« first day (257 days earlier)      last day (4683 days later) » 

10:34 AM
Quick and dirty question: why my index doesn't sort accented words correctly? For example, some portuguese words like ábaco or ébano (which start with an accented letter) are sorted wrong.
 
Just writing on my thesis and found an application of an \marginbox which adds some margins to a box or image. Like the inverse of trim / \trimbox. This saves some \hspaces and simplifies vertical adjustments...
@PauloCereda Most likely the underlying code (Unicode) value is used and that isn't anywhere near 'a' or 'e' in this case. Also this letters are often active and implemented using macros, so string compare and sorting in LaTeX itself is also tricky.
 
10:51 AM
@MartinScharrer Hm I see.
I think this a nice question. Should I post it?
 
@PauloCereda You can try passing the -L option to makeindex.
Assuming that you have your locale set to Portuguese.
 
@egreg Wow, it worked! o.O
@egreg: What does the L flag do?
 
11:15 AM
@PauloCereda It uses the locale settings for collating. Not all implementations of makeindex are able to do it. See "man locale" for details.
 
@egreg: Got it! Thanks, you saved my life! I owe some drinks too. =)
 
11:27 AM
@PauloCereda Hurray! I can drink for free also in Brazil! The only problem is to get there. :)
 
12:08 PM
@egreg: I'm gonna host a TUG conference in here, so everybody will have a good pretext to be here. =P
 
Hmmm. This isn't the first time I've hit up against KOMA's TMOS: Too Many Options Syndrome
The documentation is ok, but it's not clear which option I have to change to ahcieve what I want...
 
12:49 PM
Ah. Another "Answering my own question while trying to ask it". I like it when that happens.
 
user19161
@Seamus I see you have not fixed the 10k bug, but getting there.
 
user19161
@PauloCereda But you have no agenda for the conference. :)
 
@JasperLoy True. I need to fix that. =P
 
@JasperLoy Which 10k bug?
 
user19161
1:06 PM
 
user19161
Does anyone know what happened to tug.org.in? That site has been down for months and nobody knows why.
 
Aah. That 10k bug. Yes. I should have a fix in the next month or so.
I was confused, because I've been working on a package and had a problem where one of the macros stopped working for numbers bigger than 100,000. But I didn't think I'd mentioned it on here...
 
user19161
The Indian TUG tutorial is my favourite free tutorial. Fortunately it still can be accessed online even though tug.org.in is down.
 
user19161
Of course I have the hardcopy already. :)
 
3:37 PM
ctan.org seems to be down.
 
@MartinScharrer Odd
 
@JosephWright: It's there again. But seems to be slow.
@JosephWright: Say, do you know any packages for test cases, i.e. to test macros of your own packages?
 
@MartinScharrer There is qstest, or you can adapt the LaTeX3 Project's test system (if you feel brave)
I just use the documentation itself
 
@JosephWright the documentation itself?
 
@MartinScharrer Read the siunitx documentation carefully, and notice how many odd things there are in the examples :-)
 
4:00 PM
@JosephWright Ah, I get it. This doesn't work for me. I have 50+ test cases for svn-multi already.
The qstest package seems to be what I want. I need to add some support for box test, i.e. correct box dimensions etc.
 
@MartinScharrer In the longer-term, I'll probably go down the LaTeX3 test suite route.
We should really make it available as a general tool - I'll perhaps raise that on the team list
(When you get to 1000s of tests then it needs to be automated)
 
LaTeX man, LaTeX man. Doing the things LaTeX can. What's he like? It's not important. LaTeX man. Is he a dot, or is he a speck? When he's underwater does he get wet? Or does the water get him instead? Nobody knows, LaTeX man. (With apologies to They Might Be Giants)
2
 
I'm just thinking about putting my next package or all of them on bitbucket.org and us Hg as version control too. At the moment I use SVN and my own Track server.
Any other good alternatives? I just know of Github.
 
@MartinScharrer I use BitBucket myself, but also GitHub for some other projects
Personally, I prefer BitBucket, but I see why GitHub is popular
 
@JosephWright Is it suitable for LaTeX packages?
I mean, they seem to prefer "real software" :-)
 
4:12 PM
@MartinScharrer Both are perfectly suitable, as are other alternatives
 
Would this be as suitable question for the main site? Or do we get a "My VC software is better than yours" flamewar?
 
@MartinScharrer I'd say it's off-topic for TeX.SX (the detail of version control is not a TeX question)
The main differences are at the website level, in terms of things like bug tracking, etc.
Both Mercurial and Git are full-featured version control systems, and if you know the correct commands can do many of the same things
(There are some more complex areas where there is a difference, but the basics are all there in both)
 
@JosephWright I meant something like "how to host your package code ..." or similar. Doesn't sound so off-topic.
 
@MartinScharrer Still not really TeX-related, as all that is needed is 'set up rep, pull rep, add LaTeX file, check in, push rep'
That's the same as you'd do for any other source files (C, Python, plain text, Markdown, ...)
 
Right.
 
4:17 PM
Of course, you could ask and see what the community as a whole feels - that is just my opinion (and I can't vote to close :-)
 
@JosephWright Oh, yes, sometimes it's bad that you can't vote or flag "normally" without super powers.
 
@MartinScharrer That was asked on meta.so, and {status-declined}
 
@JosephWright I see. Doesn't surprise me. They decline a lot of stuff.
 
I use github. It hadn't even occurred to me that I could use it for bug tracking and things as well! I just use it as a remote backup (and VCS)...
 
Other thing: There is no actual archive from the Comprehensive TeX Archive Network (CTAN), is there? I mean I can't get the version from three month ago somewhere?
 
4:22 PM
version 1.01 of moreenum is floating its way towards CTAN!
 
@MartinScharrer This has been discussed before. Some people like the idea of a master rep for all TeX stuff, but that is not realistic (not enough CTAN people, not enough devs would use it, ...)
TeX Live's SVN is the nearest you get
 
@JosephWright Doesn't have to be a VC system, just some form of archive. Sometimes you need an old version of a package!
@JosephWright: Did you ever convert a SVN repository to Hg?
 
@MartinScharrer Yes: the siunitx one
@MartinScharrer Still complex to run (remember CTAN was set up a long time ago, when disks were expensive)
If you want changes at CTAN, be prepared to run it!
(Rainer and Robin are both pretty busy)
Just to be clear: both Rainer and Robin are already fully-committed CTAN-wise. So any changes will need additional time from someone
 
@JosephWright I know. I was just thinking about some server which rsyncs like a mirror but then keeps the older versions. This is actually easy to setup. I have an own server, but diskspace would be an issue over the long term. I have only 50GB, where I need 10-20G for other things.
 
What if only package deltas are stored?
Hm nevermind, at some point the complete package should be available.
 
4:37 PM
@PauloCereda I have some incremental 7ZIP archives in mind which only store the differences.
Just to get it archived. I'm not talking about having it fully accessible by everyone yet.
 
@MartinScharrer but for old versions of the manuals, which aren't plain text, even the incremental zips will get big quickly, right?
 
@Seamus Depends on how often these packages change.
Also the really big chunks like TeXLive ISOs would be archived, just packages etc..
 
5:31 PM
@JosephWright, @all: Something else:
I remembered my idea about Polls.
We could have one "question" on meta for this.
The answers are tiny polls with comments for the possible answers.
People then up-vote the comments if they apply to them or can add own comments
Things like "What TeX distribution do you use?" [ ] TeXLive, [ ] MikTeX, [ ] MacTeX, ..
 
 
1 hour later…
6:56 PM
I think we should be able to award some reputation to an answerer (for any question). I mean, it's great to upvote an answer when it's really useful, but sometimes a great answer saves our life.
- Normal upvote: "Cool, thanks! Your answer is very good and useful!"
- Award reputation: "Dude, for real, your answer saved my life!"
I know we have bounties, but I think this is quite different. =)
 
7:37 PM
Can someone remove this question from the "unanswered" list? Either upvote one of the three existing answers or write another one. :-)
 
@lockstep done
 
@MartinScharrer Thanks. (And I don't think that my answer deserves more than one upvote.)
@PauloCereda It is possible to award bounties retroactively -- just be upfront about your intentions when starting the bounty.
2
 
@lockstep Hm I see. I'm gonna try it next time. =)
@lockstep: Incidentally, you are one of the great answerers I was referring to. =P
 
@PauloCereda (blushes)
 
7:53 PM
@lockstep: This answer of yours was superb. I was struggling with indices during all morning. =P
 
8:08 PM
Right now, I have 6666 votes cast. :-)
 
This week I had 33777 rep :-)
 
I got a 5!
=P
 
8:52 PM
Have we decided yet when to celebrate our birthday? Is it the start of private beta, or the start of public beta?
 
@lockstep I don't know if that's evil or not.
 
@AlanMunn Not as evil as 33777 rep (or egreg's rep per day). ;-)
 
To anyone: isn't texdoc supposed to find locally installed documentation if it exists? I have the latest version of enumitem.sty and its docs in my local texmf but texdoc still brings up the distribution version of the docs. Other local doc files are found by texdoc. WTF?
 
9:27 PM
Friends, is there a way to dump all the macros and settings of a document?
I want to know the value of \idxtitlefont
 
@MartinScharrer svn-multi begreport sourceforge.net/mailarchive/…
 
@AlanMunn Odd, I tried texdoc enumitem and it pointed to the local file.
 
@MartinScharrer the report seems correct. I retexed my thesis in draft mode. The table of revisions used to include all the \included files, but now it does not.
 
@LevBishop: I have a look on it. I remember there was an issue with that once. I didn't changed it however since a few month.
 
@AlanMunn What's the answer of texdoc -l enumitem?
 
9:42 PM
@egreg It orders the texmf-dist as 1, the local as 2 (which explains the result, but not why the order isn't reversed.)
 
@PauloCereda Put \show\idxtitlefont somewhere after \begin{document} and compile from the shell
@AlanMunn That's strange: on my system the local file comes first. What about texdoc -f?
 
@egreg Thanks, it works like a charm! Now I owe you \setcounter{drinks}{2}. =)
 
@egreg Everything absent except /usr/local/texlive/2010/texmf/texdoc/texdoc.cnf, which is active
 
@MartinScharrer seems to work fine with \input instead of \include, if that gives a clue.
 
@LevBishop Yes, it does.
@PauloCereda You mean \refstepcounter{drinks}!
 
9:54 PM
@MartinScharrer Ouch, now I owe drinks for everybody!
 
@AlanMunn Try texdoc -d -l enumitem, which should tell what's the reason (among many other things)
 
@PauloCereda Both \setcounter and \refstepcounter are global assignments, so YES!
 
@MartinScharrer (dramatic music) nooooooo!
 
Still about 250 unanswered questions. :-( I suggest that people skim their favourite tags for unanswered questions where crucial information from the OP has been missing for a long time and vote to close such questions. (This is exhibit A for .)
 
@lockstep Closed
:-)
 
9:59 PM
@JosephWright Thanks. BTW, shouldn't the reason for closing be "not a real question" rather than "too localized"?
 
@lockstep I went with the consensus ('not a real question'), although I commented first and would personally have picked 'too localized'.
 
@egreg You weren't kidding about the "among other things" :-) So wading through this, the final score for the texmf-dist is 7 and for the local file is 4.5. In the temf-dist version it give Heuristic score 4, exact match; heuristic score 4.5, directory bonus, but then New score 7 from metadata, which I guess is why it wins. Is there something else in there that is relevant?
 
I suspect that this is really a feature request rather than a question
 
@JosephWright I'm not sure if that's a valid reason to close it.
 
@lockstep Hence asking for thoughts!
 
10:05 PM
"Get the weekly newsletter!" - Since when do we have this ad in the front page?!
I noticed it just now.
Oh nevermind, AdBlock seems to disable it. =(
 
@JosephWright I myself have asked questions about biblatex that turned into an official feature request (and eventually a feature), e.g. this one. And didn't Konrad Rudolph write the "minted" package as a self-answer to a question at stackoverflow? So I'd say, potential feature requests are appropriate questions.
 
@AlanMunn I get score 7 for both, so the local one wins, because the local tree comes before in $TEXDOCS. You can limit the output with texdoc -d=score -l enumitem.
 
@egreg Is there something in my setup that I'm missing that should cause the local file to give me 4.5 and not 7? (I thought there might have been a local ls-R file but that doesn't seem to be the case.)
 
10:33 PM
@AlanMunn I really don't know; I get score "7 from metadata" for both. How's the directory in the local tree called?
 
10:48 PM
@egreg It's just ~/Library/texmf/doc The enuitem.pdf file is in its own enumitem folder. If I move it out to the doc folder its score reduces by .5
 
@AlanMunn That's the problem! It should be ~/Library/texmf/doc/latex/enumitem
 
@egreg Thanks! I guess I'd never noticed it before because most of my local stuff is really local and then the folder structure doesn't matter since there's no competing file in texmf-dist. Thanks for persevering with this.
 
@AlanMunn I learnt something too, for example how to limit the torrential output of texdoc -d.
 
11:57 PM
@JosephWright I think the auctex package question is a valid one. I just think the current best answer we can give is "No there is no such package." If I didn't have three papers to finish writing this summer I'd have a bash at learning a bit of elisp to start a proper answer. Give me a week and I could throw together a "No but here's how you'd start to write one for yourself" answer...
 

« first day (257 days earlier)      last day (4683 days later) »