this evening probably, i wanted to test with minted but didn't quite get that to load this morning

@DavidCarlisle Reasonable plan

@DavidCarlisle L3 code will be easier: no direct access to \write so shell escape doesn't need to hack about

@PauloCereda That would be just to remember the old times. After StoneTeX it was a big leap forward. ;-)

@egreg :)

@JosephWright in 2e need to decide whether to detect \immediate. write18 is almost always used with immediate so could ignore it and use directlua but if you can detect immediate then you could use latelua for the non-immediate case

@DavidCarlisle Hmm, tricky: I doubt Hans has to worry about that

@Johannes_B I gave a look at the page: it starts quite well, but ends in a complete mess, mixing up disparate things. The syntax with \lhead, \rhead and \fancyplain has been obsolete for at least a decade.

@egreg i would separate in geometry and package for changing header/footer and explain a few things instead of bringing up several points and saying over and over again: package geometry does help here. And i wouldn't tell people how to set lengths manually that can be changed with geometry.

@JosephWright I did wonder if i should ask for a callback on \write to avoid hacking at the macro level if a callback got the stream number an immediate or not flag and the string to write, it would be somewhat easier, there are callbacks on file access but only for the filename it seems

@DavidCarlisle I wondered about a callback and had a feeling I'd not seen one

@DavidCarlisle Hans will love us (and probably not document it!)

@JosephWright he should be happy to have two more users

@DavidCarlisle See mail

@PauloCereda oh no! I'm so glad I can make the thesis to my liking

@yo' It's all incredibly odd: who cares about this stuff

@JosephWright exactly. Funny thing: In France they more care about the opponent reports, and these are done by the scientists and usually look incredibly horrible :D

@yo' Indeed

@yo' Here in the UK it's the two examiners, who just want something they can read and which (probably) looks like a classical manuscript

@JosephWright yes, replied...

@JosephWright in France, you send the PDFs (ok, one examiner got a printed and stapled copy, printed 2-on-1). Once you get the reports, you can make changes to the thesis based on the reports, and you give it to the university to print it.

@yo' I did a viva recently and we are now allowed to get a PDF copy but the assumption is still a print one. As we do oral exams, you really want a copy in front of you (with notes).

@JosephWright "viva" is the thing in the middle of the studies, or the defense?

@Johannes_B I agree.

@Joseph But in general, for the examination itself, you bring several printed copies; in France, the thesis is printed in >60 copies usually.

@yo' Viva voce: oral examination. In the UK you submit your thesis (nowadays only soft bound though I just overlapped with the period when it was hard-bound), two copies. One goes to the internal examiner, one to the external. The viva then takes place with just the three of you in a room for (normally) 2-3 hours.

@JosephWright wow, that sounds like a good thing to have.

In Czechia, you have to pass "state exams" -- submitting a short ~20page manuscript, it's with 7 examiners and lasts about an hour.

@DavidCarlisle LuaTeX manual updated:

> You can now open up to 127 files with \openout. When no file is open writes will go to the console and log. As a consequence a system command is no longer possible but one can use os.execute to do the same.

@DavidCarlisle Doesn't apply to \read

@JosephWright Good news for all packages that use \write18

@egreg Indeed

@egreg Well that can fail anyway, so ...

@JosephWright And, as far as I know, os.execute has no protection.

@JosephWright how comes?

@egreg No, but that's been the case for some time (with LuaTeX you can already do \directlua{os.execute(<whatever>)}

@yo' If shell escape is not enabled

@JosephWright ah ok, but that's what you want, for the sake of safety...

@JosephWright Which I find extremely bad.

@yo' Yes, what I mean is most packages using \write18 do look for it being on or off

@egreg Complain at Hans

@JosephWright well, I do not. I probably should :)

@yo' Yes :-)

@JosephWright not epstopdf.

@JosephWright You're overoptimistic.

@egreg tru dat. On the other hand, my code is a personal tweak for the dirtiest of dirty tricks

@JosephWright oh for goodness sake:(

@egreg I think that got fixed, hang on I'll check manual...

@JosephWright In imakeidx I use pdftexcmds to get the value of \pdf@shellescape. If Heiko changes it to return 3 (or any suitable signal value) in case LuaTeX > 0.85 is used, it should be easy to fix everything, because I use \imki@exec (presently \immediate\write18) and not directly \write18. So I can define \imki@exec to be \directlua{os.execute{#1}} in that case.

@egreg I just tried: in the current luatex os.execute only works if --shell-escape is used, so there is protection.

@egreg Certainly

@UlrikeFischer Ah, sounds sensible :-)

@egreg plan I think is to make write18 work if you load such a package

@DavidCarlisle Yes. I wonder why just 127 and not 255.

@JosephWright The manual describes it also, but I don't know if (and how) is possible to execute some commands when --shell-restricted is used.

@DavidCarlisle But if \write18 called indirectly, it's easy to branch the cases.

@egreg sure

@egreg sorry for the formatting but ^^^

@JosephWright Maybe one can convince Hans to provide a sensible return value for \shellescape (which they presently don't), say 10, 11 or 12 (no escape, full escape, restricted escape).

@DavidCarlisle --shell-restricted seems currently to affect write18 (as with pdftex), but will it do anything with write 18 gone?

@UlrikeFischer given that the engine seems to be changing by the day, who knows. It would be easier if there was some public announcement/discussion of requirements but that doesn't seem to be how things work... I don't see why in principle it can't allow os.execute with a fixed path, but ....

@DavidCarlisle yes discussion in a distribuated system like latex is more open. That's one of the reasons why I prefer it over context ;-)

@StefanKottwitz Do you have a minute to use the moderator hammer for some questions? I can add them here and you can close them?

@Kurt sure

@Kurt you could also flag, I can see

not necessarily public chat I mean

as you like, site discussion is public and free anyway :-)

@StefanKottwitz Ok, I will flag, please close when you saw it!

@Kurt Sure! If I think the same :-) of course. I also always acted on your remarks to obsolete comments for cleaning.

@StefanKottwitz Well Then I can write in german ;-) That's much easier for me. My school English is not so good ...

@StefanKottwitz Just joking!

@Kurt I got school English besides 3 Grappa plus 2 beer, in the Pizzaria after work. I'm building a ship in Port of Maghera, Venezia.

so, some minutes before I fall asleep :-)

@StefanKottwitz No problem; I have flagged that one I think they can be closed.

@Kurt I don't see a new one in the small (<10) mod queue