@WillRobertson No, only one.

@egreg Huh, there you go

@PauloCereda ^^^ Ciao!

To the other chat friends: sorry for spamming, but I couldn't resist the temptation of making another photo album!

LaTeX2e <2017/01/01> patch level 1

@CarLaTeX Good morning Carla.

I love this space. There is so much to learn and so friendly colleagues!

:-)

And of course ducks! :-)

Although I am related to penguins and lions. And sometimes GNUs.

@Jan Hallo Jan! Yes, this is a very friendly place!

@Jan There was also a penguin in my previous photo album :)

@CarLaTeX So thank you for your kind words and the inverse arrows. :-)

@CarLaTeX Oh, what a pitty, I missed it.

@Jan ^^^ see the last photo

Could anyone answer the comment of this post: tex.stackexchange.com/a/350684/101651? Thank you!

@WillRobertson I think it has to do with the order of the key processing. The Ligatures-Key is handled after the Mapping key:

\defaultfontfeatures[\rmfamily,\sffamily]{}
\setmainfont{Cambria}[Mapping=text-tex,Mapping=arabtex-trans-dmg]
%"Cambria/OT:script=latn;language=DFLT;mapping=text-tex;mapping=arabtex-trans-dmg;"

\setsansfont{Arial}[Ligatures=TeX,Mapping=arabtex-trans-dmg]
%"Arial/OT:script=latn;language=DFLT;mapping=arabtex-trans-dmg;mapping=tex-text;"

@CarLaTeX would it help if I answered along the lines, "tikz is weird, just copy and paste snippets you find on the internet until it looks about right" ?

latex2e 2017/01/01 Patch Level 1 is in texlive

@WillRobertson 2e PL1 is in texlive but fontspec hasn't been updated yet for the emph thing, did the upload go OK?

@DavidCarlisle Read all the tikz-pgfmanual is humanly impossible!!!

@Jan OK, I'll remind you that, ciao!

@DavidCarlisle Hahaha I suppose your love towards tikz is the same that the one towards Vim...

@CarLaTeX no tikz is OK, it's just picture mode with a few extra commands, vim is vi so just an evil trick to confuse people and divert them from the one true editor.

@DavidCarlisle :):):) So you only don't like the tikz manual, do you? I can't even imagine how they managed to write it!

@michal.h21 I am here, if I can provide more info. Thanks.

@CarLaTeX no to be honest tikz is fine it's just that I (mostly) dropped out of tex for about 10 years and when I came back almost the only thing that was different was tikz (about which I know nothing even if I have a gold badge) almost all the non-tikz questions could have been (and were) asked last century on comp.text.tex so are easier to answer.

@michal.h21 do you mean you can not reproduce this problem with the math SVG on your end? I am using Linux and TL 2016 and did full update. When I removed the adjustbox, everything now works just fine. No problem. As I show in the MWE I posted.

@Nasser if I use the basic configuration for dvisvgm, then it works fine:

 Make:image("svg$","dvisvgm -n -p ${page} -c 1.4,1.4 -s ${source} > ${output}")

I can imagine that there may be some issue with the method where dvisvgm creates all pictures in one rum.

It is definitely corrputing the math images. This is the same file I have been building for years. Always worked. Now I found it is missing all the math images from it. I did not change my main.mk4 for years, since I got it from you.....

Here is screen shot, so you see

But I need to make MWE....

@DavidCarlisle Ok, I didn't know your background.... :):):)

@Nasser I don't think we have any configuration for adjustbox, only package which was recently updated and is used in your mwe is listings

@DavidCarlisle Oh really? That's a pain :( Petra wanted to clarify something with me, and I was a bit slow to respond but that was, oh, almost 24 hours ago!

@michal.h21 Ok. I think in this case, may be I should delete the post. I remove adjustbox any way. I used it once. So no need to waste time on this? I need to spend more time to make MWE which use main.mk4 again, and I already spend hrs on this.

@DavidCarlisle I played with Harbuzz and Luabidi yesterday and I was able to correctly shape the whole paragraph with multiple fonts. Reinserting different node types at the right position wil be more difficult, but probably doable. And today I figured out that I can use font data provided by luaotfload, so there is no need for custom font loader.

@UlrikeFischer Hmmm, still weird to me. Anyway, I'll change the code so that you can't set mapping more than once!

@WillRobertson What's the unicode-math status?

@JosephWright I can upload soon if you like. Will start the interminable build/fix test/rebuild cycle now :)

@michal.h21 if you think adjustbox is not supported, may be you can post that as answer if you think it will be useful. Else I will just delete the post if not needed as I am ok now. Not using adjustbox and the problem gone.

@WillRobertson It's the one outstanding before an expl3 update

@Nasser no, it will be better to update your question, so we can figure out what!s going on. it should be fixable.

@michal.h21 Ok, need to spend more time on this.

@WillRobertson the good part though is that the bug is in the wild long enough for me to get a badge for an answer saying its a bug:-)

@DavidCarlisle dammit

@DavidCarlisle like you need more badges :)

@WillRobertson two badges it seems ^^

@WillRobertson I'm not sure what you mean by "can't set mapping more than once". The original code (from arabtex) start with a font with mapping=tex-text and want to locally change this to mapping=arab-xxx. So this must be possible. Also if you ensure that only one mapping is in the feature you must also ensure that it is the correct one. That means that if two mappings are set in the keys with e.g. Ligatures=TeX and Mapping=arab-xxx) then the correct one (the last and not Ligatures) must win.

@UlrikeFischer Sorry for the confusion — I meant from XeTeX's perspective, you can only have one mapping file active at once. So yes, definitely — the code needs to ensure that the most recent invocation takes precedence

The number of tests that I always need to rebuild and manually check is the real reason I should set up continuous integration!!

It never occurred to me that I'd have a good need for it.

@WillRobertson It's quite handy, I've found :)

@WillRobertson I'm set up to do this on a local VM so if you want I can work out the requirements

(Damn line breaks right)

@WillRobertson Could change the end-of-log rule to look for END\-TEST\-LOG$

I guess that's safer; it's currently ^END\-TEST\-LOG$ ?

@WillRobertson A simple equality for the line as a whole, but yes effectively

@WillRobertson Probably we can simply change the TeS here: it avoids making the Lua more complex and there is not really much in it

@JosephWright I guess if we keep it to \LONGTYPEOUT{^^JEND-TEST-LOG^^J} then it shouldn't impact on already-existing tests

(I assume that's what you mean?)

@WillRobertson Yes and yes

@WillRobertson Give me five mins: still sorting out new work PC!

@JosephWright no problem, I have it at my fingertips

@JosephWright i.e., I can do it for once :)

@WillRobertson My Git and SVN copies of L3 are back to normal :)

@JosephWright That's a relief :

:)

@WillRobertson Certainly is for me: new Win10 machine at work, a few issues just at the moment :(

@JosephWright It's always fun/daunting setting up a new PC!! I've heard the "linux" subsystem is really great in Windows 10... it's about time :)

@WillRobertson Have that activated on my PC at home (a gaming 'console style' one, really), but not yet at work: I have GOW so don't tend to worry too much

@JosephWright GOW?

@WillRobertson GNU On Windows: small set of tools for the command line

@JosephWright Oh neat! Last time I used Windows seriously I tried to use Cygwin and do not have fond experiences... good to hear there's been a saner alternative

Okay, unicode-math has been uploaded

@WillRobertson GOW is very small, just 'key' tools

@JosephWright Right -- usually that's all you want/need

@WillRobertson Great: L3 tests will hopefully be working by Sunday, and I can update CTAN with expl3

@JosephWright wait, does it mean you finally have things like cp (rather than copy)?

@yo' Yes

@yo' github.com/bmatzelle/gow/wiki

why I'm asking: I would like to do certain things in a multiplatform way from within pdflatex (via \write18)

@yo' Nahhh… just \immediate\write18{texlua ...} :)

@WillRobertson no lua please :-)

@JosephWright Back to the CI stuff... when you say "I'm set up to do this on a local VM so if you want I can work out the requirements" — I guess I don't know what I might be getting myself into

@yo' You don't like it?? I've found it pretty neat…but if all you're doing is shell scripting I can see that it might be overkill

@WillRobertson basically I need pdf/lua/xelatex, cd, cp, mv, mkdir, rm, pwd (can be overcome by FLS), and all of the file manipulation just on one file (nothing like cp * ../foo)

@WillRobertson If you want a small-ish TL install, you need to work out which bits you need. For you, simply copying the setup script in the L3 repo will be fine as you need almost everything there anyway (the contrib tests cover fontspec)

@WillRobertson Take texlive.sh from L3, cut out e.g. CTeX and you'll be fine

@JosephWright ah great!

@WillRobertson I wrote a TUGBoat on this ;)

@JosephWright That doesn't sound too daunting at all, thanks to your hard work

@JosephWright Oops, I'm a bit behind on my TUGboat reading!!

@WillRobertson Only takes a few minutes once you know what you want to do

@UlrikeFischer @barbarabeeton I just realized I will simply need many of the features of expl3 to do my job well, so I'll likely stick to 2014 versions or alike.

@yo' :) I know the feeling

@WillRobertson the most difficult thing is avoiding the most recent features

@yo' We list them all with dates ...

@JosephWright I know, but still...

@yo' For example, siunitx needs 2015/09/11

@WillRobertson me, I use cygwin:-)

@DavidCarlisle I never doubted your ability to get arcane systems up and running :)

@WillRobertson well I need a window system and I don't think GOW comes with X windows.

@DavidCarlisle Indeed it doesn't

@DavidCarlisle That's what Xming or Vxcsrv are for

@JosephWright yes but I don't want to get X from one place bash from another, make, gcc, ssh, perl, .... we just use cygwin for everything and (mostly) out build scripts work on macos, linux and windows/cygwin even though they use a wide variety of tools

@DavidCarlisle Oh, sure

@DavidCarlisle Of course, for you it's all 'official' whereas for me I need to arrange what I can whilst 'higher up' they'd like us all to have PCs with identical set ups ...

@JosephWright apart from the building latex bit

@DavidCarlisle Well, my laptop can do that: that said, in the past I've just paid for my own PC at work to get something sensible, and one of my colleagues was telling me that she uses her (own) Surface rather than the univ-issue machine ...

@DavidCarlisle Speaking of which, my new desktop is slower at building LaTeX than my 3 year old laptop

@JosephWright lacks an ssd?

@JosephWright well, I do not depend on this package :)

@DavidCarlisle Yes, I suspect that's it

@JosephWright ssd are great for lots of small files :-)

On its way to CTAN: github.com/wspr/breqn/releases/tag/v0.98e

@WillRobertson Cool

@WillRobertson Mainly my changes?

@JosephWright ONLY your changes :)

@JosephWright Plus an update to all the test files

@WillRobertson have you got fontspec unstuck as well?

@WillRobertson Ah, 'It's all JAW's fault' in other words

@WillRobertson Let me know which repo's you want Travis-CI for and I'll send you some appropriate pull requests

@DavidCarlisle No... I'll email Petra now

It's obviously pretty trivial but I wonder whether this would almost be worth providing to people in a consistent way for various packages...

meanwhile, it's late, so good night :)

@WillRobertson Night

@WillRobertson Most users have Windows ...

@JosephWright Any plan for \bool_new_true:N?

@WillRobertson night!

@yo' Er, not at the moment

@JosephWright dammit, extra lines of code...

It is too bad to use a clist_map_inline for keys_define to horten code?

@yo' Log an issue or ask on LaTeX-L :)

@yo' No, that's fine if it works for you

@JosephWright I know I can do it :)

@JosephWright ok, then that's what I'm gonna do likely (got a lot of bool class options, some defaulting to true, some to false). I just think, what is better?

@yo' I guess the team probably favour writing them out explicitly as then the code clearly documents the interface

\bool_new:N \l_myclass_foo_bool
\bool_set_true:N \l_myclass_foo_bool
\keys_define:nn { myclass } {
  foo .bool_set:N = \l_myclass_foo_bool
}

---

\bool_new:N \l_myclass_foo_bool
\keys_define:nn { myclass } {
  foo .bool_set:N = \l_myclass_foo_bool,
  foo .default:n = true,
}

(I hope I got it right, writing without testing)

@yo' You don't need to create all of the bools: we allow the 'shortcut' that \keys_define:nn will do that for you

@yo' Also don't need .default:n here

@JosephWright really? Ok, that simplifies a lot. (But wasn't that generating warnings with check-declarations?)

@JosephWright (I had a mistake), well, you do if you want to default to true, don't you?

@yo' No: \keys_define:nn looks for the variables when you do a ..._set:N line and creates them - things get too verbose

@JosephWright ok, good then, thanks!

@yo' A boolean should always have the default as true: perhaps you are thinking of the initial value?

@JosephWright ah yes, that's it!

\keys_define:nn { myclass } {
  foo .bool_set:N = \l_myclass_foo_bool,
  foo .intial:n   = false, % If required
}

@yo' BTW, most of the time the variable is likely to be internal

@JosephWright right. Actually, these are "almost constant" (basically, constant after 30 lines of code)

@yo' That happens

so internal/external is not that important as nobody (not even the class itself) is supposed to ever change the value

it' only important from the POV of documentation

anyway, gotta go, boss' Bday party begins

see you later!

[OT]: has any of you encountered windows installations where some programme is flodding AppData/Local/temp with TTxxx.tmp files that turns out to be TrueType fonts (Arial and Arial Bold). Somehow that programme is not cleaning up after itself so after a while it ends up filling the HardDrive. Hmm, how does one debug something like that.

Hmm, since it happens every time, it should be easy to pin point which program is behaving badly

does anyone wish to help reopen? i have a working answer not dealt with in the "duplicate" question.

@WillRobertson I guess it's too late at night but does the fontspec update lose the tu*{def|fd} files from ctan.org/tex-archive/macros/latex/contrib/fontspec (which are the ones prior to the PL1 fixes)

biber doesn't turn \'{\i} into a combining acute i U+0301 does it?

@DavidCarlisle Yes it does

@DavidCarlisle I had this discussion with @plk recently

@JosephWright that is stupid:-)

@JosephWright can you post an answer to current question. what is the user supposed to do?

@DavidCarlisle Forwarding mail

@DavidCarlisle He should probably simply use \'i see also tex.stackexchange.com/questions/345476/…

@UlrikeFischer Yes: I've marked as a dupe

@UlrikeFischer or use bibtex:-)

@DavidCarlisle -output-safechars

@JosephWright yes except biber needs bleeding edge perl which I decided not to build (and the one in tl2016 doesn't work for me so the option doesn't help:-) (but outputting suffix combining characters for a tex related tool makes no sense as a default, they don't even work properly in xetex/luatex never mind pdftex)

@DavidCarlisle You take it up with PLK :) The Unicode situation on combining chars is clear: they are suffixes whether we like it or not

@JosephWright yes but taking valid tex input and generating output known not to work seems less than optimal. (I can understand if it doesn't remove combining characters that are in the input)

@DavidCarlisle As noted in the answers for the question @UlrikeFischer's pointed to, the 'expected' input here gives the right output

@JosephWright Only if you have incorrect expectations:-). I'd expect \'i, \'\i, \'{i} and \'{\i} to be equivalent input so produce the same output.

@DavidCarlisle \'\i hasn't been required for 15+ years!

@DavidCarlisle PLK was keen not to have to add more special cases to Biber: you can raise with him :)

@JosephWright I know, but that's because it's defined to be the same as \'{i}

@DavidCarlisle For macro expansion that's fine, but for other tools things don't work like that (as you know)

@JosephWright yes I was just going to say, why aren't all macros expanded before biber sees the text anyway (in which case all those forms ought to have ended up as the same LICR)

@DavidCarlisle It's reading a .bib file

Anyone got the following combination: Windows 10, TeX Live 2016, an account without Admin rights?

I think my IT people have set something weird: can't run Biber even though it's extracting into the cache fine

@DavidCarlisle It is certainly not optimal. But \myspecialcommandforiwithaccent is also (assuming a suitable definition) valid tex input and you wouldn't expect biber to handle it.

@JosephWright I have and it works. But it is possible that the temp folder is blocked thaen you can set another one: github.com/plk/biber/issues/70

@UlrikeFischer yes I would (or at least I'd have expected biblatex to handle it before passing any such strings to biber

@JosephWright oh yes so it is, hmm (@UlrikeFischer)

@DavidCarlisle Hm. Would you expect bibtex to expand such a command?

@UlrikeFischer If it did nothing I could understand, but taking \'\i and changing it to something that doesn't work can't really be justified I think.

@barbarabeeton what sans font and tt fonts do the ams plan to use with stix? (new question on site)

@DavidCarlisle Well it would work fine (assuming a suitable font) with lualatex or xelatex. I do understand your concern -- I had some discussion about this already, e.g. github.com/plk/biber/issues/65 -- but biber can't let everything unchanged as it has to convert things to unicode for the sorting. So it is really the best to avoid to many tex accent command and to use direct unicode input in the bib.

@UlrikeFischer except (as came up again recently for the things that don't work like hyphenation)

@DavidCarlisle -- don't know, but will find out.

@UlrikeFischer I have feeling what they've done is prevent non-Admins from running any executables not installed in a standard location (I can move the .exe and it still fails)

@UlrikeFischer Ultimately I can get an unpacked Biber into the right place, at which point the issue will go away, but I wanted to be sure that I understood the issue first

@DavidCarlisle I just run into an error when running the following with lualatex and now wonder about the status of xunicode (deprecated? should work?)

\documentclass[12pt,a4paper]{article}
\usepackage{fontspec,xunicode}
\begin{document}
bac
\end{document}

@UlrikeFischer Doesn't load with LuaTeX: historically fontspec 'worked around' that to load it but it's not needed anymore

@UlrikeFischer Team recommendation is that xunicode is normally not required

@JosephWright I know that fontspec did worked around it, and also that fontspec did load xunicode so that a manual loading is not needed. But I bet there are thousands of documents around which loads xunicode ... (this one came from a knitr question), so imho some question will pop up.

@UlrikeFischer Sure, but xunicode itself hasn't changed for some years and when we revisited what it does for addition to the kernel we decided a rather 'lighter' approach was best

@UlrikeFischer I never understood why they put that test in. There is virtually nothing in xunicode that applies to xetex and not luatex. I'd say that's a package bug that they can easily fix by removing the test for xetex.

@UlrikeFischer in fact texdoc xunicode says exactly that:

(When other engines become available, the package will need modifying
not to check solely for XeTeX.)

@UlrikeFischer I suppose I could mail Ross and ask him

@DavidCarlisle I should mail him too. I owe him a mail since weeks but simply don't manage it. Fighting with the handling of my father's legacy ...

@egreg: Isn't there a specialized version for @DavidCarlisle too? latexfeatures package? ;-)

@ChristianHupfer He should write one. :P

@UlrikeFischer oh sorry to hear that, I'm cc'ing you on this one anyway.

@egreg: Here it is

@DavidCarlisle It happend in the summer it is not new. But you wouldn't believe how much time it can take to sort it out.

@ChristianHupfer Best collection of hidden features.

@ChristianHupfer There was latexbug.tex but I just retired it this morning (in the latex svn)

@egreg high level code here, I thought it was one of your answers initially.

the \end{document}\bye was a nice touch I thought.

@DavidCarlisle I can write even worse code 😝

@egreg as we see regularly!

@UlrikeFischer I've tracked down my issue: our IT support have blocked all execution of binaries other than those explicitly installed by administrators. So no portable apps or anything in TEMP!

@JosephWright oops

@DavidCarlisle Ah, ah, ah!

@DavidCarlisle What's the next test match?

@egreg one day 20-20 match coming up bbc.co.uk/sport/cricket/36805572

@DavidCarlisle Indeed

@DavidCarlisle The sales pitch for AppSense (the tool for this) says something about being easy for users ...

Good evening all together. I was just reading an answer from Gonzalo Medina (tex.stackexchange.com/users/3954) and saw that he's not active anymore since the end of 2015. I really liked his answers. Do you have further information if something happened? The same is true for the creator of BiBLaTeX (texdev.net/2012/04/23/biblatex-a-team-to-continue-the-work) - also strange. I don't want to start a discussion - I was just wondering :).

@Dr.ManuelKuehner I (and others) tried hard to find out what happened to Philipp Lehman with no success

@JosephWright Ok. Thanks. Sad story. It's amazing how LaTerX relies on the effort of only a few (un-payed people) compared to the number of users.

@Dr.ManuelKuehner According to @PauloCereda (who is in contact with Gonzalo) Gonzalo is very busy and has a family to care of. And Colombia is not a really safe place

@JosephWright The same as with Carsten Heinz, the initial author of listings

@ChristianHupfer Didn't know @PauloCereda had done that: good to hear

@JosephWright Paulo told me so some weeks ago

@ChristianHupfer Ah

@ChristianHupfer Don't know Carsten at all, but I was in reasonable touch with Philipp: the odd thing is no-one knows of ever meeting him

@JosephWright I think many (La)TeX developers are still in a more or less academic work place and it's hardly believable that they vanish completely. But it seems to happen

@ChristianHupfer On the team, Will, Bruno and I are in academia

Perhaps also Javier (@DavidCarlisle, do you know what he does?)

@JosephWright Well, I did not mean LaTeX core developers 'only'.

@ChristianHupfer Oh sure: I was just pointing out that it's not universal: @DavidCarlisle, Frank, Chris, Johannes, Robin and Rainer are all not in academic jobs

@JosephWright I know ... I was thinking about most package authors. Perhaps developers was the wrong term.

@ChristianHupfer It's variable: certainly stuff supporting maths work tends to come from academics, e.g. Till Tantau

@ChristianHupfer Thanks for the information!

@Dr.ManuelKuehner Well, and sometimes users of TeX.SE need a hiatus ;-)

I make dinner now. Maybe until later.

@Dr.ManuelKuehner Guten Appetit!

@JosephWright In germany, that names are very common. There might be a thousand people with that name. So a google search turns up lots of results. :-(

@DavidCarlisle Interesting times with @WillRobertson's Travis-CI set up: the test files are broken to start with ...

@Johannes_B Yes, I've met a Carsten Heinz during my studies, but he wasn't the one I mentioned

@ChristianHupfer I bet there are far less Christian Hupfers or Johannes Böttchers in the world than Martin Müllers.

@Johannes_B By far less ... yes...

@PauloCereda, @DavidCarlisle I see I'm now the Travis-CI expert for the team ...

@DavidCarlisle yes indeed — at least I hope so!!

@JosephWright Excellent, good start :-) No need to debug ob my behalf!!

Just got in, anything exciting happening?

@DavidCarlisle you mentioned it

@StefanKottwitz seems like everyone left as I arrived:-)

@DavidCarlisle I did not know you were here :-o

@DavidCarlisle -- i got an authoritative response regarding stix 2 font sans and monospace; question answered.

@barbarabeeton yes saw thanks.

@barbarabeeton voted even:-)

Can anybody tell me what a linear transformation of a function is exactly? The function maps options to the reals. I've found a lot of stuff explaining it when applied to vectors, but I'm not sure I'm reading about the correct thing!

Note that I probably don't need a very detailed explanation because I almost certainly knew this once. Or not.

@cfr a linear transformation of a function? or a function being a linear transformation? a linear transformation is a function from one vector space to another given as a each element of the target vector is a linear combination of the elements of the source (so a matrix multiplication)

@cfr what do you mean by "option" in this context?

@DavidCarlisle Options in the decision-theory sense e.g. choices of things to spread on toast. So the function maps, say, marmalade to 3 and Marmite to 303 or something. So if f: {things to spread on toast} -> R, what's a linear transformation of f?

@DavidCarlisle I guess I'm being dumb, but I don't see how the definition of linear transformation applies here.

@cfr nor me, you can of course apply a linear transformation to the resulting vector, eg double the resulting number.