« first day (3184 days earlier)      last day (28 days later) » 

12:20 AM
Why doesn't \verb work in parboxes? It's not an issue but I'm curious
2 hours later…
2:31 AM
@texdr.aft \verb and other verbatim type environments can't be the arguments of another macro because of how they mess with catcodes. See texfaq.org/FAQ-verbwithin for some discussion.
@AlanMunn oh right. For some reason I didn't register that \parbox is a macro and not a primitive or whatever. Thanks.
2 hours later…
4:05 AM
i think it looks pretty good
using parboxes simplifies it a lot (although the code looks ridiculous)
i feel like the typewriter font is too dark and wide, and i don't really like the lambda though
1 hour later…
5:38 AM
can i get some assistance with this answer? tex.stackexchange.com/a/500522/165801
@texdr.aft i dunno what i'm looking at, but it's typset nicely and looks good
5:56 AM
@tjt263 thanks. what do you need with the answer?
@tjt263 and it's formatted Common Lisp code
2 hours later…
8:18 AM
the computer modern lambda has its stress inverted from how they usually look
vs computer modern
this variant is more consistent with the uppercase form
but i feel like the "normal" one may blend in better with roman text when pretty-printing code in programming languages that have the concept of a lambda function
textgreek's lambda works pretty well
^^^ newpxmath
^^^ newtxmath
are there unslanted variants?
@texdr.aft don't know, just wanted to show that other fonts have the same stress as computer modern
@Skillmon it's interesting. I wonder if it's to be more consistent with the uppercase form or for some other reason
8:34 AM
@texdr.aft for overall conformity. The stroke direction if you'd draw those characters decides how fat the stroke gets, just take a look at other Greek letters of the same font.
@Skillmon that's true. in almost every other letter the longest stroke is heaviest.
Chi is a similar case
but computer modern
it doesn't really matter at all, of course. I just noticed it today
9:08 AM
@JosephWright @UlrikeFischer I think more or less all that is needed is
-      \def\noexpand\Gin@base{"####1"}}}%
+      \def\noexpand\Gin@base{####1}}}%
in graphics then make sure that any external use of the file by the .def files quotes the whole name, just testing now....
9:21 AM
@DavidCarlisle I can run tests in the afternoon, if needed.
@UlrikeFischer seems to work I pushed that diff to required/graphics, which means I can probably remove the unquote calls from the def files since we are not adding quotes
@DavidCarlisle ooh
@texdr.aft it's not about stroke length, only about stroke direction. In CM the strokes from upper left to lower right are thick, the ones from upper right to lower left are thin.
9:39 AM
@Skillmon I see. That makes sense.
@texdr.aft ah. i don't know lisp
@texdr.aft and to answer your question, i'm just not sure how to incorporate it into my existing code.
@DavidCarlisle Er ...
9:56 AM
user image
@AlanMunn yay
@PauloCereda yaaaay
@JosephWright works for me?
@AlanMunn ooohhooo
@DavidCarlisle Not for Travis-CI ..
@JosephWright let me look... is that picking up the develop branch def files...
10:07 AM
@DavidCarlisle No, Travis-CI can't, unless we actively add them or they are on CTAN
@JosephWright hmm, so the " were added around ginn@base which meant 'hello world.eps' test worked but that broke epstopdf and is the wrong place really the quoting ought to be just at the filesystem end in the def files.... I could put a copy of dvips.def in the testfiles/support in the develop branch test?
@JosephWright or I suppose we make sure the .def files work with either format and update them on ctan now....
@DavidCarlisle Yes, sounds good
@DavidCarlisle Also a plan
@JosephWright which? copy to support or push to ctan (which means removing the unquote command or adding a stub)
@DavidCarlisle Probably copy to support at this stage
@DavidCarlisle yes (ooh)
10:16 AM
@JosephWright OK
@texdr.aft you should probably use a monospaced font though, like courier. for code i think i \usepackage{minted}
@DavidCarlisle even the comma is at the correct spot.
@JosephWright oh also should I be using a texlive installed l3build rather than one from gh?
@DavidCarlisle ooh
@DavidCarlisle TL one: I've made one minor change just for zip creation
10:27 AM
@tjt263 do you not think the typeset version looks nicer?
10:38 AM
@JosephWright OK, (but tests fail here anyway, looking....)
(\end occurred inside a group at level 4)

### semi simple group (level 4) entered at line 44 (\begingroup)
### semi simple group (level 3) entered at line 44 (\begingroup)
### semi simple group (level 2) entered at line 39 (\begingroup)
### semi simple group (level 1) entered at line 39 (\begingroup)
### bottom level
(\end occurred when \ifx on line 44 was incomplete)
(\end occurred when \ifx on line 44 was incomplete)
(\end occurred when \ifx on line 39 was incomplete)
(\end occurred when \ifx on line 39 was incomplete)
@JosephWright oops
11:00 AM
@texdr.aft no, not really. I mean sure it looks nice generally speaking. I like cursive and I like calligraphy too, although I do also like the fixed-width typewriter-style aesthetic. but it doesn't immediately look like code (which is probably fine) and there's a practical/functional reason why it's usually done that way. It's not really a big deal, and you're probably just experimenting, plus i don't wana tell you what to do, but if its for other people to read, i'd just go with tradition.
@tjt263 well... here's the source of TeX, as Knuth intended for it to be read (skip to page 138 to see some interesting parts):
@tjt263 note that (unless I misunderstand) it's more psuedo-code than code, in that the lambda will not be in greek script in actual code (unless common lisp is more unicode aware than other lisps that I know better)
i like the curly lowercase lambda too and don't like the ⋋ looking one; on another note
@DavidCarlisle right; the same goes for the logical and/or/not and multiplication symbols
(but that's beside the point)
11:13 AM
@tjt263 I'll make a better example of what I'm ultimately going for
@texdr.aft i think i understand completely. i just wouldn't do it that way. i have a few books where it's done that way and in practice it's just not all that conducive
@tjt263 that's fair
i like how some is bold and some is italic etc
that's useful
in the same way that colours and highlighters are useful
i find it more elegant than a monospaced listing (even with syntax highlighting)
and it is
way more elegant
11:21 AM
and for the record you're not alone
a lot of people really strongly disagree with the idea of literate programming etc.
what's literate programming? kanji is arguably more elegant than english and latin letter forms, it's not necessarily very easy to read though. i don't think. actually i don't read kanji, so i don't actually know, but you get it
essentially you write your code organized as a collection of what Knuth called "modules" according to the logical order of human understanding and then software makes it machine-readable by re-ordering it. So for instance you might outline the whole program at the beginning and then progressively "fill in" parts of it. Each module also has text explaining it, and another program formats the code and explanations
is it a real thing
in TeX, for example, as various data structures are introduced, global variables and types are progressively defined, instead of having to do it all at once
it is
it's how TeX and metafont were written
i can't tell if it sounds ingenious or stupid
i think it's a different topic altogether though
the pretty-printing element is relevant but yeah
but in a way it also kind of sounds like ordinary programming
this is the program that makes a program-readable file:
it's really nice to look at
i mean it sounds kind of how i write a program
11:31 AM
this method also documents what you're writing completely
well it forces you to
that sounds kind of not good. the forcing. not the documenting
but i dont know really
@texdr.aft so anyway are you good at tikz and things
not really
i wrote my first postscript thing the other day. btw
how is postscript as a language?
it's logical. and i spoke to a guy who used too work for HP back when postscript was a big deal and a new thing
and asked him about metafont
and it sounds like metafont should be good for your music notation thing we talked about
11:37 AM
yeah definitely
he said it's good for drawing basic lines and giving them profiles
@DavidCarlisle I just pulled latex2e and graphics-def (develop) and l3build check utf8-filenames-002 passed.
i was trying to modify computer modern's lambda to invert the stress
but it is very very difficult because of how tightly-constructed it is
11:55 AM
it's confusing
@UlrikeFischer it does?
@UlrikeFischer on travis it doesn't use the new graphics-def unless they are copied to the graphics testfiles support area, but I'm getting issues with the rolled back code not working (actually the tlg file I suspect capture some non-working rollback as "foo bar".eps used to work but doesn't in the rollback version for me
@texdr.aft why not just use a different font?
@DavidCarlisle yeah that's what I'm doing. I just wanted to see if I could (I can't)
@DavidCarlisle later ^^^^ ;-)
@UlrikeFischer duck pate?
12:11 PM
Q: How to draw Venn diagrams (especially: complements) in LaTeX

daroczigWhat I am up to is to write some exercises dealing with logical formulas for my students, like: And the students should draw these formulas on Venn diagrams. At the end of the lesson, I really would like to print the correct answer for them. I found a great resource on a forum thread at latex...

12:24 PM
@DavidCarlisle oh no
12:35 PM
@CarLaTeX I've seen that one
12:57 PM
@tjt263 If you don't manage to apply it to your case, please ask a question on the main site
1:09 PM
what's this site for
1:30 PM
@DavidCarlisle when I copy dvips.def it fails, but only because of different log messages. You could sent me the log of l3build check -eetex utf8-filenames-002, then I could compare. Is it normal/intented that the test loads dvips.def also with luatex?
yes there are different log I fixed that but `\openin\@inputcheck"#1" %` for example in luatex.def can't work unless the `"` have been removed from `#1` which as far as I can see they will not be in the rollback cases, I'm wondering about defining \def\quote@name#1{"\quote@@name#1\@gobble""}
in the rollback version of graphicx so you can replace `"#1"` by `\quote@name{#1} ` in the def files, might need to leave it until later today though....
@DavidCarlisle a lot work for spaces and a lot work for a rollback ;-(.
@DavidCarlisle if one of these has duck hidden in it: L‘Affine au Champagne Rosé, Chèvre Crottin en Piment d‘Esplette, Trou de Cru avec Marc de Bourgogne, Tomme Morgé refined with hops and beer, Morbier „paysan“ du Jura, Le Vigneron d‘Alsace refined with Gewürztraminer, Gouda, five years old, „Belper Knolle“ - No way to translate that. Something really special from Switzerland.
2:02 PM
@UlrikeFischer Within an ini file, LuaTeX will be in DVI mode
@JosephWright yes, I was only surprised that it then uses dvips as driver as you can't really use the dvi with dvips, but thinking about it, none of the other drivers makes sense too.
2:47 PM
@tjt263 TeX.SE, it is a questions and answers site
@UlrikeFischer You're doubting @DavidCarlisle 's linguistic abilities? I'm sure he can translate it.
@AlanMunn ooh
@CarLaTeX i know it. i'm asking about this site. why can't i ask a question here
3:02 PM
@tjt263 you mean, the chat room?
@UlrikeFischer What is the date format specified in the PDF 2.0 standard?
@MarcelKrüger Date values used in a PDF shall conform to a standard date format, which closely follows that of the international standard ASN.1 (Abstract Syntax Notation One), defined in ISO/IEC 8824-1. A date shall be a text string containing no whitespace, of the form: (D:YYYYMMDDHHmmSSOHH'mm).
@MarcelKrüger do you need more?
@MarcelKrüger there is a note at the end: PDF versions up to and including 1.7 defined a date string to include a terminating apostrophe. PDF processors are recommended to accept date strings that still follow that convention.
@MarcelKrüger only the D:YYYY is required.
@UlrikeFischer Thanks. Nice to hear that they added a note about the change, I just discovered that the trailing apostrophe is specified in the Adobe 1.7 reference but not in the ISO version, so I got a bit confused.
4:03 PM
@tjt263 Sometimes it's ok, but it's often not the best place for it. See tex.meta.stackexchange.com/q/7289/2693 for some discussion.
@PauloCereda yea
4:36 PM
@AlanMunn duck. definitely translates as duck.
@DavidCarlisle /QUACK
@DavidCarlisle What's the plan on LaTeX2e? I think the only issuue is the rotex docs ..
@JosephWright just got back in...
@DavidCarlisle Cool
@DavidCarlisle Had fun?
@JosephWright well yes and no, some of the other ,def files are not right (they just don't get well tested) so if people try the advertised feature of spacy filenames they will be disapointed, so I hope to fix up the quoting in the def files for "normal" formats then just do whatever is needed to make the tests pass. Should be done this afternoon
@JosephWright no. I hate shopping:-)
@DavidCarlisle Where? Oxford?
4:41 PM
@DavidCarlisle amazon.co.uk
@JosephWright banbury
@DavidCarlisle Lego?
@DavidCarlisle Ah
@PauloCereda yourself excepted of course, children grow up and stop playing with toys:-)
@DavidCarlisle bah :)
4:47 PM
I can't stand Arial anymore :( Why is everyone requesting that stupid ugly font?
@Skillmon Huh?
@Skillmon Comes standard with MSWord, and likely various funding agencies require it.
@JosephWright it's just that every bit of text I have to produce for the university in the next year will be in Arial.
@JosephWright And I had to complain somewhere to vent my spleen (did I use that phrase correctly?)
@AlanMunn DFG (German reseach community) requires it for requests
5:06 PM
@Skillmon Ah, right: has to be sanserif for accessiblity?
@Skillmon :)
5:23 PM
@JosephWright not sure why. I thought serif is easier to the eye for longer texts. Never understood why everyone wants sanserif full-text.
@Skillmon I have various funders who insist on sans; the argument is that for dyslexics and others with impairments, serifed text is not as accessible
@JosephWright in that case: Comic Sans, known to be best for dyslexic people.
@Skillmon I don't make the rules ...
@Skillmon To be fair, most people seem to completely ignore all of the rules about submissions
@AlanMunn Sounds about right
@JosephWright why are the rules always made by people not sharing my personal preferences? :)
5:39 PM
can i get a hand with this? i want tikz nodes to clip like paths tex.stackexchange.com/q/500747/165801
does this have anything useful?
Q: Clipping a path using a node

Joseph CooperI've searched, but I can't find a previous question sufficiently close to this one to find an answer. At the same time, this problem seems like it must have been solved by someone before now, so apologies if it's a duplicate. I want to clip a path (particularly a circle, but a general solution w...

useful to someone probably, not so much for me
i don't know why i can't just clip it like a path. i mean, the nodes have paths drawn
6:36 PM
Hi folks. I was looking (again) at:
Q: How can I add "page # of ##" on my document?

LipisI want to add a footer to my document displaying the current number out of total number of pages. How do I do that?

I was wondering if anyone can recommend what the current best approach is, when using KOMA classes. I've pretty much switched to them in recent years.
With 12 answers, the options are a bit confusing.


@JosephWright ignore me, I had save_size = 1000 which means you can't re-read the unicode files....
6:52 PM
@DavidCarlisle Erm, disaster
@JosephWright all my rollback tests were giving ! TeX capacity exceeded, sorry [save size=1000].
@DavidCarlisle Ah, that would be ... bad
@JosephWright I was experimenting with how far you could push save stack for a question the other day....
@DavidCarlisle So what's the plan for the graphics stuff? We force a kernel build then do a CTAN release of epstopdf later?
@DavidCarlisle I'm intrigued
@JosephWright I don't think epstopdf will need much change
A: “Tex capacity exceeded, sorry [save size=8000]” on Texpad

David CarlisleAlmost always if you use up the save stack you have some construct making local and global assignments to the same thing, which uses up an arbitrary amount of stack, so increasing TeX's stack size (if that is possible on the ipad implementation) probably won't help and you will need to correct th...

@JosephWright I'm working on these in the rollback case, they used to work, and they work with the new code but currently rolling back graphics but using new def files kills them (\includegraphics{"foo bar".png} should be fixable though (the old fix quoted earlier, but broke epstopdf...
6:57 PM
@DavidCarlisle Oh, right, yes
@DavidCarlisle I'm going to check I can get as far as graphics: if I can, probably I only need to rebuild that zip file later
pdftex rollback fixed, so others should follow now I think...
@DavidCarlisle OOoh
@DavidCarlisle So perhaps fixed today ...
@DavidCarlisle Reminds me of the time I tried to do \expandafter\gdef\csname ...\endcsname inside a group for all the Unicode chars ...
7:22 PM
@JosephWright hmm not as fixed as I thought I wonder how bad it would be to not offer rollback it's adding a lot of complication for little gain, unless we also want to roll back the def files
@DavidCarlisle I don't think we can do that: they need to load in plain
@JosephWright well.... they only work in plain as it loads miniltx which was half of latex as it was in 93, so that could be updated for at least the space handling, but yes, I need to check they still work in plain b@@gr (who's idea was it to offer plain tex support for graphics...)
@DavidCarlisle well let's say it this way: I never used a rollback and never saw someone using it. Also as we are uploading a dev-format, whatever rollback is perhaps needed could also be added later if someone complains ...
@JosephWright I'll drop out a bit realtime update not working:-) if you want realtime updates follow apolloinrealtime.org/11
@UlrikeFischer yes but as Joseph says it relates to plain tex support as well as we don't want to inflict the utf8 support on plain so graphics needs to work without that at least
7:40 PM
@DavidCarlisle yes, but is this related to rollback?
@UlrikeFischer related yes as the rolled back format is more or less what the miniltx code for plain sets up, but it's actually different in other ways too so that's three different cases:(
8:09 PM
@JosephWright Object Oriented oh?
@JosephWright or Open Office oh
@DavidCarlisle the idea to run and use l3image instead gets very attractive ;-)
@UlrikeFischer Yup
8:24 PM
I used to have a FF script that implemented a "copy the code" link in the upper right corner of code blocks on the site. It seems to have disappeared. I thought it was a Greasemonkey script, but I guess it wasn't. Does anyone use it and have a link? I have no idea where I got it.
8:38 PM
Talk about nostalgy: booting up my Lion system...
@PauloCereda WOW
@JosephWright everything works, surprisingly. :)
@Skillmon I know you're not working at Overleaf anymore, but do you know why files created with filecontents don't show up in the project area (even though they are created properly)?
@AlanMunn They are dynamic? Like the log?
@AlanMunn Overleaf is mean oh no
8:50 PM
@AlanMunn overleaf copies imho everything to a temporary folder, it doesn't compile in the project folder.
@UlrikeFischer Ok, that makes sense. It must be doing something like that. It's quite frustrating to deal with Overleaf users sometimes... :(
@DavidCarlisle Just remembered: the test failing on Travis-CI is one I have to disable locally anyway (the Windows command_line_utf8 business), so I can in principal build for release
@AlanMunn Ulrike's right. But I'm not that familiar with Overleaf's exact inner workings, Most of my work was debugging LaTeX projects. I know that Overleaf doesn't completely clear the virtual machine that compiles your project, it does have a cache (and most likely your file is stored inside that cache). But I don't know enough about that to be of real help here, best would be to contact the support directly.
@AlanMunn You can find them if you click on "Other logs & files"
@AlanMunn Click "Logs and output files" and then "Other logs & files" ^^^
9:09 PM
@CarLaTeX I see. Yes, the clearing works somewhat differently from the way we're used to on a local installation. So stuff can appear to persist longer than you would expect unless you explicitly clear the cache.
@AlanMunn If you use \usepackage{filecontents} you can rewrite the csv without deleting the previous version
Just curious. How doable is the approach described in
A: Reading the contents of a box

Charles StewartLuatex has what looks to me like nice support for this, having primitives for traversing over the contents of boxes (which it represents internally using the more general concept of typed nodes). If you have a smattering of Lua and read the first three chapters and section 4.10 of the Luatex m...

? The author doesn't provide any actual code.
@CarLaTeX Yes, I know how it works; I was just commenting on the fact that the file it creates doesn't show up in your project folder.
@AlanMunn No, but you can download and upload it if you want it there
@AlanMunn but it is not updated, unfortunalety
10:11 PM
@FaheemMitha that is what tagpdf is doing to inject literals: it traverse the shipout box. lua-visual-debug is doing the same. ctan.org/pkg/lua-visual-debug.
2 hours later…
11:43 PM
if anybody here happens to be familiar with common lisp i have a question
it seems that the order of (let ...) (labels ...) (macrolet ...) etc. in a function doesn't matter, but is there a "standard" order?
i get that if functions defined in (labels ...) refer to variables defined in (let ...) then the let has to come first, but what happens in the case that everything's independent

« first day (3184 days earlier)      last day (28 days later) »