TeX, LaTeX and Friends - 2014-05-01

« first day (1278 days earlier) ← previous day next day → last day (3657 days later) »

12:03 AM

@DavidCarlisle: I like it. I feel better. Thanks David. Phew, you almost killed Karl. :)

3 hours later…

3:06 AM

Where can I find the code that defines the behavior change.case$ for .bst files? Shouldn't it be easy to add an option to "u", "t", and "l" that does headline case? </naivete> I'm sure seeing the source will change my mind...

3 hours later…

5:37 AM

@repurposer texdoc tamethebeast at the Command Line/Terminal will bring up the best docs there are on BibTeX

@repurposer Title casing is at best language-dependent and poorly defined (e.g. in English different publishers have a different list of ignored words). As such, it's actually very hard to do programmatically, and BibTeX doesn't offer it.

I am who I say I am

6:08 AM

user image

2 hours later…

8:20 AM

@JosephWright is it possible with biblatex/biber to use alternate style dependent titles in the bib file with various capitalizations?

@StrongBad Yes: you can do field manipulation in Biber

8:58 AM

@repurposer it's in the (web= documented pascal) sources of the bibtex program, so you would have to change that and then run web2c to get C and compile the generated C to get a modified executable.

@JosephWright have you read the "extensive" documentation about file encoding support in the xetex manual (section 4) :-)

@DavidCarlisle I believe that only Ulrike has made a sense out of it.

@egreg It turns out that if you give it a malformed utf8 stream (eg it's iso-8859-2) lualatex stops with ! String contains an invalid utf-8 sequence but xetex magically carries on, interpreting the bytes as latin1. This would produce nonsense (as the new inputenc warns you) but the old inputenc fluked its way to give active definitions to the mis-interpreted bytes and it sort of worked in xelatex (but not lualatex)....

@DavidCarlisle I believe this feature is used in babel-greek.

@egreg so does babel-greek not work in luatex or does it do something different there?

@DavidCarlisle I've stopped trying to understand it. ;-)

9:09 AM

@egreg good plan, perhaps I should do the same...

@DavidCarlisle Lines 324-343 of greek.ldf, where ^^9f is used

@egreg hmm at least it is entered by number not as a literal character

9:48 AM

@David: are we already at plan B? :)

@PauloCereda Yay! :-)

Of course the real solution is to ban spaces in filenames ;-)

Frank Mittelbach

@PauloCereda plan B for what?

does anybody know if the latest version of xetex (the one after TL2013) detects invalid utf8 files? (like luatex does?

@DavidCarlisle @egreg but the situation with xetex is worse ... it magically drops all major accents from latin1 encoded files

10:05 AM

@FrankMittelbach With this file

% -*- coding: latin-1 -*-
\documentclass{article}
\usepackage{fontspec}
\begin{document}
àèü
\end{document}

I get in the log file

Invalid UTF-8 byte or sequence at line 5 replaced by U+FFFD.
Invalid UTF-8 byte or sequence at line 5 replaced by U+FFFD.
Invalid UTF-8 byte or sequence at line 5 replaced by U+FFFD.
Missing character: There is no � in font [lmroman10-regular]:mapping=tex-text!
Missing character: There is no � in font [lmroman10-regular]:mapping=tex-text!
Missing character: There is no � in font [lmroman10-regular]:mapping=tex-text!

And nothing shows on the PDF file

Tried with TL 2013 and 2014

@FrankMittelbach David had a lot of plans. :)

Frank Mittelbach

@egreg and that is lualatex I guess?

what happens with xelatex?

@FrankMittelbach No, XeLaTeX

@NicolaTalbot I've been talking to the Apache Commons folks, so we are all investigating the issue. :) Meanwhile, they gave a suggestion of another library in which the error does not hold. :)

Offtopic: Firefox 29 reached Fedora repos.

@FrankMittelbach With LuaLaTeX I get

! String contains an invalid utf-8 sequence.
l.5
  ???

Frank Mittelbach

10:13 AM

@egreg ahh I see so luatex makes it an error and xetex just records it in the log

@FrankMittelbach On the other hand, if I add \XeTeXinputencoding="latin1", the file is compiled correctly.

@FrankMittelbach I could ask on the xetex list....

Frank Mittelbach

@egreg is that just a xetex thing? or is this understood by luatex as well?

@PauloCereda That's great. More seriously, it's not just an issue of spaces in filenames, as I think it more generally affects arguments with spaces, which some of my applications needs (such as jmakepdfx which tries to pass some postscript code to ghostscript).

@FrankMittelbach With LuaLaTeX there's a luainputenc package.

Frank Mittelbach

10:17 AM

@egreg i meant as some sort of primitive

@FrankMittelbach its a xetex primitive

Frank Mittelbach

@DavidCarlisle i guessed that from the name, just wondered if somehting similar exists for lua

but i guess the luainputenc is doing it differently

@FrankMittelbach I couldnt see one but I think the model there is not to add primitives if you can do it in lua

@NicolaTalbot I've made a couple of experiments and found out that everything goes right until the proper call inside the library. I was thinking to rewrite the tool for v6, and we could probably include a test suite with most of our use cases, so we can ensure (most of) all situations are covered. :)

Frank Mittelbach

anyway ... plan c) ... what encodings are supported by the xetex primitive and by luainputenc?

10:19 AM

@FrankMittelbach the documentation for \XeTeXinputencoding consists of one line in section 4 of texdoc xetex so it's not exactly clear what it supports

@FrankMittelbach This version goes through with both XeLaTeX and LuaLaTeX

% -*- coding: latin-1 -*-
\documentclass{article}
\usepackage{ifluatex}
\ifluatex
  \usepackage[latin1]{luainputenc}
\else
  \XeTeXinputencoding="latin1"
\fi
\usepackage{fontspec}
\begin{document}
àèü
\end{document}

Frank Mittelbach

because depending on how the mapping is one could get rid of the error and just pass the right encoding to the right subsystem

@egreg and what for latin9 .... jadajada

@FrankMittelbach It's good with both:

@FrankMittelbach yes except we claim you can switch \inputencoding anywhere and write toc files (which use an encoding neutral form) etc and if you make everything catcode 11 and just switch \XeTeXinputencoding things are just different

% -*- coding: latin-9 -*-
\documentclass{article}
\usepackage{ifluatex}
\ifluatex
  \usepackage[latin9]{luainputenc}
\else
  \XeTeXinputencoding="latin9"
\fi
\usepackage{fontspec}
\begin{document}
àèü€
\end{document}

user image

Frank Mittelbach

10:23 AM

@egreg I'm not going to ask you for the other 20 code pages :-) there must be a different way :-) ... currently looking at luainputenc.sty ...

FF29 looks a lot faster for me.

do you need the explicit \ifluatex there? luainputenc has: \ifxetex
\DeclareOption{unactivate}{\PassOptionsToPackage{utf8}{xetex-inputenc}}
\DeclareOption{lutf8}{\PassOptionsToPackage{utf8}{xetex-inputenc}}
\DeclareOption{lutf8x}{\PassOptionsToPackage{utf8}{xetex-inputenc}}
\DeclareOption*{\PassOptionsToPackage{\CurrentOption}{xetex-inputenc}}
\ProcessOptions*
\RequirePackage{xetex-inputenc}

@DavidCarlisle Possibly not.

@FrankMittelbach If you do strings $(which -p xetex)|less and look for latin1, you'll find hundreds of encodings understood by XeTeX. From UTF-8 on line 28254 to latin10 on line 31436.

@PauloCereda That's great. Can you include tests with a single quote at the start or end of the string "'foo' bar" or "foo 'bar'" (they're causing problems for the SQL statement option in datatooltk).

10:42 AM

@NicolaTalbot Sure. :) Could you test something for me?

@PauloCereda Sure, what d'you want?

@DavidCarlisle Do you have a xetex-inputenc package? ;-)

@NicolaTalbot A good way to see who's the culprit (thanks to the guys at Apache!) is to create a fake datatooltk file with the following code:

#!/bin/bash
echo "Current directory (\$PWD):"
echo "$PWD"
echo
echo "Arguments:"
while [ $# -gt 0 ]
do
    echo "$1"
    shift
done
echo "Done."

Then in the rule, go with the full path for /home/nicola/datatooltk which is our script. :)

We can see if the arguments are being properly expanded.

@egreg github.com/wspr/xetex-inputenc

@DavidCarlisle "I hope that it will make it into TeX Live 2009". :)

10:46 AM

@egreg not everyone's hopes come true

@PauloCereda Okay, I'll give that a try.

@NicolaTalbot Thank you! :)

@PauloCereda Okay. Here's the test directive:

% arara: datatooltk: { output: customer.dbtex, sqldb: samples, sqluser: sampleuser, sql: "SELECT * FROM people WHERE surname='Parrot'"}

@NicolaTalbot Fingers crossed!

And here's the output:

[nlct@dickimaw test]$ arara test
  __ _ _ __ __ _ _ __ __ _
 / _` | '__/ _` | '__/ _` |
| (_| | | | (_| | | | (_| |
 \__,_|_|  \__,_|_|  \__,_|

Running DatatoolTk...

Current directory ($PWD):
/home/nlct/latex/test

Arguments:
--output
customer.dbtex
--sql
"SELECT * FROM people WHERE surname='Parrot"
--sqldb
samples
--sqluser
sampleuser
--noconsole-action
gui
Done.
Status: SUCCESS

10:52 AM

@NicolaTalbot Aha! It's the YAML parser.

@egreg what do you get with lualatex with your latin9 example modified to

% -*- coding: latin-9 -*-
\documentclass{article}
\usepackage{ifluatex}
\ifluatex
  \usepackage[latin9]{luainputenc}
\else
  \XeTeXinputencoding="latin9"
\fi
\usepackage{fontspec}
\pagestyle{headings}
\begin{document}
\tableofcontents
\section{àèü€}
àèü€
\end{document}

@Nicola: Thank you, now I can investigate more precisely. :)

@PauloCereda should have used xml

@PauloCereda That's great. :-)

@DavidCarlisle I'd need a parser. :)

10:54 AM

@PauloCereda ctan.org/pkg/xmltex

@DavidCarlisle I get good results with XeLaTeX, bad with LuaLaTeX

@DavidCarlisle I was expecting that. <3

@egreg yep same here, thanks for confirming.

@FrankMittelbach ^^^

@PauloCereda There's a java xml parser :-)

Frank Mittelbach

@DavidCarlisle why is that?

10:56 AM

@NicolaTalbot shhh, I don't want to play with XML. :)

Frank Mittelbach

ok maybe I don't wanna know

@DavidCarlisle XeTeX writes off in UTF-8 and also inputs in UTF-8, by default.

@PauloCereda :-)

Only the first file is read with the specified encoding. That's what \XeTeXdefaultencoding is for. There's a question on the site.

@NicolaTalbot Actually, we do play with it when generating the hashes in arara.xml. :P

Frank Mittelbach

10:57 AM

@egreg that should be ok if LaTeX got down to use LICRs as those are deliberately 7bit

@DavidCarlisle tex.stackexchange.com/questions/153833/…

@FrankMittelbach Well, XeTeX doesn't do that, unless you activate the characters.

@FrankMittelbach it's as I mentioned before inputenc writes files in ascii LICR so it doesn't matter much what encoding is in force when you read things back, but if you don't do that you need a good tie between the files you write and the encoding you use to read

Frank Mittelbach

@DavidCarlisle I remember that from the times we designed the LICR stuff for NFSS

@FrankMittelbach @DavidCarlisle When LuaLaTeX is used on David's example, the second compilation succeeds, because the characters are indeed made active and changed into their LICR.

Frank Mittelbach

@egreg and the first did not?

11:00 AM

@FrankMittelbach The first compilation didn't succeed, because the .toc had been written by XeLaTeX.

Frank Mittelbach

luainputenc is using LICR concept in the background as it more or less emulates inputenc

@egreg ah unfair :-)

@egreg ouch sorry about that, that was an unfair test for luainputenc:-)

@PauloCereda I'm working on a new version of jpgfdraw at the moment, but I want to change the name, so what I'll probably do is make jpgfdraw obsolete and upload the new version as a new replacement application (otherwise it'll cause a problem for CTAN). Unfortunately I can't think of a snappy name for a GUI vector graphics application that generates pgf, flowfram.sty, \parshape and \shapepar code.

@NicolaTalbot vdraw? :)

@PauloCereda Already exists. As, I think, does every variation of <char>draw.

11:04 AM

navdraw -> Nicola's awesome vector drawing tool.

@PauloCereda Ooh! :-)

No, there's already a website by the name.

:(

I thought about flowframtk but that only emphasizes the connection with flowfram.sty.

@NicolaTalbot notjustflowframtk

@DavidCarlisle :-) or flowfram-pgf-parshape-shapepar-toolkit

11:11 AM

@Nicola: you could call it florist. :)

@PauloCereda Ooh, a cryptic name :-)

@NicolaTalbot :)

user image

3

@egreg yay, more palindromes!

Stefan Kottwitz

@IamwhoIsayIam I took the first 9 PSTricks examples from CTAN , intended for demonstration, because there were published under the free LPPL license. Though I referred to the LPPL, linked to the CTAN source and stated @Herbert as being the author, he critizized that I showed those examples.

@IamwhoIsayIam So I removed his examples, that's why the gallery is empty for now.

I thought it would be a nice idea to make pstricks-examples-en available like on texample.net, with an index of a lot of tags and nice gallery browsing, but it's not desired, so I can save that huge amount of work.

11:45 AM

@egreg boring: go away and earn another 55561

I am who I say I am

12:41 PM

@StefanKottwitz Oh I see. Btw, the most ideal site (or web application) as TikZ or PSTricks (or others) gallery (in my opinion) should provide the following features (that are similar to SE): 1. Membership to login/logout, 2. Each member can post a new diagram or edit the existing diagrams.

3. Most up-voted diagrams can be arranged in a grid form with paging buttons to ease searching. 4. Each diagram can have zero or more code-and-output revison, the most up-voted revision wins temporarily because someday it might be changed as the time goes.

5. Each diagram belongs to at least one tag from each group. For example, in "field" group there are mathematics, physics, chemistry, biolgy, etc tags, in "package" group there are pst-plot, pst-eucl, pst-solides3d, etc, tags in "action" group there are clipping, filling, plotting, overlaying, etc, tags in field "..." there are nodes, coordinates, etc tags. TeXexample.net is good but it can be improved with such features.

2 hours later…

2:26 PM

@HeikoOberdiek That's my speciality, ask David in chat. ;-) — egreg 16 hours ago

@egreg, @DavidCarlisle: ^^^ :)

!!/battle

@egreg wait wait wait. :)

@PauloCereda sigh don't bother

!!/battle

@PauloCereda Psmith, the TeX bot: The current score is egreg 215 vs. 160 David. So far, egreg is winning.

@PauloCereda vv

@MCH oh no, how am I ever going to catch up with egreg:-) — David Carlisle 7 hours ago

2:33 PM

@DavidCarlisle ooh! :) I just upvoted you. :)

2:45 PM

Hello from BachoTeX!

2

@topskip: Patrick! :) <3

@topskip Yay!

(I have to type fast, because the internet connection breaks every now and then)

There was an expl3 discussion yesterday evening. I wish someone like @JosephWright was here.

@topskip GM?

@egreg yes

@topskip So, was it in Polish?

2:53 PM

No, english. And it wasn't a talk, it was a discussion (a rather good discussion)

But there aren't any l3 experts here.

(many kids here, much beer and such a nice site - I went swimming yesterday - for a few minutes during lunch break)

@topskip Judging from the abstract, GM seemed not very well disposed towards L3.

@egreg True, but the discussion turned in a "towards l3" direction.

There was a nice bonfire yesterday with beer and sausages.

(the internet here connection is a game of luck)

Stefan Kottwitz

3:19 PM

@topskip Do you might have time to post a conference report blog post?

3:51 PM

@StefanKottwitz Not sure yet. I am currently skipping the workshops to do some work :(

@topskip Huh?

@topskip Ah, at BachoTeX

@JosephWright yes

@topskip I'm afraid I had a prior commitment (away with family)

@egreg Link?

@JosephWright gust.org.pl/bachotex/2014/program

Grzegorz Murzynowski, Warszawa, Polska
The expl3 bundle from a complainer's point of view. An informal workshop/BOF session.

We'll be considering (armed with examples wherever possible) the following questions (and maybe some other): What is it for? Is Plain/LaTeX 2e/ConTeXt not enough anymore? Are we able not to use expl3? Who needs yet another interface? What naming conventions are introduced? What is that excentric concept of “functions” that are claimed to “return” something? Is expl3 useful at all? Has anyone of us used it with a slightest success?

@egreg Thanks

@egreg Not like anyone has looked at this before and found that you run into issues with e.g. the LaTeX2e approach

@topskip I hope it was all amicable

4:07 PM

@JosephWright yes it was.

Contrary to what some people might believe, I was arguing much pro l3, as I've seen quite some people here adopting l3

@topskip Thanks :-)

(a lot of topics were covered: naming conventions, stability (how mature is expl3), who uses l3, what to expect from the future etc.)

Frank Mittelbach

4:26 PM

@topskip who discussed what?

@FrankMittelbach see the expl3 abstract here: gust.org.pl/bachotex/2014/program

@DavidCarlisle egreg quoted it a few posts up, as well.

Frank Mittelbach

@DavidCarlisle it doesn't havecontent other than the abstract which is a bunch of provoking questions but I would be more interested in the answers - or am i missing something

@FrankMittelbach The answers were not written down, as far as I know.

@FrankMittelbach yes I just passed by and thought you were asking what the discussion was but then saw the link had been posted and you really meant who was there discussing it

4:32 PM

... and it was not an "informed" discussions, more of people who have heard of something about l3

Frank Mittelbach

@topskip so why was it then a "good" discussion if nobody knew what they where talking about? :-)

@FrankMittelbach good is always subjective :)

Frank Mittelbach

I'm getting invited each year and i would love to come one day but May time is not the best and so far it never worked out

@FrankMittelbach its my first time here

Frank Mittelbach

@topskip sure but it would be more helpful if some notes have been taken (whether informed or not ... it gives a basis for dicussion or even correction)

4:36 PM

@FrankMittelbach perhaps there is a video recording of it... I have summarized the topics above (chat.stackexchange.com/transcript/message/15278065#15278065) - not much. We had a look at some source code, IIRC fontspec and we were discussing naming scheme and if it makes sense to us mere mortals or not

4:48 PM

Hello everyone! Does anyone know, how to typeset the DANTE logo correctly (correct font, etc.)?

5:05 PM

@PauloCereda I've answered a question about parskip telling my opinion: it makes white stripes on the page. You may enjoy the final note that not all black and white strips are bad.

Stefan Kottwitz

@HenriMenke \usefont{OT1}{dante}{m}{n}\selectfont DANTE

@HenriMenke See tug.org/TUGboat/tb29-1/tb91hoeppner.pdf

Stefan Kottwitz

Here's the font: ftp.tug.org/texlive/historic/2011/packaging

@StefanKottwitz Why is it not part of TeXlive?

Stefan Kottwitz

@HenriMenke And here the logo and related things are defined as commands: dtklogos.sty

@HenriMenke I don't know, perhaps it's a license issue

@HenriMenke at least the dtk package is part of TeX Live

@HenriMenke Did you attend the recent DANTE meeting?

5:15 PM

@StefanKottwitz Unfortunately not :(

But I'm planning to attend in autumn in Karlsruhe

@StefanKottwitz BTW: I wrote you an E-Mail some days ago. Maybe you ignored it because the subject was "TeXwelt meta" and probably you receive tons of these (I should have chosen another one).

@StefanKottwitz This might also be interesting meta.osqa.net/questions/7400/… I'd love to have a mobile view for TeXwelt in the future.

Stefan Kottwitz

@HenriMenke I read it, I was just a bit busy and did not answer yet

@HenriMenke I found a responsive layout, I'll try to find again, a second

@HenriMenke I found it - have a look at domande.curvedicrescita.com , resize the windows

@StefanKottwitz This is also great. Any mobile view would do it for me. In principle I only read questions and star them for answering later when I'm on my phone.

Stefan Kottwitz

5:34 PM

@HenriMenke I will see if I can work it out!

5:54 PM

@egreg ooh I love the image! :)

6:09 PM

@PauloCereda In one hour the match is starting.

@egreg Oh can we be champions today?

@PauloCereda No, Juve-Benfica, the return match.

@egreg Ah Europa!

1 hour later…

7:17 PM

What are the odds of getting a green avatar and have a name like:

1

datetime programming

@Werner LOL!!!

7:39 PM

@PauloCereda Oh, I've just had an idea. I could call it macawdraw and get it to write an arara directive if you use the export to LaTeX document function :-)

8:20 PM

Hey, it seems that while I've been gone I hit 5,000 karma. In other terms, what @egreg does on his lunch break.

9:02 PM

@Canageek :) A smile was just what I needed.

@PauloCereda :(

Hi everyone. I need to keep a list of files (max of around 500) along with the path. Should I just save it in a comma separated list using \g@addto@macro, or should I use the datatool package, or something else? Just thinking about any limitations I might run into.

@Werner :-) Maybe he changed the name based on the avatar he got as that it what it reminded him of :-)

9:23 PM

@egreg Glad to help. I take it things aren't going well?

@Canageek Juventus got kicked off the final.

@Canageek Nothing too serious. ;-)

@egreg Oh right, we've got some people in our lab who have been unhappy (One from the UK, who did his Masters in New Zealand)

@egreg :(

@PauloCereda Let's wait for good news next Monday.

@egreg Ah indeed. Another championship. :)

9:36 PM

@PeterGrill I think I'd probably just use etoolbox's list macros (\listadd etc). Are you intending to iterate through the list or use some fancy look-up system?

@NicolaTalbot: I'm having other crazy ideas about arara. My brain needs some sort of overclocking. :)

@PauloCereda Oh my! It's going to be a giant amongst all other parrots! :-)

9:59 PM

@PeterGrill My preference would go to a LaTeX3 sequence, more efficient than a comma separated list. Or to an etoolbox list, which is based on a similar principle.

@PeterGrill what are you going to do with the list? it's typically more efficient to store it as an \@elt list which you just execute than a comma separated list that you iterate over with \@for or similar.

10:17 PM

@DavidCarlisle \seq_map_function:NN ;-)

10:32 PM

@egreg you say tomato I say tomato

2

@DavidCarlisle And my last vote went to your \cr answer.

Although the site is full of \cr from you. ;-)

@egreg \cr...ap. There, I said it. :-| We love you @DavidCarlisle!

10:51 PM

@egreg not in the middle of an ams alignment:-)

« first day (1278 days earlier) ← previous day next day → last day (3657 days later) »

Transcript for

May '141

$TeX - LaTeX$ 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