TeX, LaTeX and Friends - 2017-01-24

« first day (2277 days earlier) ← previous day next day → last day (2662 days later) »

6:28 AM

Why is this closed as a duplicate

2

Q: Pdf_tex not found

I am trying to include a .pdf_tex file into my latex document. I saved the file from svg in Inkscape to .pdf_tex file. And I use this in my latex document: \begin{figure} \centering \input{myfile2.pdf_tex} \end{figure} myfile2.pdf_tex is in the same file as the main folder. But the error s...

graphics inkscape

The title is wrong, it is not pdf_tex that is the problem, but rather the pdf that goes along with it, the pdf is found, but apparently contains too few pages. Thus the question is not similar to the question it is marked as a duplicate of.

3 hours later…

9:21 AM

@cfr Chemistry journals publish the schemes ['pictures' :)] we submit as-is, so they have to look right. There are some technical things that are hard (line join mitring) plus for any complex scheme you need to get it right visually: an element of 'artwork'

@DavidCarlisle No: something to do with @WillRobertson moving the optional argument?

@egreg Oh goodness: really, LuaTeX + DVI mode = a bad idea

@JosephWright but same as tex4ht; what do you do if you have a dvidriver based workflow and users start sending you bug reports that their luatex documents don't work in it....

@DavidCarlisle Be honest

@DavidCarlisle These things simply aren't going to work long-term with LuaTeX: Hans is pretty clear on the fact that they are not actively doing anything with DVI mode

@JosephWright no idea: it was middle of that @@##@ xepersian document once I had removed enough control characters to get the packages to load those two commands were undefined but I found them in an old copy (@WillRobertson <<<<)

@JosephWright radical suggestion!

@JosephWright I know that:-)

@DavidCarlisle I guess I'd simply bail out with a warning

@DavidCarlisle Problem is of course the Hans approach is that people should code all of the 'stuff' they want to do in Lua and go LuaTeX-only, which is not ideal for someone writing a tool mainly used by pdfTeX users ...

@DavidCarlisle I guess it is fixable if someone wants to extend dvips, but that seems unlikely

@JosephWright possibly more likely is extend xdvipdfmx as there really can not be (or should not be) that much difference between luatex and xetex dvi(ish) output using otf fonts

9:30 AM

@DavidCarlisle Yes, true

@DavidCarlisle Not sure though that helps with the workflows here: don't they need dvips or similar (i.e. no PDF involvement)

@JosephWright no idea to be honest can't remember what asy is doing at all

@DavidCarlisle Based on asymptote.sourceforge.net I'd say 'making PostScript'

@JosephWright Unless somebody updates DVI and dvips so they can cope with OpenType fonts…

@JosephWright sure but in 99.99% of the cases isn't that postscript going to be distilled to pdf anyway

@egreg Even then, I'm not sure about the engine support for DVI mode longer-term

@DavidCarlisle I see Brexit needs an act of parliament

@DavidCarlisle Agreed

9:43 AM

@JosephWright Ah I was going to check the news around now (not that it will make any difference, apart from timing perhaps)

@DavidCarlisle Had the live stream on

@DavidCarlisle One and only time I'll watch a court verdict live!

@JosephWright maybe the PM will take it to the European courts and overturn these uncooperative UK judges:-)

@DavidCarlisle This has been suggested but depends on whether 'reversibility' of Article 50 is important. From what I heard, that's not the key so European law has no effect and there is no appeal possible.

10:03 AM

@DavidCarlisle I see Chris has woken up

@DavidCarlisle, @egreg (others) Fewer immediate issues with the kernel update than I feared we might get

@JosephWright Haven't updated yet

@JosephWright so I wonder how many people yet have actually tried lualatex or xelatex without fontspec.

@egreg :)

@egreg catch up old man!

@DavidCarlisle Us ;)

10:10 AM

@JosephWright Travis still unhappy?

@JosephWright and them?

@PauloCereda no not them, not even @egreg:-)

@DavidCarlisle ooh

@DavidCarlisle Yes, but I'm sorting that

@PauloCereda I blame you for all the mail I'm getting

10:13 AM

@DavidCarlisle what did I do?

@PauloCereda got Joseph using Travis

@DavidCarlisle oh no

@PauloCereda you should put your thesis in Travis CI with a fail condition that it doesn't have the required length, then it'll mail you every few minutes just to let you know it's still failing (in case you forget)

2

@DavidCarlisle Will be sorted soon: it's just a question of tracking down the few places that a simple switch to OT1/cmr doesn't 'fix'

@DavidCarlisle that's a very evil plan!

Is fontspec fixed?

10:17 AM

@PauloCereda Not yet

@JosephWright oh no

@JosephWright: regression bug?

@PauloCereda will posted the required patch in github issue so don't let that stop you writing your thesis...

@DavidCarlisle ~~what thesis~~ oh

Being in Australia, is it possible for Will to submit a patch and due to timezones, I apply a patch today in which Will wrote tomorrow?

10:33 AM

Hi everyone! Can we please reopen

2

Q: Unable to compile pdf_tex not on latex

I am trying to include a .pdf_tex file into my latex document. I saved the file from svg in Inkscape to .pdf_tex file. And I use this in my latex document: \begin{figure} \centering \input{myfile2.pdf_tex} \end{figure} myfile2.pdf_tex is in the same file as the main folder. But the error s...

graphics inkscape

I think it was closed as the duplicate of the wrong question.

@JosephWright @DavidCarlisle Here VVVV

This is pdfTeX, Version 3.14159265-2.6-1.40.17 (TeX Live 2016) (preloaded format=pdflatex)
 restricted \write18 enabled.
entering extended mode
(./ruiq.tex
LaTeX2e <2017/01/01>

@egreg now try lualatex small2e and get the all new TU lmr version of that document.

@DavidCarlisle I did! I tried \font\test={Arial} \test abc ;-). And it worked fine.

@UlrikeFischer you'll need to explain to @egreg what Arial is.

@DavidCarlisle the little mermaid.

10:44 AM

@DavidCarlisle I did and found the first bug

user image

@DavidCarlisle The one font name I can remember easily -- never got round to learn the name for Latin Modern.

@egreg hey novel idea that: looking at the typeset output. I only checked it didn't error.

@DavidCarlisle Another change needed in tuenc.def?

@egreg but why did it do that I wonder, same in xelatex.

@DavidCarlisle Yes, +tlig and mapping=tex-text are missing

10:48 AM

@DavidCarlisle Well you obviously didn't added the tlig ligatures when loading the fonts

@JosephWright er is that the place? don't we need to activate the ligature in the font loading, how come that didn't happen?

@DavidCarlisle Don't know: thought it was in @WillRobertson's original approach!

@UlrikeFischer yes I can see that but I thought we'd stolen the right bits of fontspec, need to check.....

@WillRobertson you around?

@DavidCarlisle You have mixed up the options: with XeLaTeX I get \TU/lmr/m/n/10=select font "[lmroman10-regular]:+tlig;" and with LuaLaTeX it is \TU/lmr/m/n/10=select font [lmroman10-regular]:mapping=tex-text;

Hello, I see I broke things again!

Where are we at?

10:51 AM

@WillRobertson no I think it was OK until I copied it to teh kernel and broke it.

Sorry I didn't catch it :(

@WillRobertson if you submit a patch today, given your timezone, will I install today a patch you wrote tomorrow? :)

@WillRobertson try xelatex small2e with the new format and there are no ligatures for tex quotes

For fontspec:
• Need to re-add some private macros to fontspec so xepersian loads?

@WillRobertson th ebase lmr fd file has \UnicodeFontFile{lmroman6-regular}{\UnicodeFontTeXLigatures} but presumably that's got the wrong definition, looking....

@WillRobertson oh that one, I had to add:

\ExplSyntaxOn

\cs_new:Nn \__fontspec_pass_args:nnn
 {
  \IfNoValueTF {#2}
   { \__fontspec_post_arg:w {#1} {#3} }
   { #1 {#2} {#3} }
 }
\NewDocumentCommand \__fontspec_post_arg:w { m m O{} }
 { #1 {#3} {#2} }
\ExplSyntaxOff

But more worried that '' doesn't ligature in the base format:-)

10:55 AM

@DavidCarlisle Well, true.

I think I get the prize for the worst bug of all though — in current fontspec, \emph doesn't do anything!

So yes I will be updating tonight :)

If only we could load expl3 in the kernel to avoid this sort of code:

\begingroup\expandafter\expandafter\expandafter\endgroup
\expandafter\ifx\csname XeTeXrevision\endcsname\relax
\begingroup\expandafter\expandafter\expandafter\endgroup
\expandafter\ifx\csname directlua\endcsname\relax
\PackageWarningNoLine{fontenc}
{\UnicodeEncodingName\space
encoding is only available with XeTeX and LuaTeX.\MessageBreak
Defaulting to T1 encoding}
\def\encodingdefault{T1}
\expandafter\expandafter\expandafter\endinput

@JosephWright @DavidCarlisle I've fixed the TU enc issue in tex4ht, all what was needed was to add \RequirePackage[OT1]{fontenc} at the right place, as Joseph suggested. Also basic Latin Unicode is supported out of the box, as in the PDF mode.

So the XeTeX=false conditional is first, but the else branch has the LuaTeX defn. Okay, just swap the two lines.

@michal.h21 Great news

@WillRobertson yep thanks. odd it isn't a syntax error rather than being silently? ignored nut at least easy to fix

I hope that we can fix some issues in the LuaTeX's DVI mode using node callbacks. There is a big issue with pictures, because DVI converters like dvipng or dvisvgm don't like unicode

@JosephWright @WillRobertson (we need some test files using TU :-)

10:59 AM

@WillRobertson thanks. BTW, we added support for previous Fontspec

Hello everyone. Is the correct way to left align the first column, and right align the second column to use:
\begin{tabular}{>{\raggedright}p{2.3in} >{\raggedleft\arraybackslash}p{2.2in}}
?

version, but it is broken in the current release

@WillRobertson Yes, just tried it out -- a good case to expl3 ;-)

@DavidCarlisle BTW I always make that mistake with \if\csname..\relax

@michal.h21 What can I do to help (if anything)? Either now or as things change in the future

@FaheemMitha yes or lr if you just want single line or perhaps use \RaggedLeft (and right) instead of \raggedleft

11:03 AM

@DavidCarlisle Yes, I saw the \RaggedLeft (and right) thing. What's the difference?

@DavidCarlisle The extended font interfaces have always been quite forgiving. Definitely arguments to make that it shouldn't be. fontspec itself is more rigorous and I suppose wouldn't be if the engine shouldered that aspect of things.

@WillRobertson I just need to figure out which macro I should redefine to do nothing. The problem is that tex4ht has a bug which causes it to break with Oepntype fonts. So we need to prevent loading of any OpenType font and use Lua callbacks or active characters for Unicode support.

@FaheemMitha it allows less raggedness so you get hyphenation, the default allows arbitrary stretch so will never hyphenate unless the word is linger than a line

@DavidCarlisle I see. Ok, I'll try it both ways. Thanks.

@WillRobertson in the previous version, we needed to redefine just \__fontspec_fontspec:nn to do nothing,

11:07 AM

@michal.h21 Ah, you should let me know about things like this and I can be more helpful :) To be honest if I were you I think it would be better to redefine all the "top level" commands like \fontspec and so on -- it means more redefinitions but generally speaking they shouldn't change underneath you.

@michal.h21 \__fontspec_fontspec:nn now has the far more sensible name \__fontspec_select_font_family:nn :)

@WillRobertson hmm

$ svn diff
Index: ltoutenc.dtx
===================================================================
--- ltoutenc.dtx        (revision 1348)
+++ ltoutenc.dtx        (working copy)
@@ -2703,7 +2703,7 @@
 %
 % LuaTeX.
 %    \begin{macrocode}
-    \def\UnicodeFontTeXLigatures{mapping=tex-text;}
+    \def\UnicodeFontTeXLigatures{+tlig;}
   \fi
 \else
 %    \end{macrocode}
@@ -2710,7 +2710,7 @@
 %
 % XeTeX
 %    \begin{macrocode}
-  \def\UnicodeFontTeXLigatures{+tlig;} % "+trep;" no longer needed
+  \def\UnicodeFontTeXLigatures{mapping=tex-text;}

rebuild format still no ligatures, will look later better do some work stuff now....

@DavidCarlisle Damn, I'll try to look immediately

@JosephWright @DavidCarlisle — Any thoughts on whether I should re-add those private macros to allow xepersian to load?

@DavidCarlisle I exchanged in tuenc.def the two settings and rebuild lualatex format and it worked fine (with luaotfload-release).

@WillRobertson no feeling, what do they do, or just mail and ask them to update, I'd be tempted to do the latter

@UlrikeFischer I now get

> \TU/lmr/m/n/10=select font "[lmroman10-regular]:mapping=tex-text;".
<inserted text> \TU/lmr/m/n/10

l.16 \expandafter\show\the\font

but no ligatures (with xelatex)

@DavidCarlisle that...doesn't seem right... I just tried \font\1="[lmroman10-regular]:mapping=tex-text;" at 10pt\relax {\1 ``abc''} and it was fine — but not in a 2017 build

(As in, my check wasn't in a 2017 build)

11:21 AM

@DavidCarlisle Work fine for me with This is XeTeX, Version 3.14159265-2.6-0.99996 (TeX Live 2016/W32TeX) (preloaded format=xelatex 2017.1.24) 24 JAN 2017 12:12

@WillRobertson @UlrikeFischer of course it helps if you look at the correct small2e.pdf after remaking a file of that name....

user image

@DavidCarlisle ;-)

@DavidCarlisle You win some you lose some :)

@WillRobertson No, they are private, unless there is some vital functionality that really should have a public interface but is missing

@WillRobertson today the latter. On which happy note I should go and apply some of my new found debugging skills to code here....

11:23 AM

@DavidCarlisle Sure thing!

@DavidCarlisle, @WillRobertson Down to the contrib test failures on Travis: I think the ctex002 one is due to the \emph bug in fontspec :)

19 hours ago, by David Carlisle

@PauloCereda just go to your thesis defence and say it's all Will's fault.

@WillRobertson ^^

@egreg thanks for that catch, fixed in svn

@WillRobertson I thought so :)

@JosephWright Nope, that was the hack code that you and egreg said I must absolutely remove :)

11:34 AM

@WillRobertson :)

@JosephWright It was the thing that implemented {mO{}mO{}} where only one of the optional arguments were allowed

@JosephWright So now it's just {mO{}mO{}} and if anyone complains... too bad

11:48 AM

Trying to re-add some of my old test files to fontspec... looks like font support for centered asterisk is spotty:

Missing character: There is no ⁎ (U+204E) in font file:texgyrepagella-regular.otf:mode=node;script=latn;language=DFLT;+tlig;!

Missing character: There is no ⁎ (U+204E) in font file:LinLibertine_R.otf:mode=node;script=latn;language=DFLT;+tlig;!

Bit of a problem considering

\newcommand\labelitemiii{\textasteriskcentered}

\def\@fnsymbol#1{%
\ifcase#1\or \TextOrMath\textasteriskcentered *\or
\TextOrMath \textdagger \dagger\or
\TextOrMath \textdaggerdbl \ddagger \or
\TextOrMath \textsection \mathsection\or
\TextOrMath \textparagraph \mathparagraph\or
\TextOrMath \textbardbl \|\or
\TextOrMath {\textasteriskcentered\textasteriskcentered}{**}\or
\TextOrMath {\textdagger\textdagger}{\dagger\dagger}\or
\TextOrMath {\textdaggerdbl\textdaggerdbl}{\ddagger\ddagger}\else
\@ctrerr \fi
}%

I guess we should (somewhere) change these definitions rather than expect people to add a weird unicode glyph to all their fonts…

@WillRobertson While you are here: there seems to be a \luatex_if_engine:T in unicode-math

@WillRobertson Causes the one outstanding failure for L3 testing (@DavidCarlisle)

@JosephWright Oh yes, a unicode-math update was next on my list — don't I also owe you a breqn update??

@WillRobertson Possibly: I don't have a test for that :)

12:05 PM

@WillRobertson is U+2217 better supported?

@JosephWright @WillRobertson we could change it before PL1....

@DavidCarlisle in those two fonts, yes — and looking at the definitions, the operator one seems to be a better match to the intention of \textasteriskcentered

@WillRobertson seems like it it's also &lowast; in html

@DavidCarlisle That's a shame considering U+204E is literally "LOW ASTERISK"

@DavidCarlisle I guess it's passé to complain about unicode naming these days

@WillRobertson yes but if it comes out literally as "missing glyph box" in your web page, it's not so good:-)

@DavidCarlisle :)

12:13 PM

Change?

@WillRobertson seeing as the current one seems to be missing even in the default lmr changing looks best to me

12:25 PM

@DavidCarlisle Agreed. Sorry that definition was inherited from Ross; not sure how it could have been "wrong" for so long without anyone noticing...

Oops, not it's not: \DeclareUTFcharacter[\UTFencname]{x002A}{\textasteriskcentered}

I guess this one IS my fault after all!

@WillRobertson while I was looking at the ascii mappings for the latex/4500 bug I also worried about a couple of definitions such as \DeclareTextSymbol{\textasciicircum} \UnicodeEncodingName{"02C6} U+02C6 is a circumflex, but it's not U+005E ....

@WillRobertson but that's * so not centred at all:-)

@DavidCarlisle Right — wrong in a different way!

Hmm, Ross had
\DeclareUTFcharacter[\UTFencname]{x005E}{\textasciicircum}% see also x02C6
\DeclareUTFcharacter[\UTFencname]{x02C6}{\textcircumaccent} % see also x005E

@WillRobertson seems a bit odd for the names with ascii in the middle to go higher than 127

@DavidCarlisle agreed.

\DeclareTextSymbol{\textasciicircum}     \UnicodeEncodingName{"02C6}
\DeclareTextSymbol{\textasciicaron}      \UnicodeEncodingName{"02C7}
\DeclareTextSymbol{\textasciitilde}      \UnicodeEncodingName{"02DC}
\DeclareTextSymbol{\textasciibreve}      \UnicodeEncodingName{"02D8}

@WillRobertson ^^^^

12:33 PM

Ross has a comment about this:
% changed \textascii<accent>, etc. to give non-accent chars
% 6 chars affected:
% \textasciitilde, \textasciiacute, \textasciigrave
% \textasciimacron, \textasciicircum, \textasciicedilla

I guess I can see the reasoning?

It would have been better to call them "nonaccentacute" and similar I guess

Is it normal for the word "Introduction" to end up as "INTRODVCTION" (note the uppercase V for U) in a chapter heading?

I wonder what TS1 really intended here

@Telemachus Sure, if you have a funny font :)

I am using Polyglossia, and some of my text is Latin, but the default is English. And I even added a specific request that this text be English.

@WillRobertson Hmm. I don't think it's a funny font, but who knows, I suppose. What would be a non-funny font to test against?

user image

The font I'm using is Baskerville, if that matters.

12:37 PM

@WillRobertson ^^^ libertine, the first is ^ the second is ^^^^02c6 probably I'd say the first is more expected for ascii ^ ?

Ok, I don't think it's the font. I just switched to "Times New Roman", which I assume is pretty vanilla. No joy. Now it's just an uglier looking "INTRODVCTION".

How can I force that V to be a U?

@Telemachus The question should more be, why is a U becoming a V?!

(Short of keeping it lowercase, which I'd like to avoid.)

@DavidCarlisle I suppose so, yeah

@WillRobertson Well, that is my question. But I have no answer to that, so I'm in damage-control mode! Happy to take suggestions to troubleshoot though.

12:39 PM

@Telemachus Impossible to say without seeing more. Can you reduce the problem down?

@WillRobertson I can try, though it's a fairly complex document. I'll see what I can do.

Okay, fontspec 2.5d uploaded — but I'm falling asleep so hopefully I didn't mess anything up

@WillRobertson What have the Romans ever done for Vs?

@WillRobertson here's a reasonably short gist. I left all the prelude material in since it's likely to be something I did wrong in there. The problem still appears: gist.github.com/5f333f3953f25c9948048068a1a6f13a

Okay: It's Polyglossia. If I remove the \setotherlanguage[variant=medieval]{latin}, then it's fine. But I kind of need that. Joy.

@Telemachus well you only need to lose %[variant=medieval] so it might be fontspec leaking the variant, get @WillRobertson to check when he's woken up:-) I didn't have baskerville so I just randomly picked Arial, not a very medieval font, sorry:-)

12:54 PM

@DavidCarlisle Ah, that's interesting. Thanks for checking so carefully.

I'll do that for now. I suppose Descartes is trying to write Classical Latin anyhow. ;)

Thanks so much.

@DavidCarlisle Does it strike you that I should file a bug in Polyglossia?

@Telemachus You can add \uccode`\u=`\U to your document (not sure if I get the markup correctly here

@Telemachus I suspect polyglossia is just deferring to fontspec for the low level font handling so my first thought would be to blame Will, not sure though. I'm an English speaking mathematician I never actually use any of this:-)

@UlrikeFischer Thanks. I'll try that.

@UlrikeFischer oh it's just the uccode? I thought it was setting a font feature thing.

@DavidCarlisle Ah, ok, so fontspec is the culprit (maybe) rather than polyglossia? Am I understanding that right? I don't know what @WillRobertson is responsible for. I'm newish here.

12:59 PM

@DavidCarlisle gloss-latin.def has a \classicuclccodes and imho it doesn't reset/group it correctly so the codes are wrong throughout the document.

@Telemachus I would guess without checking that @UlrikeFischer is right, in which case it might be a bit more delicate to fix, I thought it was using a font style variant to make U look like a V but if it's just declaring uppercase of u is V then the uppercase code is a global property so might be harder to detect which fonts happen to be in your page head and whether they need that property

@Telemachus what @UlrikeFischer says:-)

Ah, I see. But something like that \uccode declaration might help? I'm assuming I put that in the prelude, right?

@Telemachus fontspec is WillRobertson's baby

@Telemachus \uccode`\u=`\U just declares that the uppercase of u is U so it's not likely to do any harm:-)

@DavidCarlisle Why are you saying that uppercase code is a global property? {\uccode`\u=`\V \uppercase{uv} } \uppercase{uv} does what I expect.

@DavidCarlisle Ok. I added that, then put back the language variant and the problem is back. Sadly.

Would it matter where in the prelude I put the \uccode business?

1:06 PM

@Telemachus yes after the latin one has messed it up, just after \begin{document} might be best

@DavidCarlisle Cool. Retrying. tibi grātiās agō. (I.e., thank you. You get to learn a little Latin for your help!)

@Telemachus I don't think there should be so much accents in Latin sentences. :)

@UlrikeFischer oh sorry not global to tex grouping, but global to all fonts, so if you specify a heading is uppercase you normally do that without checking for font changes within the heading text, whereas if the uppercasing is done by small caps etc tehn of course it is sensitive to font choice

@DavidCarlisle @UlrikeFischer That does it. I had put it too high. Putting the \uccode fix later fixes the problem.

@PauloCereda I teach Latin for a living. ;) Those are macrons, used for helping beginning students learn which vowels are long, and which are short. They don't appear in more advanced texts (or on ancient inscriptions).

@Telemachus Ego loqui variis linguis, et ex Germanica in Ulrike

1:09 PM

@Telemachus You can also add to the head definition: \fancyhead[CE]{\uccode`\u=`\U \scshape\MakeUppercase{\leftmark}}. Or don't use \MakeUppercase in the heading but e.g. a small caps fonts.

@Telemachus oh sorry, I had no idea. :) My only Latin knowledge comes from the ordo missae. :)

@DavidCarlisle Actually lol-ing right now. Google Translate?

@PauloCereda would you care to guess an answer to the above question ^^^^

@Telemachus ~~yes~~ no, David is an expert in all known languages.

And because he's ~~very old~~ an expert, he also knows about ancient languages.

@DavidCarlisle: ^^ :)

@Telemachus As far as I can see, the default variant for Latin is classic. You can try \setmainlanguage[variant=modern]{latin}

1:12 PM

@PauloCereda No worries at all.

@egreg Yeah, the categorization of periods in Latin are not so clear to me in Polyglossia.

Anataes scribe theses bene

That is, Descartes is technically neo-Latin. There's no such category in Polyglossia. It's not (I think) modern Latin, so I picked medieval as the closest variant. That variant also fits Descartes's occasional non-classical spellings. E.g., coelum = caelum (sky, heaven)

@Telemachus Interesting — I saw those macrons and assumed because I was unfamiliar with them that my schoolboy Latin has simplified things for us :)

@UlrikeFischer Thanks for other tips about where to put it. I actually wanted to make the change globally since I don't want any other capital letter u to become V either.

@WillRobertson That may be right: that someone simplified. Some teachers leave them out to make things simpler for their students—or for themselves!

Which is fine, except for two things: (1) macrons matter a lot for scansion of poetry, and (2) the pronunciation does vary between say malum (a bad thing) and mālum (an apple).

@Telemachus I can't say I remember too much. I think the book series we learned from was called Ecce Romani

1:18 PM

But as there are no Romans around to get annoyed, I suppose the pronunciation doesn't much matter so very much anymore. Nearly every modern country has a more or less different set of beliefs about how to pronounce Classical Latin anyhow. It's a mess.

@Telemachus The once main prison in Rome is called “Regina coeli” (it used to be a monastery before becoming a prison).

Ecce! In pictura est puella. Puella nomine Cornelia. Cornelia cantat. [etc]

@WillRobertson I know that series. I've taught from it even. The most recent editions (American at least) do have macrons, as I recall. Maybe not earlier ones? Or the UK ones maybe not. (I have no idea how old you are, nor where you're from, however.)

@egreg Wow: "Queen of heaven". Very unfortunate name for a prison!

@Telemachus Interesting! I learned Latin in 1994–1996, I guess some people might call that "middle school".

(In Australia — I would guess we'd have the British editions.)

@WillRobertson Does it strike you that this u -> V business might be fontspec related? Or do you think rather Polyglossia's fault?

1:20 PM

Good to hear they're still around! I thought the conceit worked quite well

@Telemachus No, the others hit the nail on the head.

@WillRobertson Yup, it's a very popular series in the US. Second only to Cambridge Latin, I think, in overall use.

@Telemachus if it's uccode's it probably isn't Will's fault:-)

@DavidCarlisle Ah, ok. Whose fault might it be? That is, I don't want to throw blame, but I would happily file a bug if it might help others.

I left a comment on your gist (gist.github.com/telemachus/5f333f3953f25c9948048068a1a6f13a) — I think it would be worth reporting as a bug

@WillRobertson Ah, sorry. I hadn't seen that. Thanks!

1:22 PM

@Telemachus polyglossia, you were right first time

@Telemachus I'm rather confused what the point of this uc/lccode messing around with is, though

@Telemachus It's done by polyglossia: in gloss-latin.ldf there is \def\classicuclccodes{\lccode\V=\u \uccode\u=\V}, which is executed when variant=classic is chosen.

@WillRobertson Sorry? Confused about the point of my messing around or theirs?

@WillRobertson I think I might leave those ascii... names circumflex could be ^ but what would textasciicaron be?

Oh, I think I can answer that. The Romans only had one letter where we have two.

1:23 PM

@Telemachus Actually I take it back — like I said earlier, it's late :) Presumably for historic latin people are literally typing the correct symbols and you do want \MakeUppercase etc to work correcty

Same with I and J.

@WillRobertson Correct. Though it should actually apply to classical Latin as well.

That is, the poet's name would look like this on an inscription PVBLIVUS VERGILIVS MARO

@Telemachus “Regina coeli” or “Regina caeli”, depending on the spelling, is one of the attributes of the Virgin Mary: “Regina caeli laetare, quia quem meruisti portare resurrexit sicut dixit” is one of the prayers of Easter time.

@Telemachus Oh, right! In a different universe, this would actually be handled by the font (hence more of a fontspec thing) but that would require the font designers to take this sort of thing into account — which they rarely do

@egreg I'm embarrassed to say I didn't recall that, though I taught at a Catholic school for 12 years. I'm not Christian by background. Thanks.

@Telemachus @WillRobertson @DavidCarlisle: Even if I have some doubt that polyglossia will do in the next future something about it I made a bug report github.com/reutenauer/polyglossia/issues/172

1:25 PM

@UlrikeFischer thanks!

@UlrikeFischer Ditto what Will said. Thank you all very much. Most fun I've ever had talking about a software bug!

@Telemachus Getting off topic here — so "VU" is uppercase "uu" ? Or is "PVBLIVS" is what you meant?

@UlrikeFischer Out of curiosity, why do you doubt that they'll do much about it?

@WillRobertson Nope, that was my typo. Fingers automatically changed. It should be PVBLIVS.

Here's a nice little image with happy Vs for VERGILIVS: upload.wikimedia.org/wikipedia/commons/1/1b/Vergilius.jpg

@Telemachus Ah good, that is indeed more sane :)

@Telemachus I like the happy Vs :)

Though when they had VV they would in earlier inscriptions sometimes prefer to write VO rather than VV. As an example QVOM (which is a variant of an earlier version of the conjunction CUM with the initial C- represented by QV-).

There's a bunch of those QVOMs in this, for example: en.wikipedia.org/wiki/Senatus_consultum_de_Bacchanalibus

And now I'll shut up about all that. Thank you again, everyone.

1:33 PM

must... type... my... thesis... in... Latin...

@Telemachus Which is what the words with “uu” were pronounced, IIRC: “equus” was pronounced somewhat like it was written “equos”.

@Telemachus Very interesting though, thanks :)

Okay all, must depart. Good night

@Telemachus “qu” represented “kʷ”

@Telemachus Because the polyglossia maintainer is not very active currently. Quite a number of issues are open. Btw babel is doing it much better

\documentclass{book}
\usepackage{fontspec}
\usepackage[latin.medieval,english]{babel}


\begin{document}

\uppercase{sum} \MakeUppercase{sum}

\selectlanguage{latin}
\uppercase{sum} \MakeUppercase{sum}

\selectlanguage{english}
\uppercase{sum} \MakeUppercase{sum}
\end{document}

user image

@egreg Yup, that sounds right to me.

@UlrikeFischer It's been a while since I chose between Polyglossia and Babel, but as I recall with Babel my whole document wouldn't compile. It conflicted with something else, and I think I gave up. That said, if Polyglossia is stagnant, I may look at that again. Thank you.

Or maybe Babel didn't work as well with ancient Greek? (I have a little of that as well, though not much.) Too many wrinkles...

1:45 PM

@Telemachus Imho with your three languages babel should work fine. All have been adapted to lualatex/xelatex. And in case of problems you can get support.

Latin is difficult.

The new version of fontspec break \emph command

Is it a new bug ?

@Maïeul It is new but already known. There is an issue in the github tracker and it should be sorted out soon.

ah ! there is an issue track! that is nice, but the handbook don't mention it

Many thanks Ulrike

I will post a message on my website

@Maïeul Add an issue to the issue tracker that the docu should mention the issue tracker (if you can find the tracker ;-))

1:51 PM

@UlrikeFischer That's good to know re Babel. When I get a little time, I'll test out switching from Polyglossia to Babel and see how it goes.

@egreg @DavidCarlisle @JosephWright: Btw I made a feature request that miktex's pdflatex ignores the BOM sourceforge.net/p/miktex/bugs/2570. Wouldn't it also be good if the non-windows texlive would do this too?

@UlrikeFischer apparently, the correction commit is done. I hope the release will be done quickly.

@UlrikeFischer I have open an issue about issue https://github.com/wspr/fontspec/issues/255

1 hour ago, by Will Robertson

Okay, fontspec 2.5d uploaded — but I'm falling asleep so hopefully I didn't mess anything up

Thanks Ulrike!

I go back to my thesis

@Maïeul ooh good luck to you!

@DavidCarlisle ^^ see, someone is actually working on his thesis.

1:59 PM

@DavidCarlisle I'm doing some experiments for a set of TU-based parallel tests for the kernel: I think the Lua is simple enough

@JosephWright thanks, seems like they are needed:-)

@DavidCarlisle We might be able to cover it by adding a switch to l3build in the longer term (one recognising 'test sets' or similar): I'll get it working by hand then we can discuss

I just broke a record I think: 10000 views (and a golden badge) for a question of score 1.

3:00 PM

@Telemachus -- it's a little late in the season, but if you're teaching latin, you might enjoy this: youtube.com/watch?v=x8qB-PUw2AQ (i have the words, if you're interested.)

@barbarabeeton Thank you. I'll take a look.

@UlrikeFischer Is there a way for a user to find out what attributes different languages have in Babel? I'm not seeing a list in the Babel documentation. latin.medieval is mentioned in an example, but not otherwise. (I'm wondering more about Greek attributes.)

@UlrikeFischer Ah, wait, I think I see. The languages are in CTAN separately under babel-contrib. Greek's documentation dates to 1997. Not so encouraging a sign, but we'll see.

@Telemachus The language files have their own maintainer. You get e.g. the docu for latin with texdoc babel-latin. It mentions four attributes medieval, withprosodicmarks, classic, and ecclesiastic

@UlrikeFischer Yes, sorry. I took a minute to catch up on that. Unfortunately, the babel-greek seems out-of-date. It doesn't mention the polytonic option that the main babel documentation does. Anyhow, I'll try it out again. Thanks.

3:15 PM

@Telemachus My babel-greek documentation is from September 2016 and mentiones polytonic.

@UlrikeFischer Huh. Welp, that sounds like I made a mistake. I was looking on CTAN rather than locally. I looked at the Usage Documentation here ctan.org/pkg/babel-greek

@Telemachus Sounds as if CTAN made an error. babel-greek.pdf is in the file list and is the current documentation. Usage.pdf is really a bit outdated ;-). Make a report (there is a feedback button on the page).

3:38 PM

@UlrikeFischer Will do. Good news is that I was able to switch to Babel without much trouble. Now I just have to test how well it works compared with Polyglossia (in terms of hyphenation, etc.)

Thanks.

@UlrikeFischer Reported.

4:11 PM

@yo' Link?

4:34 PM

@yo' Tracked it down ;)

4:45 PM

@DavidCarlisle Given you somewhere to start collecting up TU tests :)

@yo' What site?

@yo' Oh, I found it! They're strong voters, it seems.

@JosephWright @egreg sorry I was away...

5:09 PM

@JosephWright yes thanks. Do you want to do the ctan dance for pl1 or should I risk it this time?

@JosephWright yeh posted to c.t.t just like old times (apart from not using emacs/gnus)

@JosephWright Yes, it is another world. By the way, the students really don't like ChemDraw.

5:25 PM

@cfr No? Prefer ChemDoodle, ChemSketch, ...? ChemDoodle is OK and if we didn't have a site license for ChemDraw it's what I'd pay for (much cheaper)

 Tree conflicts: 1
Tree conflict on 'build-TU.lua'

@JosephWright you just got there first:-)

@DavidCarlisle Give me 1 minute

@DavidCarlisle l3build now has a -t option

@JosephWright I guessed. Actually you've got 30, I'm driving home:- now:-)

@JosephWright ChemDoodle /giggles

@PauloCereda chemdoodle.com

5:29 PM

@JosephWright I am laughing more than I should with this name...

@cfr chemdoodle.com/wp-content/uploads/2014/08/macWord-150x150.png shows nicely some of the reasons a GUI is needed for chemistry structures (getting the 'front' ring right), plus the business about bond mitring (see the 'wedged' bond joining the 'stick' one)

@JosephWright: Animal Crossing has a place named Shampoodle: animalcrossing.wikia.com/wiki/Shampoodle

6:26 PM

! \hbox(7.5+2.5)x45.84
--- 1,7 ----
! \hbox(7.5+2.5)x45.84, direction TLT

@DavidCarlisle Saw that

@JosephWright seem to have lost the luatex normalisation

@JosephWright ah OK:-)

@DavidCarlisle LuaTeX wasn't quite right when I set things up: need to work a bit more on it

@JosephWright you mentioned it wasn't working but it looked like it was so I tried it anway, it normalised on save but not on check it seems, can back luatex out of the checkengine lists if you prefer

@DavidCarlisle Nah, don't worry: we know this is a bit of a work-in-progress

@DavidCarlisle I think it will all be sorted within the next couple of hours

6:35 PM

@JosephWright looks like I'd better add some discretionary tests:

5

Q: Letters after discretionary disappear (seems an issue of package fontspec)?

For the following case, % !TeX program = xelatex \documentclass[12pt]{article} \usepackage{fontspec} \begin{document} \hsize =2in shelf{}ful shelf{}ful shelf\-ful di\discretionary{f-}{fi}{f\kern0ptfi}cult \end{document} I get, Actually, if there's a kern followed by anything in the thi...

xetex fontspec hyphenation ligatures

@DavidCarlisle Aye

7:03 PM

My latest attempt at math in tikz:

in Mathematics, 3 mins ago, by Danu

user image

Dr. Manuel Kuehner

@DavidCarlisle You recently answered one of my questions (tex.stackexchange.com/questions/348827) about the graphicx package.

@DavidCarlisle I would have a follow-up question - maybe you can have a look at it.

7:22 PM

@JosephWright actually as there is a luatex.tlg isn't not doing all the luatex normalisation expected?

@Dr.ManuelKuehner you mean you edited the old question or there is a new one?

Dr. Manuel Kuehner

I aopened a new one - it was too mich change. I send you the link - wait a min

@DavidCarlisle tex.stackexchange.com/questions/350258

@DavidCarlisle If it's very complicated then just say it - it would be great to have that but if not then I will find a manual solution.

@Dr.ManuelKuehner posted answer:-)

@egreg do you want to do a xetex sourceforge bug, or should I?

@Dr.ManuelKuehner example-a.png.txt or example-a.txt ? (first is easier)

Dr. Manuel Kuehner

7:52 PM

@DavidCarlisle example-a.txt would be easier to handle (didn't see the chat notification).

@DavidCarlisle But I am happy with both. So far I have to do it manually.

@DavidCarlisle I have to go and buy some food - I'll reply later. Thanks for the help. This feature really helps me to save time and have a better quality in my presentations (when I distribute them for example)

8:25 PM

@DavidCarlisle Please, go on

@egreg Ok

2 hours later…

10:43 PM

@DavidCarlisle @JosephWright what do you think about the recent court brexit statement?

@yo' interesting historical case affirming the power of the courts over parliament but will no make no practical difference to the actual process of brexit

@DavidCarlisle well, isn't it the power of the courts over the government in this case?

the court didn't diminish the authority of the parliament, FWIV

@yo' well same thing really, parliament has final say because the courts said so. But they'll rush a vote through and only a handful will vote against so it'll all be back on track by next week....

@DavidCarlisle well, I have no idea on which side the parliament is currently, and I have no idea how much MPs feel bond by the referendum. Here, they wouldn't...

@yo' the vote will end up being 600-50 in favour. (Even though a majority of MPs campaigned to remain)

10:54 PM

@DavidCarlisle ah ok. Well, that's one of the things that really makes UK different from my country. (Good we at least share similar sense of humour.)

« first day (2277 days earlier) ← previous day next day → last day (2662 days later) »

Transcript for

Jan '1724

$TeX - LaTeX$ TeX, LaTeX and Friends

\begin{chat}

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

$TeX - LaTeX$

site design / logo © 2024 Stack Exchange Inc; legal

mobile