« first day (3754 days earlier)      last day (1231 days later) » 

1:06 AM
@Rmano and @samcarter_says_quack -- Thanks for the pointers. I tried a google question, and got a few pseudo-hits, but when they were all dated before 2010, I didn't think that was going to be useful.
@DavidCarlisle -- And you know that my fingers can't function correctly in anything but emacs.
 
 
7 hours later…
7:51 AM
@Skillmon editor is part 2. :)
 
8:06 AM
@JosephWright expanscape.com
 
8:25 AM
@UlrikeFischer @DavidCarlisle interesting hyperref issue, I can confirm it from my update yesterday. Seems a named anchor is not added unless cleverref is loaded: tex.stackexchange.com/q/582598/3929
 
@daleif I know that, that is a problem in amsthm (github.com/latex3/hyperref/issues/48). The interesting question is what cleveref patches to get it working.
 
@UlrikeFischer ooh cleveref seems to live up to its name, being clever. :)
 
8:48 AM
@UlrikeFischer Done
 
@JosephWright it better isn't it?
 
@UlrikeFischer that's strange then that I haven't encountered that problem
 
9:14 AM
@daleif well it only happens with lists at the begin.
 
@UlrikeFischer Ahh, I often use some cheats in that case, which might have prevented the issue
 
@UlrikeFischer you don't think ctan will be missing you?
 
@DavidCarlisle oh
 
@UlrikeFischer we should probably fix amsthm (haven't looked how)
@PauloCereda ask @UlrikeFischer about accents
 
@DavidCarlisle those meddling L3 kids are up to something!
@DavidCarlisle óôḧ
 
9:30 AM
@DavidCarlisle well cleverref seems to insert or move a \refstepcounter somewhere. But I didn't check if this breaks something else.
@DavidCarlisle perhaps it the next round we could also add this here: github.com/latex3/hyperref/issues/167 ?
 
@DavidCarlisle, @UlrikeFischer, @PauloCereda Now we have parts, I can put in a PR to move l3color - do we want to just do l3pdf too?
 
@JosephWright I think it would be nice. What happens then with all the files doing \RequirePackage{l3pdf} currently?
 
@UlrikeFischer yes but can probably be simplified a bit, I suspect you want to treat an explicit space of 0pt like any other space so can drop \ifdim\lastskip=\z@ that was just a pre-etex version of \ifnum\lastnodetype=11 so I think ypu just need that latter test and if tis previous node is glue save and restore it, otherwise don't, perhaps...
 
@UlrikeFischer We have two options: we either say 'tough' (as it's experimental and hic sunt leones) or we bring back l3obsolete.dtx to DocStrip a one-line package for it and l3color (for l3cctab, I took the first route)
 
@JosephWright I don't have a problem with "tough".
 
9:43 AM
@UlrikeFischer I'll get the PR sorted then we can talk about this :)
@UlrikeFischer Makes sense to me: the old l3obsolete was about the combination of everything in l3kernel into one file, not explicitly experimental modules
 
@DavidCarlisle I don't want to do anything here ;-) It is your code.
 
@JosephWright you probably meant to ping @PhelypeOleinik? :)
 
@PauloCereda he had to ping you as @Fubá didn't work.
6
 
@DavidCarlisle ooh
> A font no one asked for, with no discernible use case, Refraktury (inspired by Volker Busse's F25 Blackletter Typewriter font) is a monospaced black letter font, suitable for typesetting math the hard way, illegible code listings, and oddly kerned newspaper marquees. It's a simplified mix of several different styles, mostly Fraktur and Schwabacher. It covers the ISO Latin-1 block, because why in the world would you care if it covers anything else?!
 
@JosephWright but I would then need something to trigger the backend loading. Currently I'm doing \RequirePackage{l3pdf} which gives me \c_sys_backend_str and with it I can load my extra backend code.
 
9:59 AM
@PauloCereda :)
 
@DavidCarlisle as ctan asked to note, if backref or nameref has changed, I was wondering if one should use my sword and move them somewhere else or use the same version number everywhere.
 
@UlrikeFischer Well \RequirePackage{expl3} will then work
 
@PauloCereda No, the ping was correct: @JosephWright wants you (or @Fubá) to proofread all of source3 :-)
 
@PhelypeOleinik oh no
@PhelypeOleinik I then need to read source1 and source2 to catch up!
 
@PauloCereda source1 is The TeX book, I guess :)
3
 
10:08 AM
@PhelypeOleinik ooh
@PhelypeOleinik /Appendix G intensifies
Meanwhile, people create transparent wood:
> Scientists Develop Transparent Wood That Is Stronger, Lighter Than Glass
@JosephWright ^^ opacity and transparency are now trendy, I see
 
@PauloCereda :)
@PauloCereda Appendix D all the way ;)
 
@JosephWright ooh :)
 
@DavidCarlisle, @UlrikeFischer, @PhelypeOleinik I'm also thinking about dropping \__kernel_msg_...: it doesn't seem to help us very much :)
 
@JosephWright It's been causing some trouble :)
 
@PhelypeOleinik Indeed, plus I think it's not clear what 'counts' as being kernel level
 
10:16 AM
@JosephWright But I do find “Package hooks error” slightly weird
 
10:35 AM
@JosephWright what would be then the error message, say from a tl command?
 
10:46 AM
@UlrikeFischer That could stay the same: all that's needed is the messages have a name that starts LaTeX/: currently that's hidden inside \__kernel_msg_...
 
11:03 AM
@PhelypeOleinik Isn't the TeX book more of a "interface1" document? "source1" sounds more like "TeX: The Program".
 
@PhelypeOleinik \prop_gput:Nnn \g_msg_module_type_prop { hooks } { } would give you just hooks error
 
@JosephWright That seems a bit odd too IMO. Could we get something like hooks module error?
 
@MarcelKrüger Well then I'd add \prop_gput:Nnn \g_msg_module_type_prop { hooks } { Module } to get Module hooks error
 
@JosephWright I would prefer that one. It keeps the general structure while also making clear that we don't have a separate package.
 
@MarcelKrüger Just add the line, then: we have the mechanism there for a reason (it is documented)
 
11:15 AM
@JosephWright Hm... That's much better, yes. But I think from the user viewpoint that's all LaTeX, regardless of some internal separation of sources. The idea with the message starting LaTeX/ seems good
 
@PhelypeOleinik Well yes, that's what happens with the ones currently as kernel: they are internally LaTeX/<module>/<msg> - all I'm suggesting is to make that explicit by dropping __kernel_...
 
@JosephWright Seems reasonable to me
@MarcelKrüger Ah, indeed! We should make @PauloCereda read both though :)
 
@PhelypeOleinik oh no
 
@UlrikeFischer backref was such a hack and nameref ought to have been replaced by one of the other *ref packages by now but I suppose moving them out would be reasonable could put them in tools? possibly simpler to use same number everywhere?
First Test, Chennai (day five)
England 578 (Root 218) & 178 (Root 40, Ashwin 6-61)
India 337 (Pant 91, Bess 4-76) & 192 (Kohli 72, Anderson 3-17, Leach 4-76)
England win by 227 runs
 
@DavidCarlisle :)
 
11:26 AM
@JosephWright There was a time you could rely on Psmith to post these things
 
@DavidCarlisle :)
 
@DavidCarlisle I'm not sure if tools would be a good idea - imho some place where fast updates are easier is better. Perhaps changes will be needed when we start to improve the references. And I'm not sure if the same number helps ctan, both have their own catalog entry so they would have to update everything all the time.
 
@UlrikeFischer could move them back to ho-tex
 
@DavidCarlisle ah right, it was there with hyperref. That could be a plan.
 
@UlrikeFischer dpctex is another possibility (I have a suspicion I wrote them:-)
@UlrikeFischer or of course just split them but keep them under latex3 org at github (maybe that is best) like iftex
 
11:32 AM
@DavidCarlisle backref has your name, nameref only Rahtz and Oberdiek.
 
@UlrikeFischer yes but I think Sebastian and I may have exchanged a few emails first (all a long time ago:-)
 
@DavidCarlisle ooh email was already invented
 
@PauloCereda JANET probably
 
@DavidCarlisle I think I like the ho-tex or dpctex idea best for now.
 
@UlrikeFischer hotex then
 
11:35 AM
@DavidCarlisle ooh UK
 
@DavidCarlisle ok.
 
@DavidCarlisle ooh
 
11:55 AM
@JosephWright I tried to install the new branch and it told me:
Running l3build with target "install" for module l3color
Error: Cannot find configuration build.lua
 
@UlrikeFischer That's odd: it's not coded in anywhere other than in l3draw, which I removed
 
@JosephWright let me try again with a clean build folder.
@JosephWright I found the reason, I had a pdf in the l3color folder, and so it didn't disappear when I switched the branches.
 
@UlrikeFischer Ah, that would do it, yes
 
12:14 PM
@JosephWright it works ...
 
@UlrikeFischer hmm hyperref includes backref and nameref (given suitable options) and I'd rather move hyperref towards core and reduce dependencies on ho-tex rather than the other way round. If we split I think keeping it under latex3 might be best
 
@UlrikeFischer Phew
@DavidCarlisle Sounds good to me
 
@DavidCarlisle well I think that most of nameref should be replaced in due time by kernel methods based on l3ref so it looks to me like some tempory move. Perhaps ask on the team list if we want them under latex3?
 
@UlrikeFischer yes as I said above both packages could go in the end, given that splitting just annoys people with minimal installations and hand managed dependencies, leaving them in hyperref is another option....
 
12:30 PM
@DavidCarlisle yes, ok. lets leave it for now, and I will try to remember to tell ctan that they haven't changed.
 
1:13 PM
quack
 
1:58 PM
@PauloCereda d.....
 
@DavidCarlisle ooh duck
 
@JosephWright so Module is hooking error into what? ;-) (/hide for the very bad joke)
 
2:18 PM
@Rmano "nuisances" :)
 
 
1 hour later…
3:30 PM
Quote by a colleague: "git problems are no problems, they are part of deal of using git" :)
 
@samcarter_says_quack :)
 
4:08 PM
@JosephWright - Use new (internal) interface for kerns I'm not amused. Now I'm still not allowed to use kerns in my L3 code!
 
@Skillmon Who needs keming anyway?
3
 
@Skillmon Raise on LaTeX-L :)
@Skillmon I keep asking, the rest of the team keep telling me 'later'
 
@JosephWright I don't really blame you :) I just think it's funny that there is a release which is seemingly only about using an L3 \kern interface, but doesn't provide a user interface.
 
@JosephWright @PhelypeOleinik look what I found in lipsum.sty and guess what happens: \cs_generate_variant:Nn \__xparse_split_argument:nnn {nne}
 
@UlrikeFischer Yay, I broke lipsum! :D
@UlrikeFischer Why does it need that though?
 
4:20 PM
@PhelypeOleinik no idea, but I will make a bug report and complain.
 
@UlrikeFischer why does that break? It should work, shouldn't it? (Or is \__xparse_split_argument:nnn no longer defined in dev?)
 
@UlrikeFischer Hm, \__xparse_split_argument:nnn is the same as \SplitArgument, so it can replace
\cs_generate_variant:Nn \__xparse_split_argument:nnn {nne}
\cs_new_eq:NN
  \__lipsum_split_argument_expanded:nne
  \__xparse_split_argument:nne
by
\cs_new_eq:NN \__lipsum_split_argument_expanded:nnn \SplitArgument
\cs_generate_variant:Nn \__lipsum_split_argument_expanded:nnn {nne}
(probably; didn't test :)
 
@UlrikeFischer generic bug report text: Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor
 
@Skillmon xparse is now ltcmd, so __xparse is __cmd
 
@PhelypeOleinik ah, I didn't know that you also changed the naming (but knew that it was moved)
 
4:28 PM
@Skillmon Yeah, I went for maximum breakage :)
 
@PhelypeOleinik question remains why lipsum does that in the first place.
 
@Skillmon ¯\_(ツ)_/¯
 
4:45 PM
Is there a way to get the functionality of the plain TeX tabbing macros in HTML? (Or a better place to ask that question, since it would probably not be on-topic on this site, right?)
 
@texdr.aft don't all such things end up being tables?
 
@texdr.aft If you're asking about Mathjax, then you're right, it's not in topic. But tabbing in TeX is used so rarely (if at all?). I guess I would use a tabular environment instead. (But I don't use Mathjax)
 
@AlanMunn plain tex he said
 
I suppose it could be done by just simulating the tabbing algorithms with Javascript
@DavidCarlisle The use case is for formatting programming languages nicely, using variable-width typefaces
 
@DavidCarlisle Even so, does anyone actually use tabbing?
 
4:54 PM
@AlanMunn in latex no(not many) but no plain tex users use tabular
 
About two years ago, I was whining on here about typesetting Common Lisp, and I finally realized that the tabbing macros were the way to go
Just to be clear, I'm talking about \+ and \cleartabs
 
@DavidCarlisle Yes, I know I thought people would use \halign not tabbing. But I'll stop talking now. :)
 
@texdr.aft but they are just a table really aren't they?
 
@DavidCarlisle With continually-varying tab stops
 
@AlanMunn Were you offended that Don Hosek called you an amateur linguist?
 
4:59 PM
As in this example from the TeXbook but when the code is not known beforehand:
 
@texdr.aft minor detail (or simply a new table)
 
@DavidCarlisle No, because I didn't know he had. :) But really? Where?
 
@AlanMunn universal qualifier strikes again
@DavidCarlisle Given it's needed for Polish, I assume that he's looking for ę (lower case e with ogonek), I thought all TeX-folks were amateur linguists. — Don Hosek 12 hours ago
 
@DavidCarlisle Maybe I'll just make a question about this. I'm certainly open to easier ways to do it
 
5:00 PM
@DavidCarlisle lol
 
@texdr.aft do you just want varying indent or possible mid-line alignments?
 
@DavidCarlisle mid-line alignments
 
5:30 PM
@texdr.aft possibly
<!DOCTYPE html>
<html>
 <head>
  <style>
   table {display:inline-table;vertical-align:baseline}
   .indent {margin-left:1em;}
    td{vertical-align:baseline}

  </style>
  </head>
  <body>


   <div>
    <table><tbody>
    <tr><td>if <i>n</i>&lt;<i>r</i> <table><tbody><tr><td><b>then</b> <i>n</i>:=<i>n</i>+1</td></tr>
    <tr><td><b>else</b> <table><tbody><tr><td><b>begin</b> <i>print_totals</i>; <i>n</i>:=0;</td></tr>
         <tr><td><b>end</b>;</td></tr></tbody></table></td></tr></tbody></table></td></tr>
 
@DavidCarlisle Very nice
 
@UlrikeFischer Other than xparse itself, mathcommand and spectralsequences use __xparse macros, but I'm not sure how much we need to worry about those
 
@PhelypeOleinik Well we know the rules
 
@JosephWright GLOB
 
5:40 PM
@PauloCereda Oh behave
 
@JosephWright :)
 
@JosephWright Yeah, ideally those misuses wouldn't exist, but... :)
 
@PhelypeOleinik I mean more that we have been 100% clear about the rules, and these were never documented even as internal, so really ...
 
@JosephWright Exactly. I was thinking of a case like lipsum, that everyone uses, so it's better if they get a chance to update before breaking. But the other two I've never seen any usage, so I won't lose hair over them
 
5:43 PM
@PhelypeOleinik well this spectralsequences is quite horrible. But I guess we should at least sent a short note. I will do it.
 
@PauloCereda I heard some birds outside my window. Could it be an Arara?!?
 
@PhelypeOleinik oh no
/quacks in despair
 
@UlrikeFischer It is! mathcommands I think uses only \__xparse_check_definable:nNT, which is quite easy just to copy over
 
@UlrikeFischer Yes, just reminded myself
@PhelypeOleinik No, it messes with the parsing to fiddle with space behaviour
 
@JosephWright Oh :(
 
5:49 PM
@PhelypeOleinik Look over the whole directory it's in!
@UlrikeFischer I suspect you read it
 
@JosephWright what did I read?
 
@JosephWright There's only mathcommand there...
 
@DavidCarlisle The plain TeX source for the table is
\vbox{\+\bf if $n<r$ \cleartabs&\bf then $n:=n+1$\cr
\+&\bf else &{\bf begin} ${\it print\_totals}$; $n:=0$;\cr
\+&&{\bf end};\cr
\+\bf while $p>0$ do\cr
\+\quad\cleartabs&{\bf begin} $q:={\it link}(p)$;${\it free\_node}(p)$; $p:=q$;\cr
\+&{\bf end};\cr}
 
@texdr.aft yes that's what I started from (but dropped the bold on the first if by the look of it)
 
@DavidCarlisle So every \+ becomes <tr>, every & that sets a new tab stop adds a new <table>, everything else becomes a <td>, the \quad adds a margin, and \cleartabs just ascends the table structure
except when the \cleartabs is unnecessary, as it is in the first row
For some reason I didn't consider starting a table inside a table cell, which actually makes perfect sense given the concept of "mid-line alignments" (just create an alignment in the middle of the line)
Thank you
 
6:12 PM
@DavidCarlisle the hyperref build creates hyperref-doc.html and manual.html ;-(
 
6:25 PM
@UlrikeFischer -- I'm late on the scene, so this may be addressed already. In amsthm, a theorem environment expects some text at the beginning of the environment. The example offered has none; instead it starts with \begin{enumerate}. If either text or \leavemode is inserted before the \begin{enumerate}, all labels result in linkable references.
 
@DavidCarlisle I found the reason for the additional manual: there were ignored files in the doc folder.
 
@UlrikeFischer -- To elaborate on the preceding comment, the "problem" is fully explained in section 2.1 in the user guide (texdoc amsthdoc). In other words. the cause of the issue is "user error".
 
@barbarabeeton well I know that adding text or a \leavevmode helps and that the problem is known. But I wouldn't call it an user error, if the code can't correctly handle this, more a deficiency of the code. And the fact that it works better if cleveref is loaded shows that some fix is possible.
 
@DavidCarlisle -- You could probably insert something like $ $\unskip at the beginning of all theorem environments, but I don't think that's a very good idea. I believe that @egreg answered a question in this forum to the effect that it's "better style" to start such a theorem with a brief text, a position with which I agree.
 
6:49 PM
@UlrikeFischer ah yes that caught me out when I first renamed
 
@DavidCarlisle I also broke package options by delaying the puenc. So I think tomorrow is a new update date. Can you make a suggestion for the space code?
 
@UlrikeFischer let me look...
@UlrikeFischer Oh I see what you mean I answered the tex.sx question. When you said it was my code I thought you meant there was a comment saying I did the original
 
7:05 PM
@DavidCarlisle @samcarter_says_quack twitter.com/NintendoAmerica/status/1359215530499067908
 
@UlrikeFischer actually I think that code is OK, I'll add
 
@DavidCarlisle ok, thanks.
 
I wrote it on the internet so it must be right
 
@DavidCarlisle Hi, I think the puenc-extra.def file is missing from build.lua.
@DavidCarlisle Affects the creation of ctan, it is not listed in the latest update :(
 
@PabloGonzálezL yes, that should be corrected too.
 
7:28 PM
@UlrikeFischer you or me... (just in tdslocations I guess)
 
@DavidCarlisle I can do it. Did the xetex driver really never set the pdfversion?
 
@UlrikeFischer ?
 
@DavidCarlisle I was just testing if 2.0 can be set everywhere to close github.com/latex3/hyperref/issues/87, and realized that pdfversion doesn't work in xetex at all.
 
@UlrikeFischer neither hxetex not hdvipdfm ?
 
@DavidCarlisle doesn't look so, I see no code. I will add this to both drivers. That should work:
\special{pdf:majorversion \Hy@pdf@majorversion}
\special{pdf:minorversion \Hy@pdf@minorversion}
 
7:37 PM
@DavidCarlisle Question, more or less how much is the level (in TeXLive versions) of hyperref backward compatibility?
 
@PabloGonzálezL by what measure? I don't understand the question.
 
@DavidCarlisle Sorry, I misspoke, let me explain, at the end of hluatex there is a check for the KOMA class, is it necessary today?
@DavidCarlisle If I read correctly the other day @UlrikeFischer said he was cleaning up the code a bit.
 
@PabloGonzálezL I'm a she ..
 
@UlrikeFischer A thousand apologies, I always confuse your name :)
@UlrikeFischer Anyway, I blame the translator, in Spanish it is neutral "Ulrike Fischer dijo" and it was well written.
@DavidCarlisle One more fore you github.com/latex3/hyperref/blob/…
Add The :)
 
7:57 PM
@PabloGonzálezL I have no idea, but leaving it in is safer than taking it out
 
8:19 PM
@DavidCarlisle we lost a few \penalty 10000 in some tests, so I think your code does something
 
8:30 PM
@UlrikeFischer ah sorry I should have updated the results
 
@DavidCarlisle no problem, I had to test anyway.
 
 
2 hours later…
10:43 PM
Hello everyone. A brand new version of mandi, tentatively called mandi3 (which I hope is permissible even though it is not written in LaTeX3), is in testing with my students and will soon be released. Thanks to the help of everyone here, it is far better than it would have otherwise been. Since July you have taught me so much about fonts and font management, Unicode, key-value interfaces, expansion programming, LaTeX2e's internal macros, LaTeX3, modern TeX engines, and so much more. Thank you.
 
@LaTeXereXeTaL do you have to change the name? packages like revtex that end up with revtex, revtex4, revtex4-1, revtex4-2 are such a pain.
 
@DavidCarlisle Oh sheer joy
 
@PhelypeOleinik sorry I forgot to blame you at the end of that sentence.
 
@DavidCarlisle I replied just to remind you of that <3
 
@PhelypeOleinik you see I tried to save the world from that pain by not completing revtex4
 
10:52 PM
@DavidCarlisle Yet someone thought they knew better than The Oracle you
 
11:14 PM
@PhelypeOleinik ^^
 
11:25 PM
@DavidCarlisle I keep irritating people despite trying not to. LOL I want to somehow keep the current version available and separate from the new version, and the only way I could think of to do that was slightly change the name. I have absolutely no intent to changing it thereafter. I really did think hard about it.
 
@PabloGonzálezL You are mean!
 
@LaTeXereXeTaL sometimes it's unavoidable but if every package kept all its versions live, ctan (and everyone's tex installation) would be considerably bigger.
 
@LaTeXereXeTaL I don't think you irritated anyone :-) But it's a maintenance nightmare to rename packages across versions (I know, last week I made 4 different versions of xparse :)
 
@LaTeXereXeTaL one less intrusive way of keeping both versions is to rename the old one as say mandi-v2.sty and then set up so \usepackage{mandi}[=v2] uses that. (see how array.sty works for example. That way you keep a legacy snapshot available but the "good" name goes to the new version and you don't have to distribute documentation etc for the old
 
@DavidCarlisle You definitely have superpowers, I was trying to write the same thing :)
@PhelypeOleinik Entretenido con xparse :D
 
11:33 PM
@PabloGonzálezL Oh, no, that is over thankfully!
Ç_Ç
 
@DavidCarlisle I'll see if I can get that working.
 
@PhelypeOleinik There is always some leftover :) ... reminds me that you have to look at scontents at some point.
 
@DavidCarlisle With all due respect, that's getting out of one mess and into another :)
 
@PabloGonzálezL Ah, yes, it won't find what it expects in xparse.sty anymore...
 
11:39 PM
@PhelypeOleinik Yes, some things will have to be removed (or added), but there is no rush yet.
 
@PabloGonzálezL yes well personally I could recommend the longtable versioning strategy, don't update for 25 years
 
@DavidCarlisle This is a solid development model :D
 
@DavidCarlisle hm
 
@UlrikeFischer you rang?
 
@DavidCarlisle anniversary update?
 
11:43 PM
user image
2
With a nicely misplaced comma, just to annoy @DavidCarlisle ^^^
 
@UlrikeFischer not sure when the last substantive cange was actually, the date got bumped a couple of times when Karl spell checked the distribution
user image
3
@PhelypeOleinik I am at peace and not annoyed at all
 
@DavidCarlisle You fixed it! <3
 
@DavidCarlisle Question, do you have any quick line (grep/seed) to detect the ` hyphenation` added by CTAN when copying and pasting, my editor doesn't show them :(
 

« first day (3754 days earlier)      last day (1231 days later) »