TeX, LaTeX and Friends - 2016-08-03

« first day (2103 days earlier) ← previous day next day → last day (2831 days later) »

12:00 AM

ouch, I figured as much and am downloading TL2015 Thanks for the confirmation. I guess I could post a question about how to modify the code.

@AFeldman The list of removed stuff is listed in section 9.3 of the LuaTeX manual.

Ok, is there a list of substitutions?

equivalents

@AFeldman Before you do that try loading the luatex85 package, which is a sort of compatibility layer.

ok as in \usepackage{luatex85} ?

@AFeldman Yes, that makes the document compile for me with TL2016

@AFeldman I put it right after the \ifluatex line.

12:08 AM

Thanks Alan. Much appreciated. Would you like me to post it as a question?

I think that it would be helpful for anyone using that bates solution.

@AFeldman I would add a comment to the answer and maybe suggest that it be updated to reflect that.

Okay. I'll do that and attribute it to you.

@AFeldman But I would be wary of depending too heavily on LuaTeX, since it is really under development, and (as with this last change) the developers are not too concerned about backwards compatibility.

@AFeldman What OS are you using? If you're using MacTeX, it's really easy to maintain multiple years on one system (that's how I quickly tested your document).

You know I would love to not depend on it, but the task that I had, described in that post is really complex, and Touhami was the only one to come up with a solution. Ubuntu based, debian, so I'm not able to take advantage of the considerable Mac tools.

@AFeldman You can do it in principle with Linux too, but just not with a click or too. :)

12:14 AM

Thanks for saving me the pain.

@AFeldman You're welcome.

Now I have no excuse to procrastinate doing actual work any further. :)

@AFeldman :)

@egreg A bit late to the party, but Happy Birthday!

@DavidCarlisle I see you're grumbling about thesis classes. While I agree that CTAN may not be the best place for them, it's still the most visible. I'm much happier having my class there than depending on google searches to find my personal web page. And my institution will not sanction any 'template' so I can't get a link e.g. on the grad school page.

bye all thanks for eliminating my excuses ;)

6 hours later…

6:44 AM

@AlanMunn that's the problem though. I agree it's better/easier for your institution to have your thesis class on ctan and pushed to your students via texlive tlmgr, I'm not at all sure it's better for your institution to get (what could easily soon be) hundreds or thousands of thesis classes via the same route.

@AFeldman if all else fails you could write a new tool to automate tex workflows

@PauloCereda ^^^

@AFeldman that's what I was referring to in my comment on the question. luatex post 0.8 is mostly completely incompatible with previous luatex and with pdftex, which means that luatex-specific packages (that really use lua features rather than just work cross engine) mostly have to update to work just on the new code and not support older releases (which were flagged as beta anyway) so it's not like classic tex or pdftex where actually you can use a 20 year old distribution and actually almost..

@AFeldman everything will work, either you need to freeze completely and not update any packages or you need to bite the bullet and update and get over the incompatibility issues. Unless you have a major document to finish in a couple of weeks, I'd do the latter.

7:47 AM

@AlanMunn Why won't they? There are already thesis classes for various institutions in CTAN. Can't they see that and be convinced?

8:38 AM

@wilx it's the thesis classes on ctan that I was complaining about (and @AlanMunn was defending:-)

@DavidCarlisle Oh.

8:54 AM

@wilx the problem is that texlive and miktex update work "too well" so it's easier for an institution to get a thesis template that only works in one department to its students by pushing to ctan and so the whole world. It seems like every month texlive gets yet another thesis class and it seems wrong. hard disks may be cheap, and searching input trees may be quick comparatively, but it still seems like abusing the good will of the community to use a global distribution system in this way.

9:12 AM

anyone got a non updated 2016/03/31 latex release (unpatched) and can try the test file at latex-project.org/cgi-bin/ltxbugs2html?pr=latex/… I can't reproduce the error with a current texlive 2016 system, and don't really want to build an old one....

forget the above, I see the problem

(./cambridge6A.cls
Document Class: cambridge6A 2010/09/09, v2.10 LaTeX2e document class

@DavidCarlisle ^^^^^

@DavidCarlisle The loading of packages is wrong: bm is before amsmath, for instance, and amsmath is after hyperref. No way that's a good bug report.

@egreg yes lower down I extracted a better test file but anyway I just realised it's the amsmath \dots\left bug fixed already in texlive

@DavidCarlisle Here's the diff of the two snippets

6c6
< \left( \frac{1}{2} - n + 1 \right)}
---
> \big( \frac{1}{2} - n + 1 \big)}

@egreg yes, what I hadn't noticed until just now is that the line above end in \dots

@DavidCarlisle Yes, that's the reason.

amsmath 2016/03/10 v2.15b

9:25 AM

@egreg I think it's a bug in tex that \long isn't the only primitive that starts with l

@DavidCarlisle :)

10:00 AM

@DavidCarlisle o.O

10:33 AM

@egreg: I have a new duck!

@PauloCereda :)

@egreg Want to see him? :)

11:10 AM

Hello, everybody!

I need help. I can't figure this one out. I think it is a bug. I keep getting
[507
! Huge page cannot be shipped out.
\@EveryShipout@Output ...@Org@Shipout \box \@cclv

l.48165 \]

and I see nothing wrong in the file. The problem is the Latex file is too large, about 60,000 lines. I can't make a MWE. Can I post this question and give a link to the file?

TL 2016 on Linux.

@Nasser well, you can (technically), but it's not a good idea; the question would be pretty bad.

@yo' so what should I do then?

@Nasser well, debug the problem yourself; it's never impossible to cut down the file to the minimum required.

@yo' I do not know where the error is coming from. There is nothing wrong with the line in error. if I remove code, the problem goes away.

11:15 AM

@ Sometimes the user will generate a huge page because other error messages
are being ignored. Such pages are not output to the \.{dvi} file, since they
may confuse the printing software.

@<Update the values of |max_h| and |max_v|; but if the page is too large...@>=
if (height(p)>max_dimen)or@|(depth(p)>max_dimen)or@|
   (height(p)+depth(p)+v_offset>max_dimen)or@|
   (width(p)+h_offset>max_dimen) then
  begin print_err("Huge page cannot be shipped out");
@.Huge page...@>

@Nasser ^^^^^

The line where the error is reported is not relevant, that is just where tex got to but tex is always working ahead of itself to optimise the page breaks, it is the earlier code that is wrong, you have generated a page that is bigger than \maxdimen

@nasser the code goes on to say:

  error;
  if tracing_output<=0 then
    begin begin_diagnostic;
    print_nl("The following box has been deleted:");
@.The following...deleted@>

so after that error message it should show you the box that got zapped,

@DavidCarlisle thanks. But the earlier line it shows in the error message is line 507. When I look at line 507, I see nothing wrong there. so I really have no idea why Latex is making a page from line 507 to line 48165?

Overfull \hbox (461.54228pt too wide) in paragraph at lines 48140--48141
[][]\TU/lmtt/m/n/10.95 DSolve[2*Derivative[1][f][x]*(-(a^2*Derivative[1][y][x])
 + Derivative[3][y][x]) + f[x]*(a^4*y[x] - 2*a^2*Derivative[2][y][x] + Derivati
ve[4][y][x]) == 0, y[x], x]
[507
! Huge page cannot be shipped out.

Here is around line 507....

what is after the ?

user image

no in the log

@DavidCarlisle Ok, will look at log now..

I hit X first to stop the compile.

11:22 AM

@Nasser so you ask not to get the error help....

@DavidCarlisle there is nothing new in the log file. I hit X, the compile stops and I see in the log file the same as on the terminal.

Overfull \hbox (461.54228pt too wide) in paragraph at lines 48140--48141
[][]\TU/lmtt/m/n/10.95 DSolve[2*Derivative[1][f][x]*(-(a^2*Derivative[1][y][x])
 + Derivative[3][y][x]) + f[x]*(a^4*y[x] - 2*a^2*Derivative[2][y][x] + Derivati
ve[4][y][x]) == 0, y[x], x]
 []

[507
! Huge page cannot be shipped out.
\@EveryShipout@Output ...@Org@Shipout \box \@cclv

l.48165 \]

I do not know what to do.

I do not know what you mean by "ask not to get the error help"

@Nasser don't type x which means you don't want to get the further help (or anything else) just hit return

@DavidCarlisle ok, will try again and will not hit X ...

@DavidCarlisle I hit return, and now it is compiling... and finished ! what does this all mean?

@Nasser type h woul dbe best

@Nasser why the ! that was expected, tex simply refuse to output one page that is too big to output, so it will then carry on now look at the log

@DavidCarlisle this is now in the log file ( I hit return instead of X)

Overfull \hbox (461.54228pt too wide) in paragraph at lines 48140--48141
[][]\TU/lmtt/m/n/10.95 DSolve[2*Derivative[1][f][x]*(-(a^2*Derivative[1][y][x])
 + Derivative[3][y][x]) + f[x]*(a^4*y[x] - 2*a^2*Derivative[2][y][x] + Derivati
ve[4][y][x]) == 0, y[x], x]
 []

[507
! Huge page cannot be shipped out.
\@EveryShipout@Output ...@Org@Shipout \box \@cclv

l.48165 \]

?
The following box has been deleted: []

]
Overfull \hbox (326.14421pt too wide) detected at line 48198

@DavidCarlisle but I do not understand which page that is too long. And where it came from. Line 48165 does not have large page. Just one equation that is 300 chars long

11:31 AM

@Nasser OK well you could increas \showboxdepth and \showboxbreadth (to say 1000) so that The following box has been deleted: [] showed the box contents not just [] but anyway i assume it is the display math that ends on line 48165 (oh you said line 507!!! it is page 507) that is simply too big to fit on a page

@DavidCarlisle oh I see. So 507 is page number. I thought it was line number. Ok. I got the pdf file now and will try to see what is around that page number...

@Nasser that is the line at which tex considered making a page break, it may be the code immediately before that there is something bigger than 18metres high (or wide) it can't be that hard to spot, surely

@DavidCarlisle He could buy iTeX* and the problem would be solved. :)

@DavidCarlisle you are right. I see page 507 missing from the PDF file! I see page 506 then page 508. No page 507. WIll now zoom in and see which equation was in the middle.

@Nasser which is exactly what the error says: [507 ! Huge page cannot be shipped out.

@PauloCereda or Word

11:38 AM

@DavidCarlisle Microsoft wanted to charge close to 50% more for Office for Mac than the Windows counterpart. I believe MS think Mac people have a lot of money. @JosephWright @AlanMunn @egreg :)

@PauloCereda They are always right

@DavidCarlisle :)

@DavidCarlisle I found something. I found an equation that is 20,378 letters long on that page. I think this was the problem. [ ... 20378 letters ..] too big for latex.

@JosephWright: The mail bloke just delivered my TeX Collection 2016 DVD. Thank you! UK-TUG rocks!

@Nasser number of letters is not very relevant x\\[170000pt]y only has two letters and would make the same error.

@PauloCereda a new mat for your coffee mug?

11:43 AM

@DavidCarlisle That's what thesis drafts are for. :)

@Nasser or \[x\hspace{170000pt}y\] if it is too wide rather than too tall

@PauloCereda I submitted my first draft, so didn't have that luxury.

@DavidCarlisle Ok thanks, will remove this equation from the file now.

@PauloCereda here at work I use an acrobat 3.01 CD

@DavidCarlisle Oh my! :)

@PauloCereda i think I've been using it for a while now....

11:48 AM

@DavidCarlisle Acrobat of the coffee mug? :)

@PauloCereda both

@DavidCarlisle ooh is the NAG coffee good?

\noindent \flushleft \subsection*{\Large{Profile}}\\ spot the command that shouldn't be there (or easier question for beginners: spot the command that should be there)

@PauloCereda we have a filter coffee and grinder, and a nespresso capsule thing (that I have not used) but normally I just use the instant as it's hot and brown and does the job (don't tell @egreg:-)

@DavidCarlisle :-)

btw, I'm on a summer camp and we've got several cases of head louse :-(

@DavidCarlisle ooh

@yo' oh no

12:12 PM

@DavidCarlisle \\ is extra?

12:25 PM

@wilx \\ , \noindent, \flushleft and \Large should not be there, the rest is Ok

@DavidCarlisle Ah, heh, ok. :)

@DavidCarlisle Never drink anything brown

@egreg How about Coca-Cola? :)

@egreg so that's coffee tea beer cocoa out

@DavidCarlisle What if the glass is opaque and with a colour other than brown and we never spot the liquid's original colour? It's like a Schroedingan beverage. :)

12:38 PM

@PauloCereda simpler: just drink what you want, and if egreg asks, lie.

@DavidCarlisle ooh

user image

2

This is the new duck I won!

@DavidCarlisle, @PauloCereda I see Will has moved his blog stuff to GitHub pages: I'm thinking of looking at this, though he's warned me off!

I suspect 'link stability' might be tricky!

@JosephWright where was it before, wordpress?

@DavidCarlisle Yes, it's currently Wordpress (self-hosted, not Wordpress.com)

@DavidCarlisle anything brown that advertises as coffee

12:48 PM

@DavidCarlisle At the mo', I think that might continue: moving it looks more complex than perhaps I fancy!

@egreg I never drink coffee

@PauloCereda see ^^^

@JosephWright It looks like an interesting plan, I've seen something along these lines.

@DavidCarlisle ooh :)

@JosephWright let's see if we get latex-project.org sorted first:-) (seems if we want to preview we need to have two instances of jekyll somewhere)

@DavidCarlisle If it's brown then it's definitely not coffee

@egreg yeh yeh and if it's written after 1750 it's definitely not music, I know.

12:53 PM

@egreg what if I put milk inside the coffee? :)

@JosephWright I was wondering of it would be simpler just to have one branch at the latex3 github and anyone editing and wanting a preview could run their own gh-pages fork and then pull request things to latex3 repo once done. But see what Jonas found from gh support..

> Paulo, people are paying attention to your research.

@DavidCarlisle: ^^ sounds fishy

@PauloCereda s/people are/you should be/

@DavidCarlisle oh no

@DavidCarlisle Damn. By redefining \KV@errx. The #1 argument is ended by the string "undefined".

12:57 PM

@wget quite possibly. It's not like I've looked at that code more than a couple of times since I wrote it in 1993:-) I assume you mean this bit

\ifx\@tempc\relax
  \KV@errx
   {\@tempa\space undefined}%

Damn. 1993. By a few days, this would have been by year of birth :D

@DavidCarlisle: ^^ he called you old.

@PauloCereda we shouldn't allow children on the site

@PauloCereda not as old as you-know-who

1:23 PM

@DavidCarlisle Good plan

1:38 PM

@DavidCarlisle I have some doubt about the * too ..., normally his means that secnumdepth should be changed.

@PauloCereda Yes, we are all super rich. And I buy a new version of Office monthly.

@AlanMunn Just because you can, right? :)

@PauloCereda Exactly.

2:18 PM

Am I too crazy for getting me one of these?

The slide rule, also known colloquially in the United States as a slipstick, is a mechanical analog computer. The slide rule is used primarily for multiplication and division, and also for functions such as roots, logarithms and trigonometry, but is not normally used for addition or subtraction. Though similar in name and appearance to a standard ruler, the slide rule is not ordinarily used for measuring length or drawing straight lines. Slide rules exist in a diverse range of styles and generally appear in a linear or circular form with a standardized set of markings (scales) essential to performing...

@egreg, @DavidCarlisle: ^^

@PauloCereda we had to use those at school.

@DavidCarlisle ooh so it's ~~old~~ a cool device. :)

@PauloCereda then moved on to log tables, but by the time I left school both had been more or less completely replaced by a new fangled device called an electronic calculator.

@DavidCarlisle it will never catch on. :)

@PauloCereda I think I had one about like this datamath.org/Sci/Slimline/TI-25.htm

2:24 PM

@DavidCarlisle Sorry I didn't mean you were old. ;-) I wouldn't have allowed myself any way.
I debugged my issue a bit further. 1) xkeyval is still buggy and I completely removed it 2) pgfkeys is easier to use, and *CAN*, ctrontrary to what I announced, handle key={value,value}. As soon I add babel, pgfkeys cannot handle that use case any more :-/

@PauloCereda Do you really want one? I have a pocket version here. But there is no documentation ...

@UlrikeFischer he needs one. If he had a slide rule he could do calculations like 50000 words at 2 words per day = ... days... so he could work out when he'll finish his thesis.

2:40 PM

@UlrikeFischer I got one mostly because I like those mechanical things. :) Yesterday we were having fun talking about engineering and stuff, so I had an itch to get me a slide rule. :) I think I am nerd. :)

@DavidCarlisle ooh a TI

@PauloCereda Penitentiagite

@PauloCereda there were only TI and Casio, really.

@PauloCereda I took great pride in using them in high school. It was cool (if you were a nerd). And my father had a couple of really fancy ones.

@AlanMunn Really? How awesome!

@Alan: when you come to São Carlos, remember to show the nice slide ruler they used to have in the CS department in UFSCar. It was a great sight: dc.ufscar.br/imagens/regua_calculo_s.jpg/image_view_fullscreen

@egreg :)

@DavidCarlisle oh

@PauloCereda actually that model was 1978 which is a bit late I think I had the 30 datamath.org/Sci/MAJESTIC/TI-30.htm

2:53 PM

@DavidCarlisle No Roman numerals? What was this world going to.

@DavidCarlisle Surely there were also HPs? That's what I had, although maybe I had a TI first. The HP had special cachet because not only did it look so much cooler and high quality, but it also used RPN which gave it special nerd status.

@PauloCereda what have the romans ever done for us

@DavidCarlisle :D

@AlanMunn yes I had access to an HP RPN box somewhere but I don't recall anyone having one at school there was also the programmable sinclair my maths teacher had one of those and let us play sometimes.

Yesterday, in the cafe (cute girl approaches):
- Is this seat taken?
- Not at all.
- Thanks (grabs seat and leaves).

2:58 PM

@DavidCarlisle @PauloCereda I had one of these: hpmuseum.org/hp65.htm

@PauloCereda What did you expect? :P

@AlanMunn ooh I like the keys.

@AlanMunn magnetic card strips, forgot all about that. I did have access to one of those, just can't remember where:-)

@AlanMunn "Please, tell me more about macro expansions!"

Great, @David and @Alan are about to perform the Four Yorkshiremen sketch. :)

@PauloCereda <3

@AlanMunn A man can dream, right? :)

3:01 PM

@PauloCereda Nah, because you kids today do believe us. :)

@AlanMunn LOLOLOL

@PauloCereda go away and finish your thesis

@DavidCarlisle oh no

1 hour later…

4:30 PM

@PauloCereda That was the first calculator I used.

@egreg :)

@DavidCarlisle I have finally filled a questionabout my issue: tex.stackexchange.com/q/322217/66964

@egreg since you have posted a bunch of answer wrt. pgfkeys, your help on this topic is much appreciated ;-)

@wget Sorry, no pgfkeys ;-)

@egreg are you sure? :D tex.stackexchange.com/a/167214/66964

@wget That was just fixing some given code. :)

4:35 PM

Maybe indeed pgfkeys is not your preferred topic ;-)

4:56 PM

@wget As far as I can see, languages={dutch,french} confuses the parser, whereas language=dutch alone doesn't.

Yep. But as soon you remove the line with babel, this is working fine.

I think this is due to babel which modifies catcodes and fools the parser. But I may be wrong.

And obviously the purpose here is to be able to define several languages. Most of my document have several languages in it.

5:30 PM

@wget The problem is that babel is the only package in your list that uses \ProcessOptions*. Packages that use \ProcessOptions (without the *) have no problem.

5:41 PM

@egreg I have a feeling I might need to look at this ...

@DavidCarlisle I see the LaTeX team website has updated

Just updated MacTeX and TeXLive Utility is saying "unable to find a valid update server".

"Either a network problem exists or the TeXLive version on this server does not match. . . . You may need to try a different repository."

5:58 PM

@JosephWright oh I looked a while ago and it hadn't:-)

@Gregory Are you updating MacTeX 2015? Or do you mean you just installed MacTeX 2016?

@wget I get the same error with or without babel

Just installed MacTeX 2016.

I set mirror.ctan.org/systems/texlive/tlnet as my home repository.

@JosephWright I see _site has been re-committed, still not convinced that's right, we'll see...

@JosephWright Ask @DavidCarlisle; it surely is his fault.

6:01 PM

@Gregory That's certainly the correct choice. It could just be a transient error. You could also log at the TL Utility log window to see if there's something there.

@AlanMunn it's less transient than I'd like :-/.

From the log: 2016-08-03 18:06:10 +0000 Notice -[TLMOperation main][3644] Failed executing

/usr/local/texlive/2015/bin/x86_64-darwin/tlmgr --machine-readable --repository http://mirror.ctan.org/systems/texlive/tlnet update --list --all

(error 25)

Seems it's looking in texlive/2015. Should it be?

@DavidCarlisle You are right. It's not entirely babel fault. xeCJK, csquotes, are also concerned for example.

@Gregory no

@wget so what do I need to remove to get a working version of your test, the question says it works without babel

@greg @DavidCarlisle Why the hell by default without using these package this is working then...?

@wget can't parse that

6:10 PM

I presume this is happening because I just upgraded from 2015, but how do I fix it?

@DavidCarlisle It seems using dichotomy to search for the reponsible package wasn't enough...

@Gregory No, but that's helpful information. It shouldn't happen, though.

Also from log: 2016-08-03 18:06:10 +0000 Notice -[TLMOperation main][3644] Unknown directive ...containerchecksum c59200574a316416a23695c258edf3a32531fbda43ccdc09360ee105c3f07f9fb77df17c4ba4c2ea‌4f3a5ea6667e064b51e3d8c2fe6c984ba3e71b4e32716955... , please fix it! at /usr/local/texlive/2015/tlpkg/TeXLive/TLPOBJ.pm line 210, <$retfh> line 5579.

@Gregory Can you open a Terminal window and type tlmgr --version and see what it says?

tlmgr revision 41258 (2016-05-19 17:23:37 +0200)
tlmgr using installation: /usr/local/texlive/2016
TeX Live (http://tug.org/texlive) version 2016

6:13 PM

@Gregory Ok, that's correct, so it seems the problem is with TL Utility. Try going to the Configure menu and choose Change Default TeX Live version.

That should bring up both TL 2015 and TL 2016 (assuming you didn't delete 2015).

Configure > Reconfigure Distributions brought up a menu including TL 2015 and TL 2016. TL 2016 was already selected.

@DavidCarlisle I modified my question accordingly and have retested all packages: tex.stackexchange.com/q/322217/66964

@Gregory :( even stranger. Sorry I'm out of ideas a the moment.

@wget I guess I'll have to have a look: this might be my fault in pgfopts

@JosephWright Yes, please have a look.
damn indeed. you are the author of pgfopts, I should have asked to you first...

6:24 PM

@AlanMunn in the TeXLive Utility, if I click the packages tab then I get a list of installed packages. It's just if I click the update tab that I get "listing failed". Does that add any info?

@Gregory Actually, one more idea: download the latest version of TL Utility instead of using the one that came with MacTeX? There's a Check for Updates menu item in TL Utility.

@Gregory No, that's consistent with the other behaviour you're seeing.

@JosephWright Btw, I notice, the documentation I have been reading dates back from 2010-05-01. I don't know if your package has been updated since.

@wget Nope: there's been no need to make any changes. The actual idea is copied from kvoptions, it's just a question of where the output gets sent (pgfkeys or keyval or ...)

@JosephWright In my use case what can I do to prevent the output from being sent to these packages which (according to what you say) are using the faulty keyval, xkeyval, etc. ?

@wget Sorry, I don't follow: the point of pgfopts is to allow option processing (\usepackage and \documentclass optional arguments) by pgfkeys. That's all it does and doesn't interface with any other package.

@wget I may have messed up some aspect of grabbing the data to pass through

6:33 PM

@JosephWright So why are these packages (the one I mentionned in my question like babel, csquotes, etc.) complaining while I'm not even passing data to them?

*the ones

@wget Like I say, I need to check if I've messed up the data structures that get passed through

@JosephWright Do you have the time to check that now? ;-)
Otherwise, I'll need to find another solution, requiring me parsing manually, etc. :-/

@wget I am looking

@AlanMunn I've updated the TeXLive Utility. Now it keeps prompting me to upgrade my TeXLive version from 2015 to 2016, which I believe that I already did when I installed MacTeX 2016. Could this be a path issue?

@Gregory Yes, it's looking like that. What's weird though is that when you ran tlmgr from the terminal it found the correct version.

6:42 PM

@wget It's not my fault :-)

\RequirePackage{filecontents}
\begin{filecontents*}{testtemplate.cls}
\ProvidesClass{testtemplate}
\LoadClass[twoside]{report}
\RequirePackage[english]{babel}
\end{filecontents*}
\documentclass[someOtherKey=hello world,languages={dutch,french}]{testtemplate}
\begin{document}
    Hello world
\end{document}

@DavidCarlisle ^^^

@Gregory In the preferences of TL Utility, does it show this:

@JosephWright Yep obviously ;-) I said using the pgfkeys pgfopts stuff like kthat with babel or the packages I mentioned is not working...

user image

@wget You've lost me

@AlanMunn I got a message from TL Utility that there was a TL infrastructure update available so I installed it. Not the updates tab shows a bunch of packages with updates available, but I'm still getting the message that I need to upgrade from 2015 to 2016 and the URL shown in the utility window is for 2015. I'll check the settings you sent.

6:45 PM

@JosephWright --> my question is here --> tex.stackexchange.com/q/322217/66964

@wget Yes, I know, but my point is that the error is not caused by using pgfopts but rather as parts of the LaTeX kernel don't like braces in options, and that shows up when you load babel. The 2e kernel was never designed to allow keyvals at all, so it's all a bit delicate

@AlanMunn. The path given in preferences is /usr/local/texlive/2015/bin/x86_64-darwin

@JosephWright Ok. Now I assume the only solution I have is to modify lightly my structure
instead of
\documentclass[someOtherKey=hello world,languages={dutch,french}]{testtemplate}
use this:
\documentclass[someOtherKey=hello world,languages=dutch|french]{testtemplate}

But in that case, I'll need help to help me parse the dutch|french list and convert it to dutch,french ;-)

@AlanMunn, but my shell path includes :/Library/TeX/texbin

@Gregory Ok, so click on the Choose button, and then do the following (don't navigate to a folder with the mouse, but instead use Command-Shift-G and then enter /Library/TeX/texbin/ and maybe that will fix it.

6:53 PM

@AlanMunn You're my hero. Everything appears to be working now.

@Gregory It's odd that it happened in the first place, but I'm glad we got it fixed.

@JosephWright

         \@expandtwoargs\in@{,\CurrentOption,}{,\@declaredoptions,}%

@DavidCarlisle Yes, I guessed \in@ was involved

can we blame Leslie for \in@ when \CurrentOption has a brace group

@DavidCarlisle Not saying it's an error (as this has never been supported)

@DavidCarlisle Blaming Leslie a great plan

7:01 PM

@JosephWright well that's what I told @wget a day or so back, not to use \processOptions but just take \@classoptionlist` and pass it to \setkeys (if using keyval)

@DavidCarlisle Still chokes if babel is loaded though

@JosephWright well not necessarily (I'd use \setkeys, handle any keys i wanted and rebuild a clean no-keyval \@classoptionlist for other packages to handle

@DavidCarlisle OK, doable if you don't use \in@: l3keys and the known keys business can do it

@JosephWright but since it's just an in@ test probably can make it safe with a version that adds \unexpanded or \detokenize somewhere, i'll have a look

@DavidCarlisle In other news, just finished a very interesting Skype-type meeting concerning publisher support of maths (LaTeX to XML-based)

7:06 PM

@JosephWright eventually we'll get there and everyone will be using tex and xml (and emacs of course)

@DavidCarlisle Perhaps

@JosephWright using a tex to xml convertor that we know?

@DavidCarlisle Not sure: will mail some info

@JosephWright thanks. meanwhile i make processoptions sort of half safe but babel reparses it all again by hand with more in@:

\ProcessOptions*
\ifx\bbl@opt@main\@nnil
  \edef\bbl@tempa{\@classoptionslist,\bbl@language@opts}
  \let\bbl@tempc\@empty
  \bbl@for\bbl@tempb\bbl@tempa{%
    \@expandtwoargs\in@{,\bbl@tempb,}{,\bbl@loaded,}%

plan b I think...

@JosephWright Re: meta.tex.stackexchange.com/q/6813/2693 Can we change the text that is displayed in the box when you Ask a Question?

7:17 PM

@DavidCarlisle l3keys it is (speaking of which, I really must finish fixing the performance of l3keys)

@AlanMunn Powers That Be would need to do it

@JosephWright That's too bad. Is this something that they would listen to if you asked?

7:29 PM

\makeatletter
\let\oldin@\in@
\def\zzz#1#2#{\oldin@{#2}{#1}\zzzb}
\def\in@#1#2{\zzz{#2}#1{}\@nil}
\def\zzzb#1\@nil{}
\makeatother

\RequirePackage{filecontents}
\begin{filecontents*}{testtemplate.cls}
\ProvidesClass{testtemplate}
\LoadClass[twoside]{report}
\RequirePackage[english]{babel}
\end{filecontents*}
\documentclass[someOtherKey=hello world,languages={dutch,french}]{testtemplate}
\begin{document}
    Hello world
\end{document}

@JosephWright @wget ^^^^^^^^^

@DavidCarlisle Hum. Before claiming victory.... I need to keep calm :D
I executed that piece of magic and yes this is working.
*BUT* I'm now testing deeper ;-)
removing debug info and stuff like that....

7:54 PM

@wget note it essentially removes the brace (and anything after it) so your languages={foo,bar,whizz} gets tested by \in@ as used by classic processoptions as languages= but since any package using classic \processoptions isn't expecting key=value options at all this is probably safe enough

Don't hesitate to bring these details on the answer you made. ;-)

On my side all is compiling. I'm just adding these changes to the production template aI have and I'm compiling some document to see if some things get broken.

8:10 PM

user image

Steven B. Segletes

Is it just me or have other noticed the monthly rep competition is screwed up: tex.stackexchange.com/users?tab=Reputation&filter=month? After 3 days of the month, many are listed with over 1000 rep, as much as 8000 in egreg's case. Some counter did not reset, it would seem.

@PauloCereda LOL

@AlanMunn <3

@PauloCereda If I'm not back again this time tomorrow, carry on.

@AlanMunn as if nothing really matters? :)

8:15 PM

@PauloCereda Easy come, easy go.

@AlanMunn ooh I like the looks of this following conversation. :)

user image

Steven B. Segletes

@PauloCereda Mama Mia!

I always get confused with the Fandango thingy. :)

@StevenB.Segletes Let me go!

Steven B. Segletes

@AlanMunn we will not let you go!

8:17 PM

@StevenB.Segletes Thunderbolt and lightning, very, very frightening me!

@PauloCereda Where's Galileo when you need him?

Bohemian Rhapsody | Muppet Music Video | The Muppets

Steven B. Segletes

@AlanMunn So who has the nerve to say "Beelzebub has a devil put aside for me"?

@AlanMunn :)

I need to bring this back:

Jul 19 '12 at 21:53, by Alan Munn

I see a little silhouetto of a duck
Does it bark, does bark does it do the fandango?
Compilation errors, control sequence terrors
David Carlisle, David Carlisle
David Carlisle, David Carlisle
David Carlisle Enrico. Gregorio.

This is pure gold.

Steven B. Segletes

@DavidCarlisle LOL!

8:21 PM

@StevenB.Segletes some band calling themselves Queen did a cover version later but I think that's the original

@DavidCarlisle is the bloke of Edinburgh part of the band? :)

Steven B. Segletes

@PauloCereda Freddie Muppetry?

@StevenB.Segletes LOL

@PauloCereda yes here he is playing guitar on the roof of his house

Brian May - God Save The Queen (HQ).MP4

@DavidCarlisle ooh

8:31 PM

@PauloCereda one of @egreg's musical favourites.

1 hour later…

9:37 PM

@DavidCarlisle Ok working on my side. Passed on my tests. It's party time \o/
Joke aside, a BIG thank to you and to @JosephWright et @egrep who helped me on this topic.
@DavidCarlisle I tried to add the comments you made in this chat to the answer you made, but my comment cannot format the latex code properly
http://tex.stackexchange.com/a/322238/66964

Answer accepted then ;-)

@wget you can't format code in comments other than the inline backtick

Ok :-(

@DavidCarlisle then maybe adding that code below your answer would help others

I don't want to edit your answer this is is YOUR answer ;-)

*since this is

@wget not sure it helps much: the main issue is \processoptions* and the fact that babel also has a couple of "by hand" uses of \in@ is true but very specific to babel. I don't know what you meant by " your hack is actually removing it." since your second comment said that you didn't say it removed the code.

What I meant is that your code renders ineffective (or somewhat replacing thanks to \def and \let statements, this is black magic for me) the \in@from babel code And apparently other packages working the same way (csquote, etc.)

@wget well no: \in@ is doing the same tests it was (so tests for classic options such as a4paper or german will work, it just removes brace groups from any options before testing so languages={a,b,c} is tested as languages= the test still happens and the true or false branch is executed

9:52 PM

@DavidCarlisle ok and this is YOUR code that removes the "braces groups from any options" before LaTeX could even test the package/class options. Right ? now? ;D

@wget during rather than before, but yes. (I considered pre-processing the list to remove the groups but this doesn't do that it just adjusts the tests so that the {} group is ignored while testing the actual option list isn't changed and the option itself still has the group

Ok understood.
Again, if you could add that explanation to your answer below my question this would be great ;-)

Going to bed now. Bye. And thanks again to all o/

@wget that was the intended meaning of "this just removes brace groups from options as seen by the standard option parser " :-) but I might expand it a bit:-)

10:27 PM

@JosephWright I closed tex.stackexchange.com/questions/322251/… as duplicate of the canonical tex.stackexchange.com/questions/31091/… but it was reopened. I can't see where it's different.

11:00 PM

@egreg you've turned into a regular expression matcher "a BIG thank to you and to @JosephWright et @egrep " :)

« first day (2103 days earlier) ← previous day next day → last day (2831 days later) »

Transcript for

Aug '163

$TeX - LaTeX$ 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