TeX, LaTeX and Friends

Paulo Cereda

00:22

@yo' Archery? :)

Gonzalo Medina

02:26

This one can be closed, since it was a problem using an outdated version of listings (see OP's last comment):

0

Q: issue with listings package

I am trying to use listings package with Miketex 2.9 but when I try to compile using pdflatex I get the following error: ! Package caption Error: For a successful cooperation we need at least version (caption) `2004/02/13 v1.2' of package listings, (caption) but onl...

listings

Paul Gessler

@GonzaloMedina voted; the saga has ended then! Haha, I guess you saw the long string of comments :-)

Paulo Cereda

07:57

São Paulo awaits. :)

michal.h21

09:08

Hi all. I have possible duplicate:

2

Q: Can equation resolution in TeX4ht be increased

When TeX4ht compiles my latex code it generates a png of my equation and includes this in the html file. Until a better alternative for this comes along I am quite happy with this way of managing/presenting my equations. I would however prefer to have a higher resolution image that my collabo...

equations tex4ht html resolution

is duplicate of:

7

Q: LaTeX --> (X)HTML with tex4ht: bad quality images of equations

I'm using tex4ht to produce XHTML files from a LaTeX file. The XHTML file I obtain uses images for some equations which tex4ht could not convert directly to XHTML. Problem: images of equations have very bad quality. What I want: Better quality images. More info on attempts & system: I'm usi...

tex4ht

yo'

09:38

@PauloCereda well, yes of course I mean this...

yo'

09:55

@michal.h21 zapped, a good catch!

egreg

10:22

@JosephWright The \mskip bug doesn't show with LuaLaTeX, probably because of the other bug (wrong setting of the mu unit).

@JosephWright I suspect that something that should be global is not.

\documentclass{article}
\usepackage{unicode-math}
\begin{document}
{\tiny$a$}

$1\mskip20mu1$

$1\mskip20mu1$
\end{document}

@JosephWright Even smaller ^^^

David Carlisle

10:50

@egreg mskip bug?

Joseph Wright

11:22

@egreg I guess that (but the code may be very different anyway)

David Carlisle

@JosephWright the comment under @PauloCereda's blog, is that spam? (If not, what is it?)

Joseph Wright

@DavidCarlisle Spammed

David Carlisle

@JosephWright thanks

Joseph Wright

@egreg Not sure if that's the case: you can add e.g. \normalsize$\relax$ after your {\tiny$a$} and the problem does not go away. It's only solved if you insert some text in math mode.

karlkoeller

11:40

@DavidCarlisle thanks for the edit.

David Carlisle

@karlkoeller :-)

Paul Gessler

So @JosephWright, reading your comments on the xe-lua question: The diff I posted shows nothing that would cause the problem? So something is handled differently in the engine implementation that causes this?

Joseph Wright

@PaulGessler I'm still not sure: I want to check with Will what a 'minimum' set up is to use a Unicode math font, then I can put together a plain TeX version of the same issue

@PaulGessler See @egreg's comment above: he's suspicious of a global assignment somewhere

Paul Gessler

@JosephWright ah, ok. Thanks for looking into it. It's way above my head now... :-)

Joseph Wright

@PaulGessler These issues are tricky: there was another one from @UlrikeFischer some time ago that was equally hard to track down but was in the end an engine thing

Paul Gessler

11:46

@JosephWright yeah, I can see that. Previously it had been narrowed to amstext without amsmath, so I figured something could be found from that diff. But now I see we can produce the problem without any ams stuff at all.

Joseph Wright

@PaulGessler The fact there is a bug in LuaTeX with handling the size of mu makes me think that the difference here is engine-related

David Carlisle

I'm more suspicious of a local one, all the NFSS math settings are global: % We must still define the hook |\every@math@size| we used in
% |\glb@settings|. We initialize it to nothing. It is important to
% remember that everything that goes into this hook should to global
% updates, local changes will have weird effects.

Joseph Wright

@DavidCarlisle Nothing in the hook I can see

David Carlisle

@JosephWright I know but... (started to trace but need to do some real work so I'll leave it to you, or this evening:-)

egreg

12:38

@DavidCarlisle I guess this is it: something that unicode-math wants to be local, but instead is made global by \globaldefs=1.

cis

13:12

So I just decided, making a good english course:

►

Vrouvrou

Hello, i want to install the new takmaker i have a macbook pro Mac OS 10.6.8 how to do

cis

►

I choose a good course?

Vrouvrou

hey please

yo'

@Vrouvrou hello. Sorry, but the chat doesn't work like this, that you shout "hey" and people hop in and answer. Yes, sometimes this works, but if it does not, you should simply ask question on the site in the standard way.

Vrouvrou

i just want a link that's all

sorry

yo'

13:26

@Vrouvrou you mean this? lmgtfy.com/?q=texmaker+mac+os+10+download

Johannes_B

15:46

@PauloCereda youtube.com/watch?v=UGeNoml0OHc

Johannes_B

16:03

2

Q: Update l2tabu and translation?

The document l2tabu gives overview about the sins in latex: obsolete commands, packages, bad style. However this document in it's most recent (german) Version is from 2011, the last translated english version even from 2007. Are there new sins to consider? which packages also have been succeed...

best-practices

Any thoughts on that? I think it should at least go to meta. And even there it's ....

Paul Gessler

@Johannes_B I don't think it should be on Meta, it doesn't have anything to do with the site other than that people here would work to create/update it.

Johannes_B

@PaulGessler We had a few other questions asking for help which got migrated on meta. To be honest, i think it should be closed.

Paul Gessler

@Johannes_B I know, I was one of the close voters on one that was migrated to Meta. :-) Whether it should be closed is a separate issue, but I don't think it belongs on Meta because it isn't a question about the site.

Johannes_B

@PaulGessler So close it right where it is? :-)

Paul Gessler

@Johannes_B I don't have a strong enough opinion on that to vote one way or another. On the one hand, requests for help on projects are off-topic, but on the other hand, questions about {best-practices} are historically allowed. This question has elements of both.

Johannes_B

16:20

@PaulGessler best-practice is a good keyword. I would have called it opinion-based. The high rep users will stumble over this and give an opinion. I'll let it stay for now.

Paul Gessler

Anyone know how you make the tag marker/link in chat? I thought it was [] or {} but neither worked.

David Carlisle

@PaulGessler longtable [tag:longtable]

Paul Gessler

@DavidCarlisle ah, I missed the tag:... thanks!

David Carlisle

17:21

@JosephWright going in circles:-)

Joseph Wright

17:36

@DavidCarlisle Yes

@DavidCarlisle Have been busy so need to catch up

@DavidCarlisle As there seems to a subset of the team to actually write the code, I suspect that subset will carry the day :-)

percusse

17:54

@DavidCarlisle You were mentioning once that Matlab is spending too much time on the protocolling the Fortran functions. I am now using tic, toc for measuring time on Matlab and it is giving me 0.007~ on average. Python version is 0.0007 on avergae. Do you think it is possible? I didn't even optimize yet.

It sounds a little bit too good to be true

David Carlisle

18:27

@percusse this a function that calls back to a user function in matlab?

yo'

19:27

@egreg I've got a pearl for you: \begin{align*} \textbf{\emph{f}}_{(a,b)}:=

and: is given by Ind(\textbf{\emph{f$_k$}})=$2+k+1/r_{k,1}$, where

yo'

20:15

Which catcode is ignore, please?

Joseph Wright

@yo' 9

yo'

20:32

@JosephWright ok, thanks :) however, setting \active and then \deffing did the job, too :)

Joseph Wright

@yo' Different thing

yo'

@JosephWright for my testing, it was fine :)

egreg

20:58

@yo' TeX is too difficult!

cslstr

I just came across the error "undefined control seq \p@page" ... never seen that one before and can't find it in the source code anywhere. What is \p@page supposed to be?

Paul Gessler

@egreg you should ask why that is so on the main site—easy rep boost!

yo'

@egreg well, I'm finished just now. It seems that I started at 20:15, so it was only 1:45 for 9 pages. Not bad.

@cslstr MWE?

cslstr

@yo' yeah... that's why I'm asking here first .. example is far from minimal at this point.

@yo' I am sort of using parts without chapters and trying to label the part ... \ref then fails with that error.

yo'

@cslstr ah, well, do you use \pageref ?

David Carlisle

21:03

@cslstr every counter has a p@... macro, although not normally page as that's not really defined via \newcounter

cslstr

@yo' I'm using varioref ... that's the only thing I can think added something along those lines.

yo'

@cslstr well, can be.

David Carlisle

after \refstepcounter{section}, \p@section` expands to whatever \label should write as the current reference string.

cslstr

@DavidCarlisle Oh ... I think that explains it hten

I've incremented my page numbers artificially, but I think I used \refstepcounter instead of just \stepcounter

David Carlisle

@cslstr why? you shouldn't be generating \p@page

@cslstr bad idea:-)

cslstr

21:07

@DavidCarlisle Indeed :)

@DavidCarlisle I was restarting page numbers each \part - so I stepped the page on each \part - but used \refstepcounter instead for some reason. Anyway, fixed problem now. Thanks!

David Carlisle

@cslstr just think if you'd have asked a question on site i'd have answered it, got a 1000 votes, a good answer badge, passed @egreg. But here in chat i get nothing, such is life.

cslstr

@DavidCarlisle All of that, right before it gets closed for being "too localized". But you're a good sport anyhow :)

Ramanewbie

$\langle\rangle$

David Carlisle

@Ramanewbie ?

Ramanewbie

@DavidCarlisle don't care, just tests...

David Carlisle

21:17

@Ramanewbie why here?

Ramanewbie

@DavidCarlisle how to write an angle in $LaTeX$ ?

@DavidCarlisle because it's the $LaTeX$ room, and I know this one isn't frozen

@DavidCarlisle Can you help me with that ?

David Carlisle

@Ramanewbie as I think i said the other day we don't have mathjax here so there is no online rendering (not that mathjax uses tex anyway)

Ramanewbie

mathjax ?

anywau I can see the latex interpreted here

$LaTeX$ how do you see it @DavidCarlisle ?

David Carlisle

@Ramanewbie $LaTeX$ I think you are looking for \angle ?

Ramanewbie

$\angle$

what's the syntax ?

David Carlisle

21:21

@Ramanewbie mathjax is a javascript library to interpret tex-like syntax and render it it is used on some of th estackexchange sites but not on tex.stackexchange

@Ramanewbie what is the syntax for what?

Ramanewbie

for \angle

David Carlisle

@Ramanewbie what you wrote (it's defined by amsmath package)

@Ramanewbie you must have a local configuration to use mathjax in chat if you see $\alpha$ as an alpha, it is not the default.

Ramanewbie

@DavidCarlisle I already told you $LaTeX$ is interpreted for me here !

what does $\angle$ mean ?

David Carlisle

@Ramanewbie Then you must have configured it to be that way, it is not the default.

yo'

@Ramanewbie then you have got some chat script running in your browser.

Ramanewbie

21:24

I do, and want to ...

so what is it ?

David Carlisle

@Ramanewbie ?? it doesn't mean anything it makes the symbol normally used to denote an angle, it looks a bit like a squashed <

Ramanewbie

@DavidCarlisle Can you give an exemple using this ? $\angle$

Joseph Wright

@Ramanewbie Something like

\documentclass{article}
\usepackage{amsmath}
\begin{document}

$\angle ABC = 45^\circ$

\end{document}

David Carlisle

@Ramanewbie You asked for the symbol for an angle, and in tex it is \angle in math mode. $\angle$\bye is a complete plain tex document that prints that symbol, what examnple do you want?

Ramanewbie

$\angle ABC = 45^\circ$ @JosephWright what does that mean ? That the angle ABC is 45° ?

Joseph Wright

21:28

@Ramanewbie Yes, but that's a maths question not a TeX one!

Ramanewbie

ok

$\circ$

David Carlisle

@Ramanewbie stop doing that!

Joseph Wright

@Ramanewbie Please stop posting stuff without marking it as code (if appropriate)

Ramanewbie

@DavidCarlisle what's wrong ?

Joseph Wright

@Ramanewbie You keep posting bits of code but with $ symbols around and not marked as code

Ramanewbie

21:30

ok

Paul Gessler

@Ramanewbie if you just want to test chat features/syntax, you can use the "Sandbox" room

Sandbox

Where you can play with chat features (except flagging) and ch...

2

David Carlisle

@Ramanewbie this is not a private sandbox for you to experiment with tex syntax, it's a chatroom where you are chatting to people, so you are supposed to make the statement intelligible with a question or a statement or something, not just random tex syntax

Joseph Wright

@PaulGessler Ah, didn't know about that one :-)

@DavidCarlisle Can be random cricket news, though

David Carlisle

@JosephWright Of course!

@JosephWright I wonder if @PauloCereda brought Psmith with him, to improve the level of conversation

Joseph Wright

@DavidCarlisle Do you think we are getting anywhere on the team list?

David Carlisle

21:35

@JosephWright no

Paul Gessler

@JosephWright I didn't know about it until a few days ago either; it was posted in the new engineering chat room. :-)

Joseph Wright

@DavidCarlisle I don't think Chris really gets the point about the reality

@PaulGessler Ah

Paul Gessler

!!/battle

egreg

@PaulGessler Psmith went to SP with @PauloCereda

Paul Gessler

@egreg indeed, it appears psmith is afk :-)

David Carlisle

21:47

@JosephWright but given the comments about common use of legacy fonts I'd be prepared to let fontenc go unchanged this time round, although there is an argument that says if people are doing that a lot a warning is more important, but there are enough big picture structural changes can leave smaller issues until next time

Joseph Wright

@DavidCarlisle I see that

@DavidCarlisle My feeling is if you do need a legacy font with XeTeX then you have to work 'by hand'

David Carlisle

@JosephWright yes that would be fine except for the hyphenation issue, which given current font loaders used by texlive means you've got to rebuild the config files and language.dat and format all by hand which is rather a lot of work given that almost everyone has forgotten how to build a format by hand these days

Joseph Wright

@DavidCarlisle You slightly miss my point

David Carlisle

@JosephWright it's been a day for that kind of conversation:-)

Joseph Wright

@DavidCarlisle With a legacy font and XeTeX you really have to do everything by hand and that includes hyphenation, but that means it's for small pieces of specialist output

David Carlisle

21:56

@JosephWright that is certainly the case now (it's not clear it has to be that way) but if that's the view fontenc probably did ought to scream at you to check you know what you are doing. I fear that's not going to be universally popular though. Not to mention that the default encoding is OT1...

Joseph Wright

@DavidCarlisle Really that last point should have been addressed 6 years ago for XeTeX

David Carlisle

@JosephWright but can it be addressed now....

Joseph Wright

@DavidCarlisle You have to ask why people would be using XeTeX but not at least a Unicode font for the bulk

@DavidCarlisle Well yes, I'd do that (and for LuaTeX)

David Carlisle

@JosephWright because they heard about it and want to see if it works with their thesis

Joseph Wright

@DavidCarlisle Yes, but normally that would include \usepackage{fontspec} somewhere and certainly would not be a legacy doc

@DavidCarlisle Do you really think other team members would go for such a radical change?

David Carlisle

22:03

@JosephWright There is no real point in using xetex or luatex on a document using legacy fonts but if you want a publisher or arxiv or anyone else with a back catalogue of 1000s of documents they would be a lot happier if they could switch engines and get no diffs or at least managable number of diffs, but if the default behaviour is silent mis-encoding of legacy encoded text then it's never going to be possible

Joseph Wright

@DavidCarlisle Yes

David Carlisle

@JosephWright might, if you could wind it back with [2014/01/01]{latexrelease}

Joseph Wright

@DavidCarlisle But then they should really stick to pdfTeX

@DavidCarlisle True

@DavidCarlisle You going to suggest this?

David Carlisle

@JosephWright but if that's not maintained and stops working at some point they will need to switch to an engine that is maintained

@JosephWright your turn:-)

Joseph Wright

@DavidCarlisle At present unlikely, and if it does there are very big questions all round

\makeatletter
%\def\f@encoding{UC}
\expandafter\font\csname UC/lmr/m/n/10\endcsname
  ="[lmroman10-regular]:mapping=tex-text"
\makeatother

@DavidCarlisle ^^^ :-)

Could not be bothered to get NFSS tables to work, hence commented out line :-)

David Carlisle

22:10

@JosephWright It doesn't have to actually stop working it just needs to be "unmaintained" and people will want to know an exit route. Although I suppose someone is always likely to step in to maintain it. But if normal authors have all switched to xetex or luatex using pdftex just for legacy archive data means an incompatible switch point at some point in your catalogue which isn't pain free

2

@JosephWright perhaps....

Joseph Wright

@DavidCarlisle There would still be issues: as I say, I think there would be a lot of issues under such conditions

David Carlisle

@JosephWright trouble is it's thin end of the wedge, that would leave you with lm text and cm math, not sure unicode math is ready for default yet??

Joseph Wright

@DavidCarlisle Even if we worked very hard, LuaTeX is deliberately not 100% the same as pdfTeX, and the 8-bit vs. Unicode changes to primitives mean some things are different with XeTeX (e.g. the test on the ConTeXt website for a Unicode engine)

@DavidCarlisle Could have picked CMU

@DavidCarlisle ConTeXt do exactly that

@DavidCarlisle Latin Modern Math is 'done'

@DavidCarlisle I've had some interesting discussions with the TeX Gyre people about the math mode fonts :-)

David Carlisle

@JosephWright yes I know the font is there but changing the default text encoding from Ot1 to UC/EUx seems a lot less scary than defaulting all the math codes to unicode slots not to mention questions about following or not unicode-math's treatment of math alphabets

Joseph Wright

@DavidCarlisle Yes, hence the fact I went with something simple :-)

@DavidCarlisle Plan B: fix xor, write LSS, release LaTeX3 with this all sorted from the start. Sounds much like the ConTeXt plan.

David Carlisle

22:16

@JosephWright but that's what I meant about thin end, is the simple thing enough to be worth doing on its own, not sure...

@JosephWright that was the plan in 1993, 2e is just the stop-gap until we're done.

Joseph Wright

@DavidCarlisle Yes, I can see that, hence not having suggested it even though it's basically my position!

@DavidCarlisle Not quite: the ConTeXt plan and the reality for L3 now is that you are talking about a system with similar (but not identical) interfaces and entirely different code behind them

David Carlisle

22:33

@JosephWright yes perhaps, but meanwhile we probably can't get that otherwise excellent plan completed by TL2015 so I guess we have to update 2e.

Joseph Wright

@DavidCarlisle Yes

David Carlisle

@JosephWright we have a zip file, we have a ctan upload form. we could just do it:-)

Joseph Wright

@DavidCarlisle Yes again

@DavidCarlisle Probably I will end up actually doing that task in any case!

Faheem Mitha

22:56

I see Paulo finally got the interview up.

David Carlisle

@FaheemMitha time to pick the next victim!

egreg

23:16

@DavidCarlisle I propose @HarishKumar

David Carlisle

@HarishKumar ^^^

Christian Hupfer

Good maen...

@PauloCereda: Typing from Fedora 21 ;-)

Harish Kumar

@egreg @DavidCarlisle First message of this day is scary!

:-)

David Carlisle

@HarishKumar It only took two years and the threat of a noose to drag @PauloCereda to the interview room, so it's not that scary

@ChristianHupfer hi from WINDOWS 8

Christian Hupfer

@DavidCarlisle: No comment ;-)

Harish Kumar

23:26

I think we have @Mark winbrow, @Christian Feuersänger, .... before me.

@DavidCarlisle: Does that mean the threat is not so near? ;-)

@ChristianHupfer Good morning from windows 7, the great!

Christian Hupfer

I can see dead operating systems

Harish Kumar

@ChristianHupfer Dead operating system have the advantage that they can not die again! ;-)

Sandbox

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends