TeX, LaTeX and Friends

subham soni

12:40 AM

After updating all the packages in MiKTeX, it finally worked :)

One question, does it really matter if I compile a document with PdfLaTeX or arara????

David Carlisle

@subhamsoni arara isn't a typesetter it is a command interpreter that runs pdflatex or whatever is needed

Werner

1:49 AM

15

Q: A word for a joke so poorly told and so unfunny that one cannot help but laugh

There is a word for this in Indonesian language: jayus. (Maybe, it is used in Filipino and Malaysian language also.) It is a joke that is so bad, it's funny. It is often mentioned as untranslatable. I wonder if there is any English equivalent though. (and yes, English does not have to have a w...

Sean Allred

3:56 AM

@Nasser Wait, which university?

@egreg I can almost follow that—I think I'm getting better at this :) I wonder if there are any graphical animations that show how some of these trickier cs concoctions work.

mbork

6:57 AM

Hi people, I have a problem: something gets inserted in the MVL (as \tracingpages=1 shows me), but I have no idea what's that. Is there any way to see the contents of the MVL in the log?

("Something" looks like a skip or kern, since it's invisible in the pdf, of course.)

David Carlisle

7:17 AM

@mbork \showoutput

mbork

Thanks! And how did I restrict the depth of the boxes shown? (I have my copy of The TeXbook at work and I'm at home now - sorry for the question...)

(OK, got it: I've just massaged the log using Emacs;))

Wow, now it works like a charm. What a stupid error. It was a depth, not a skip...

Joseph Wright

7:36 AM

@PauloCereda Ignore the file stuff: I've decided it's more sensible to do OS detection then use the 'native' commands

@PauloCereda At least for e.g. file copying

1010011010

8:12 AM

Morning.

I have a really small question about a piece of code too big to explain here... Yay or nay to post on main page . . .

Joseph Wright

@DavidCarlisle Did you see mail about keyval on the list?

Tarass

@PauloCereda @HarishKumar Ok, thank you

Joseph Wright

8:28 AM

@DavidCarlisle Also tex.stackexchange.com/questions/175982/…

1010011010

@DavidCarlisle I've tried doing that \@tempcntb thing but such expression didn't yield anything. So I instead tried \def (to also avoid the need of \relax for expandable \edef, cleaner code) and although it stores the correct value, I cannot reuse it in any ifnum. I can't say I can explain why it doesn't have its intended effect.

Joseph Wright

@1010011010 Context for this?

David Carlisle

@JosephWright Not got to email yet, I was a bit distracted by the longtable question on site (in which @egreg gets an honourable mention). Will look now,

Joseph Wright

@DavidCarlisle Summary: changing brace stripping behaviour is a bad idea

David Carlisle

@JosephWright hmm will look

1010011010

8:36 AM

@JosephWright tex.stackexchange.com/questions/175850/… , I'm determined to fully understand David's answer.

Joseph Wright

@DavidCarlisle For 'sensible' brace stripping people really should use l3keys

Harish Kumar

@Tarass See you :)

1010011010

8:50 AM

@JosephWright I just posted on TeX front page.

Joseph Wright

@1010011010 Eeek! You've redefined a temporary count as a macro!

@1010011010 \@tempcntb\count@ is presumably what you wanted

David Carlisle

@JosephWright replied....

@JosephWright arg latex-l bounced my mail it decided I wasn't me because of our new mail server, did my second attempt get through?

1010011010

@JosephWright No dice with that solution. I think I already tried that as well, which is why I moved on to \edef.

Stephan Lehmke

9:33 AM

@1010011010 See my comments. I think you should start much simpler until you understand exactly what \@for and \ifnum do.

@egreg I think the \ifnum is supposed to be inside the \@for loop to test for "last element in list".

David Carlisle

@JosephWright nah the idea was good the implementation was less so

@StephanLehmke I blame you!

1 hour ago, by Joseph Wright

@DavidCarlisle Summary: changing brace stripping behaviour is a bad idea

Stephan Lehmke

@DavidCarlisle What? Did you really use my code from 2002?

David Carlisle

@StephanLehmke ^^^

Stephan Lehmke

@DavidCarlisle I tried to look at the error message, but the list archive requires a login.

David Carlisle

@StephanLehmke No I changed it a bit (it's possible that your original worked, I haven't tried this morning, but even if it did I still blame you, I have to blame someone:-)

Stephan Lehmke

9:46 AM

@DavidCarlisle :-)

David Carlisle

@StephanLehmke inherit={german} works as before but inherit = {german} now keeps the brace and apparently there isn't a language called {german} in biblatex.....

@StephanLehmke so I need to change round the order of space keeping and brace dropping somewhere but I need to do some day job things first

Joseph Wright

10:21 AM

@DavidCarlisle Take a look at the code in l3keys :-) Actually, probably better to look back in the sVN to the older l3keyval module: I guess I'd look around 2009 (say r1000-ish)

David Carlisle

10:38 AM

@JosephWright you sure l3keys gets all cases right:-)

Joseph Wright

@DavidCarlisle No, but the older code probably does :-)

David Carlisle

11:44 AM

@JosephWright hmm current l3keys one strips two braces (like the old keyval, but does it more consistently)

Joseph Wright

@DavidCarlisle Eeek

David Carlisle

\documentclass[]{article}

\usepackage{expl3}

\def\j{g}

\ExplSyntaxOn
\keys_define:nn{zz}{
  inherit .code:n = \toks0{#1}\typeout{[\the\toks0]},
  inherit .value_required:
}

\let\setkeys\keys_set:nn

\ExplSyntaxOff


\typeout{^^JShould be [\string\j\space erman]}
\setkeys{zz}{inherit={\j erman}}
\setkeys{zz}{inherit= {\j erman}}
\setkeys{zz}{inherit ={\j erman}}
\setkeys{zz}{inherit = {\j erman}}

\typeout{^^JShould be [german]}
\setkeys{zz}{inherit={german}}
\setkeys{zz}{inherit= {german}}
\setkeys{zz}{inherit ={german}}

Joseph Wright

@DavidCarlisle I'm taking a look at the code now myself: I have a feeling we can tighten it up a bit

David Carlisle

@JosephWright see the last two blocks in the above

Joseph Wright

@DavidCarlisle Yes, point taken

@DavidCarlisle OK, how about I fix l3keys and you fix keyval :-)

@DavidCarlisle I didn't actually write much of the low-level part of l3keys!

David Carlisle

11:50 AM

@JosephWright It turned out that the old keyval while documented as always removing two sometimes only removed 1 if there was a space after the = and the new one removes one less than that so if there is a space after = 0 braces get stripped and biblatex breaks. It isn't clear how to fix at moment, may need to wait until I'm at home. Or may need to revert that bit and leave the \long bit.

@JosephWright you inherit the code, you inherit the blame

Joseph Wright

@DavidCarlisle I must be being a bit thick: doesn't your demo have the same thing twice (e.g. blocks 2 and 5)

@DavidCarlisle Defined behaviour of l3keys is to drop exactly one set of surrounding braces if present: isn't that what you are seeing?

David Carlisle

@JosephWright if a test's worth doing it's worth doing twice:-)

8

@JosephWright well I didn't write the test I intended as you observed so my interpretation of the results may be flawed. I think I'll stop I haven't got time to debug my own code today, not sure why I'm trying to do two at once:-)

Joseph Wright

@DavidCarlisle OK

David Carlisle

@JosephWright yep sorry false alarm (having fixed the test I'll go back and see what keyval is doing:-)

\documentclass[]{article}

\usepackage{expl3}

\def\j{g}

\ExplSyntaxOn
\keys_define:nn{zz}{
  inherit .code:n = \toks0{#1}\typeout{[\the\toks0]},
  inherit .value_required:
}

\let\setkeys\keys_set:nn

\ExplSyntaxOff


\typeout{^^JShould be [\string\j\space erman]}
\setkeys{zz}{inherit={\j erman}}
\setkeys{zz}{inherit= {\j erman}}
\setkeys{zz}{inherit ={\j erman}}
\setkeys{zz}{inherit = {\j erman}}

\typeout{^^JShould be [german]}
\setkeys{zz}{inherit={german}}
\setkeys{zz}{inherit= {german}}
\setkeys{zz}{inherit ={german}}

David Carlisle

12:24 PM

@JosephWright mail sent :(

Joseph Wright

@DavidCarlisle Agree with plan, certainly for this TL release

@DavidCarlisle I'm going to tune up the l3keys code a little: I think we can save some tests/tokens

@DavidCarlisle My feeling is to get the behaviour 'right' for keyval would need a more comprehensive rewrite (see for example where in l3keys protective tokens have to be added)

Paulo Cereda

@Joseph: I have a plan of setting a Travis instance associated with all tests, so we can have a real-time feedback on stuff we make changes (coming soon). :)

David Carlisle

@PauloCereda yes although in this case getting feedback on things for which we didn't have a test would have been good:-)

@JosephWright hmm it's already fairly careful about not expanding stuff I hope but ...

Joseph Wright

@DavidCarlisle I realise that: brace stripping is tricky, though (also, I'm not sure what the rules are for braces in keys with keyval)

David Carlisle

@JosephWright although the key name has to be safe in a csname but thats a 2e feature, same true of environment names

Joseph Wright

12:31 PM

@DavidCarlisle Yes, that makes life a bit easier!

@DavidCarlisle BTW, \toks0{#1}\typeout{[\the\toks0] is easier as \showtokens{[#1]} :-)

David Carlisle

@JosephWright ooh you're so modern

egreg

@JosephWright Three default values call for a key-value interface.

Joseph Wright

@egreg Yes, most likely: I didn't like my answer much :-)

David Carlisle

@egreg one that works, preferably

egreg

@DavidCarlisle So not keyval? I gave two pieces of code, one with keyval and one with expl3, to be on the safe side.

David Carlisle

12:40 PM

@JosephWright do you know if latex-l allows attachments I was thinking of sending keyval.sty in reply to Ulrike

@egreg sorry been too distracted to look at q or a on site today:(

Paulo Cereda

@DavidCarlisle ooh heuristics ftw!

David Carlisle

@PauloCereda how much longer would it take if you tried every test while at the same time loading biblatex with every combination of language, that would have trapped this mornings bug (lucky that "german" triggered it as that meant it only lasted a few minutes:-)

Paulo Cereda

@DavidCarlisle ooh cartesian product ftw!

:)

Paulo Cereda

1:10 PM

@David: they found a bug?! /gasp

Joseph Wright

@DavidCarlisle Isn't \long\def\KV@split#1=#2=#3\relax an issue here: I was testing why we don't use that in l3keys (not currently noted in .dtx), but of course when you think it's obvious!

@DavidCarlisle Have tune-ups for l3keys in hand (I think!)

David Carlisle

@JosephWright yes but if I change that the change ripples through the whole package

Joseph Wright

@DavidCarlisle Indeed, hence thinking 'revert stuff' is the only safe approach here

David Carlisle

@JosephWright could you confirm biblatex is happy with the ammended file? (I don't actually have any biblatex examples:-)

Joseph Wright

@DavidCarlisle Will do this evening

David Carlisle

1:16 PM

@JosephWright Ok (I should be leaving this to this evening too) Given that miktex has apparently made it go live I was wondering if I should push it out now. does miktex have an open mailing list maybe I shoudl send my latex-l message there and then miktex can push out a temp fix today if they want

Paulo Cereda

@David: moral of the story: let us not update things ever again. :)

David Carlisle

@PauloCereda well if only Joseph would get on and finish xor we could freeze 2e and move on to l3...

5

Paulo Cereda

@DavidCarlisle oooh!

David Carlisle

2:24 PM

Does anyone have a contact for Christian Schenk (I got an error when using the contact form on the miktex site) (don't post the email address here:-)

subham soni

@DavidCarlisle u there

David Carlisle

@subhamsoni sorry not today:-)

subham soni

Hmmm its ok :)

Paulo Cereda

sourceforge.net/u/csc/profile

@DavidCarlisle I couldn't find it, but you could append his SourceForge username +at users dot sf dot net, it's worth a shot.

David Carlisle

@PauloCereda thanks, or I may try a different browser (but it looked like a server error)

@PauloCereda well it's not bounced at least....

Paulo Cereda

2:50 PM

@DavidCarlisle phew. :)

Stephan Lehmke

3:03 PM

I just posted a picture mode solution to a case where TikZ doesn't work ;-)

Paulo Cereda

@StephanLehmke oh my! :)

Sean Allred

3:23 PM

exsheets is now officially my new favorite thing. And by exsheets I mean xtemplate.

Paulo Cereda

3:35 PM

@SeanAllred Make up your mind. :) You can blame Clemens for exsheets and David for xtemplate. :)

Sean Allred

@PauloCereda And blame I shall! (I really do need to find my flaming pitchfork.) Honestly though, I've spent a good deal of effort trying to customize the exam package to meet my needs/desires/irrational obsessions; the fact that the (frankly fantastic) exsheets package uses templating well makes it more-or-less a piece of cake.

There's definitely terminology I need to learn, but so it is with any new thing.

Paulo Cereda

@SeanAllred Germans. :)

Sean Allred

@PauloCereda Germans.

Paulo Cereda

@SeanAllred <3

Sean Allred

@PauloCereda NS-American love <3

Paulo Cereda

3:38 PM

@SeanAllred woooo

:)

Sean Allred

:)

@PauloCereda I tried to find a gif of someone waving a tiny indecipherable flag, but alas, I failed.

Map projections are strange. Never really knew that the Americas were so askew with each other; always thought they were more-or-less stacked.

Stephan Lehmke

3:53 PM

@SeanAllred Well jokes apart, I cringe when I see people defining mathematical symbols with TikZ which could as easily be achieved by pasting together some rules. If such a symbol appears a couple thousand times in a document, the computing effort might just become significant...

Sean Allred

@StephanLehmke Joking again aside, I completely agree. When you want a scalable solution, it's imperative that it be as efficient as possible. Every computer scientist knows this. Such a limitation might be resolved with a clever use of \savebox and \usebox though, right?

Stephan Lehmke

@SeanAllred Good idea; someone should really write a TikZ meta package for defining symbols. But it would be significant effort to make this robust wrt. font (size) changes and such.

Sean Allred

@StephanLehmke Hmm………

@StephanLehmke Whelp, college is over, I have no finals, and I'm going to be bored out of my mind until May 17 when I graduate. I guess now is as good a time as any >:)

But first, the day is half over and I haven't gotten out of bed yet (let alone make anything to eat). Toodles!

Joseph Wright

4:52 PM

@DavidCarlisle As you're working on keyval, I decided to improve l3keys too: see latest checkin :-)

Stephan Lehmke

@DavidCarlisle While you were busy tidying up the mess you made out of keyval I managed to get a tick for a picture mode answer ;-)

3

Paulo Cereda

@StephanLehmke LOL

Joseph Wright

@StephanLehmke What he needs are some proper unit tests: I wonder who might have written such a thing for LaTeX :-)

Stephan Lehmke

@JosephWright Well it was David for sure. Maybe a split personality? noo noo that was the other David (maniacial snickering)

Paulo Cereda

@StephanLehmke Probably David K (K stands for Karlisle, of course). :)

Stephan Lehmke

5:15 PM

Now that I think of it, I get the feeling PicTeX is severely underused nowadays. As people aren't caring for performance any more, I should start posting more answers involving it. We don't even have a tag for it!

Heiko Oberdiek

@NicolaTalbot You might want to fix fmtcount: In German the s in eins is dropped, in some cases. Numbers with 1: eins, elf, einundzwanzig, einunddreißig, ..., einhunderteins, einhunderteinundzwanzig, ..., eintausendeinhunderteins, ... (Sorry, I don't have time to provide a fix, now).

Henri Menke

5:37 PM

Hello everyone! Does anyone know how I can slope a node along an arc in TikZ? \draw (1,0) arc (0:180:1) node[pos=0.5,sloped] {text text}; doesn't seem to work.

David Carlisle

@HeikoOberdiek hi not sure if you've been following latex-l today:( I just (a minute ago) updated ltnews21.tex and graphics/changes.txt and graphics/keyval.dtx in svn, those changes aren't signed off yet but probably those files will need updating in your latex tds thing at some point (I'll ping you by mail) sorry.

Nicola Talbot

@HeikoOberdiek Thanks. I have a feeling that might already have been logged at github.com/nlct/fmtcount It's possible I may have to drop the maintenance of fmtcount (and datetime) as I've developed a problem with my arm and I need to cut down my workload, but I'll see if I can get those issues fixed.

Heiko Oberdiek

@DavidCarlisle Thanks for pinging. Update is on its way to CTAN.

Paulo Cereda

arara people: would you like terminal colours? I mean, green for SUCCESS, FAILURE in red, usage of David's package will print an ASCII flag of UK... :)

egreg

5:53 PM

@PauloCereda Green, white and red? I like it.

Paulo Cereda

@egreg :)

Henri Menke

@PauloCereda But keep the possibility to opt-out, e.g. via araraconfig.yaml.

David Carlisle

@HeikoOberdiek er yes well that's what I meant:-) That you may want to hold off doing that until we update the base on ctan as currently that check is just "me" so rest of team could see the changes.....

Paulo Cereda

@HenriMenke Indeed, I was thinking of that. And I want to include some --color-mode hell which will bring colour madness to the console, just like those old websites full of colour and .gif animations. :)

Heiko Oberdiek

@NicolaTalbot Sorry for your arm problem. I am proofreading an article, which mentioned that fmtcount could be used because of the bug. The author has solved the issue by creating a database with a mapping of the numbers up to 9999 to their word representations. I can ask him to provide you the data to get you "some" test cases.

@DavidCarlisle Just in time. I could stop the upload and will wait until the new version hits CTAN.

Henri Menke

5:58 PM

@PauloCereda Like this website of a german party actually running for the european parliament?

Paulo Cereda

@NicolaTalbot Oh my, I'm sorry to hear about your arm. :( Get well soon.

ppr

Hi everone

Paulo Cereda

@HenriMenke aaaaaaaaaaaaaaaaaaaaaaaaaAAAAAAAAAAAAAAAA... Yes, that one is a good example. :P

ppr

I created a new question/answer in the faq spirit. I hope it will be useful.

0

Q: How to write a MWEB (Minimal working example with BibTex)?

Someone asks me to write a MWEB. How can I produce such thing? Explanation of the question's purpose (will be deleted afterwards): This question aims to complete the very useful other question intituled "I've just been asked to write a minimal example, what is that?" for the cases concern...

support mwe

There are probably some grammar mistakes (I'm no native speaker).

Henri Menke

3

Q: Mixing underline and strike-out

Word allows for all kinds of font formatting: The above example combines strike-through/strike-out (like this) and underlining. I'd like to be able to do the same, but neither ulem nor soul allows for this in a convenient way. ulem: \documentclass{article} \usepackage[a6paper,showframe]{g...

fonts formatting soul ulem

David Carlisle

6:07 PM

@HeikoOberdiek thanks:-) It seems no one (on the team) is around but Joseph and me, so I feel a bit hesitant to just peer review my own code and release it, on the other hand biblatex being broken doesn't want to stay that way too long....

egreg

@HenriMenke I want the jumping black sheep on my own site!

Paulo Cereda

@egreg LOL

egreg

6:26 PM

4

Joseph Wright

6:44 PM

@DavidCarlisle I'll take a proper look in a second

I am who I say I am

Where is the downloaded MikTeX saved by default?

In my experience in the past 3 years, the downloaded files are saved in a folder which is sibling to the downloader application.

But now everything seems to be changed.

I am who I say I am

7:16 PM

C:\Users\Garbage Collector\AppData\Local\Temp <=== here. Problem solved.

Joseph Wright

7:34 PM

@DavidCarlisle Fixed keyval working for me with a real biblatex use case

@DavidCarlisle, @HeikoOberdiek Latest commit looks fine to me: as noted in the log, this takes the code back to pre-v1.14 in terms of brace handling

David Carlisle

8:07 PM

@JosephWright do you have a ctan back door, or do you use the public web submission form?

Joseph Wright

@DavidCarlisle I believe I have a backdoor to the announcements, but not currently to uploads

David Carlisle

@JosephWright OK thanks

Joseph Wright

@DavidCarlisle As you are not currently on the GitHub mirror: github.com/latex3/svn-mirror/issues/166

David Carlisle

@JosephWright error:-)

Joseph Wright

8:24 PM

@DavidCarlisle Agree: want to see what others think, and as you are the keyval expert ...

@DavidCarlisle Do you have a GitHub account?

Paulo Cereda

@JosephWright github.com/davidcarlisle

Joseph Wright

@PauloCereda Great: I've added him to the team as far as GitHub is concerned :-)

Paulo Cereda

@JosephWright Yay!

Joseph Wright

@DavidCarlisle Thanks to @PauloCereda, no escape now!

(I'm in charge of the latex3 GitHub 'user')

Well, actually everyone is down as an owner, but I set it up!

David Carlisle

@JosephWright doesn't it send me a message asking? or can you just assign jobs to anyone with a github account?

@JosephWright @HeikoOberdiek files on way to ctan

Sean Allred

8:30 PM

@DavidCarlisle From what I've done with GitHub, there's no real 'asking' going on. More like a notification, if anything.

Joseph Wright

@DavidCarlisle No apparent need for me to send a request: I can now assign issues to you :-)

@SeanAllred Exactly

David Carlisle

@JosephWright seems like @SeanAllred has a github account: assign all the bugs to him

Sean Allred

@DavidCarlisle Noooooooooooooooooooooooooooooo

Wait, there are only four

Joseph Wright

@DavidCarlisle You've made the mistake of committing to the kernel repo: now I know we've got your attention, ...!

Faheem Mitha

Is there some way to add text between the items of an itemized environment?

Sean Allred

8:33 PM

…because I was limiting to documentation…

David Carlisle

@SeanAllred we could transfer these from gnats: latex-project.org/cgi-bin/…

Sean Allred

@DavidCarlisle That's… that's a lot.

Paulo Cereda

@JosephWright Once in the team, you can assign stuff. :)

David Carlisle

@SeanAllred you should have seen how long it was a month ago before the 2014/05/01 release we closed loads:-)

Joseph Wright

@DavidCarlisle Could you try print (package.config:sub(1,1))) using texlua in your Cygwin rig when you get a chance? I'm working out how to do a system test for updating the L3 build system and it would be handy to know what gets picked up!

David Carlisle

8:39 PM

@JosephWright p.lua:1: unexpected symbol near ')'

@JosephWright I guess the () are supposed to match:-)

Joseph Wright

@DavidCarlisle Did I miscount?

Oh, so I did!

David Carlisle

@JosephWright if I delete a ) I get:

$ texlua p.lua
/

Paulo Cereda

@DavidCarlisle Bah. :)

@Joseph: speaking of which, how can I help? Fork/pull request?

Sean Allred

@PauloCereda Is that a thing that can happen? I'd always thought it was a one-way mirror

Also, we should try to use Travis :)

Paulo Cereda

@SeanAllred Git repository.

David Carlisle

8:42 PM

@JosephWright I'm probably not supposed to run untrusted code off the internet, what does it do?

Paulo Cereda

@SeanAllred I mentioned this earlier today. :)

Sean Allred

(Er, well, my humble opinion :))

@PauloCereda :D

Paulo Cereda

@DavidCarlisle Seems legit.

Joseph Wright

@DavidCarlisle OK, Unix

@SeanAllred The band?

@DavidCarlisle :-)

@DavidCarlisle Tells me what the directory separator is

Sean Allred

@JosephWright Travis CI :) e.g., my research: travis-ci.org/vermiculus/ssa-tool

Joseph Wright

8:44 PM

@PauloCereda Not really: unless I can sell Git to the rest of the team then the 'primary' repo is SVN and only Will can add to the GitHub one

David Carlisle

@JosephWright ah why isn't that something readable like filsystem.config.pathseparator or something:-)

Joseph Wright

@SeanAllred Doesn't sign depressing songs then?

Paulo Cereda

@JosephWright Will, Will, Will, Will Robertson (Queen's We will rock you, probably sounded better inside my head). :)

2

Joseph Wright

@DavidCarlisle Lua doesn't really do OS-dependent stuff at a 'public' level

@WillRobertson See above!

Sean Allred

@JosephWright I never said that. travis-ci.org/vermiculus/ssa-tool/builds/20862881

Paulo Cereda

8:45 PM

@JosephWright I hate you. :)

Joseph Wright

@SeanAllred You obviously don't know the band, then!

@SeanAllred What does it actually do?

Sean Allred

@JosephWright I honestly can't say I do :)

Travis CI is a continuous integration server.

It uses virtual machines (Ubuntu 12.04, I think) to test commits as they come in against a test script.

0=pass, anything else is a fail.

Paulo Cereda

@SeanAllred Buzzword again. :)

David Carlisle

flippin heck it's like when I take over my mum's machine by remote control... Hey there, we're just writing to let you know that you've been automatically subscribed to a repository on GitHub.

4

Sean Allred

@PauloCereda Hehe yeah :) it's a strange phrase, but it continuously tests commits as they come in :) That's the only way it makes sense to me

@JosephWright Travis CI uses a pretty simple YAML doc to describe what should be tested / how to test it: github.com/vermiculus/ssa-tool/blob/literate/.travis.yml (docs.travis-ci.com/user/getting-started)

It gives you a convenient little image that says build passing or build failing so people can know if the test script succeeded without actually running it.

Joseph Wright

8:52 PM

@DavidCarlisle I guess you can always leave again

@SeanAllred How does one set it up then?

Sean Allred

You just sign in to Travis with your GitHub account and add the repo. I think you have to add a file called .travis.yml to the repo first, though.

@JosephWright (see also ^^) the documentation is really, really good.

I've got to run to a piano recital, but I'll be back :) toodles!

Joseph Wright

@SeanAllred Have fun

David Carlisle

@JosephWright no it's fine I'm used to us assigning each other work, been doing that for 20 years, I'm just surprised that github doesn't send a message asking if you want to get write access rather than writing to tell you that you have it:-)

Joseph Wright

@DavidCarlisle I guess they assume if the owner(s) are happy to give out write access, you are happy to have it

@DavidCarlisle Anyway, in our case write access to the Git side of things is meaningless unless you are called 'Will Robertson'!

David Carlisle

@JosephWright is it not possible to set up a reverse mirror to the svn?

Joseph Wright

9:07 PM

@DavidCarlisle ?

@DavidCarlisle From memory, Will and I did some experiments which seemed to show that only he could really make any changes

David Carlisle

@JosephWright do you mean his changes are the only onces that are pushed back or that we can't really commit to the git repository at all?

@JosephWright that is can't you set up a bot that checks any changes in the git mirror into svn, to keep them in sync

Joseph Wright

@DavidCarlisle Thinking a little more, what I tried at the time was syncing the Git and SVN systems, before Will had the automation set up

@DavidCarlisle Think that might lead to some form of loop!

David Carlisle

@JosephWright google suggested this:

ben.lobaugh.net/blog/147853/…

@JosephWright useful tip at the end: And if something went wrong and you are freaking out an anxiously scratching your head I also recommend Scotch :) (I could have done with that advice earlier today)

4

Joseph Wright

9:24 PM

@DavidCarlisle Not sure it we can do this 'after the fact'

David Carlisle

@JosephWright didn't read it in detail, just looked for a site that fleshed out what I had in mind. probably not worth it anyway, we should probably stick with the svn or switch to github and not try to use both for the actual file changes:-)

Joseph Wright

@DavidCarlisle Yes, my position too

David Carlisle

Rainer just pushed the update out

Joseph Wright

@DavidCarlisle Devil is in the detail:

> It should be noted that the sync is triggered by you on your local machine. This is not an automated process on Github's servers, nor will any of the syncing information be available to anyone else who later clones the git repository.

Much as I found: only one person can do the sync

David Carlisle

@JosephWright ah.

Joseph Wright

9:31 PM

@DavidCarlisle I'd be happy to switch to 'native' Git, but I can see objections that I feel would prevent it

David Carlisle

@JosephWright I must admit I still find git more annoying than helpful compared to svn but I can see that the github UI and integrated issue tracker etc has advantages, so i wouldn't be necessarily against switching.

Paulo Cereda

@DavidCarlisle I'd favour Mercurial, but Git is interesting.

David Carlisle

@PauloCereda last time I was involved in a mercurial project I just about got the hang of it and they junked it all and moved to github:-)

Paulo Cereda

@DavidCarlisle LOL

Joseph Wright

@PauloCereda I think the momentum is behind Git: I've used Mercurial in the past, but can see the way the 'wind is blowing'

@PauloCereda With something like SourcTree it's quite easy to use both anyway

Paulo Cereda

9:36 PM

@JosephWright Yep. :)

David Carlisle

@PauloCereda I think we should go back to rcs with sync between different machines by Rainer's mail server sending vvencoded diff files applied by patch invoked by the emacs mail reader. That was a real source control system.

Paulo Cereda

@DavidCarlisle :)

Stephan Lehmke

@DavidCarlisle Trac is a useful UI etc for svn.

David Carlisle

@StephanLehmke yes we use trac a lot at work but we don't have that set up on the public latex svn and it's probably as easy to move to github as set up a new public trac over an existing svn

Stephan Lehmke

@DavidCarlisle Ok, I don't know how difficult it is to set up trac. And I also don't know git. But my experience with bzr is that I haven't yet managed once to use it right.

Joseph Wright

9:41 PM

@StephanLehmke :-)

@StephanLehmke The thing with GitHub (or similar) is that you leave the UI, etc. to people who do that as their day-to-day: particularly for a small team, that's a bonus

David Carlisle

@StephanLehmke the good thing about git is that it is so arcane and hard to do even the simplest thing like check in a file, that dozens of people have asked about every step so google or stackoverflow always have plenty of step by step guides.

Stephan Lehmke

A lot of teams in our company use git, but I think there is only one single person who knows how it's used correctly. Mostly everyone else is deeply frustrated most of the time.

Paulo Cereda

@David: let's Dropbox a big folder and share files that way. :)

David Carlisle

@PauloCereda who'd do a thing like that....

Paulo Cereda

@DavidCarlisle :)

Werner

10:20 PM

Anybody have an idea for an "un-nesting procedure" that would take \ABC{some \DEF{nested} stuff} and transform it to \ABC{some }\DEF{nested}\ABC{ stuff}?

David Carlisle

@Werner yes

Werner

@DavidCarlisle Should I post something?

...this almost feels like match-fixing...

David Carlisle

@Werner nah I'll answer here in a sec

Werner

@DavidCarlisle Thanks!

David Carlisle

@Werner if my fingers had typed what was in my head it would have been done ages ago:-)

\def\zz#1#2{\zzz#2\DEF\relax}

\def\zzz#1\DEF#2{%
\unexpanded{\ABC{#1}}%
\ifx\relax#2\else
\unexpanded{\DEF{#2}}%
\expandafter\zzz
\fi}

\edef\foo{\zz\ABC{some \DEF{nested} stuff}}

\show\foo

\end

> \foo=macro:
->\ABC {some }\DEF {nested}\ABC { stuff}.

Werner

10:30 PM

@DavidCarlisle :)

@DavidCarlisle Thanks David... this is a good start... slow fingers or not!

David Carlisle

I had more } than { which didn't help. then because @JosephWright chided me earlier today for using \toks register instead of \showtokens I thought I'd use \unexpanded but of course because it's plain I tried tex and got Undefined control sequence.
\zzz #1\DEF #2->\unexpanded error:(

Heiko Oberdiek

@DavidCarlisle @JosephWright base/00readme.txt is gone?

Werner

@DavidCarlisle So this is specific to matching a nested \DEF. A generalized unnesting procedure seems not easy then.

David Carlisle

@Werner hey you changed the spec:- which things get unnesed and which things get wrapped in ABC ? and also does it have to work by expansion only (it's easier to support pultiple unnested tokens if you can do local assignments

@HeikoOberdiek 00readme.txt should be there? when did it go? et me check...

Heiko Oberdiek

@DavidCarlisle It is a link to README AFAIK.

David Carlisle

10:43 PM

@HeikoOberdiek yes or copy or whatever the ctan host supports, but base in svn has 00readme.txt why isn't it on ctan I'd better check the scripts again.....

@HeikoOberdiek it seems to be in the zip file submitted as far as I can see (Franks dropbox) I just re-submitted doc and graphics but base should be that.

Heiko Oberdiek

@DavidCarlisle I have now patched base.zip from CTAN to add 00readme.txt. Now my build script for latex-tds works and can upload latex-tds.

David Carlisle

@HeikoOberdiek looking at filedates here looks like CTAN moved onstead of copied 00readme.txt to README (I think they need that name fas it shows automatically in the web view) I'll ask them to copy it back:-)

@HeikoOberdiek anorien.csc.warwick.ac.uk/mirrors/CTAN/macros/latex/base

Heiko Oberdiek

@DavidCarlisle Upload of latex-tds is on its way (ETA one hour) -- Karl is already waiting ...

David Carlisle

@HeikoOberdiek Thanks.

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends