« first day (4410 days earlier)      last day (520 days later) » 

8:18 AM
@DavidCarlisle It seems there is some move from svn to git going on, maybe that is why the patch is not yet added.
 
@UlrichDiez you could also create yet another auxiliary that contains the name<->page correspondences. (roughly the same as tagging the pages with hyperref, but instead write to an external file)
 
 
4 hours later…
12:22 PM
Interesting LaTeX error commentary on my compile:
> Things are pretty mixed up, but I think the worst is over.
2
 
12:33 PM
@mickep yes I saw that thread, I was using the main texlive svn (actually its git mirror) Luigi said he would push all changes there while the upstream luatex repo was being re-organised (that's where the original adding of the variablefam got added the other day
@DavidPurton never trust compiler messages, assume it's going to get worse?
 
I dunno. Looking at what came previously, I'm pretty sure I'd already hit the worst
 
@DavidPurton That is a funny message indeed.
@DavidCarlisle The one who waits...
 
@mickep It's a warning to anyone foolish enough to use the expl3 \peek_ functions without really understanding what is going on…
 
12:49 PM
@DavidCarlisle Rumor says one can add
     math_eqno_gap_step_par = 1000 ;
     math_flatten_mode_par = 1; /* ord */
     var_fam_par = -1,
at the end of maincontrol.c
 
@mickep let's see.....
 
Or to add
\ifdefined\variablefam \variablefam-1 \fi
at some relevant place.
 
@mickep yes if the engine wasn't going to change I assume we'd add that to luatex and lualatex .ini files so done in the format but an engine change seems safer
 
@DavidCarlisle so, now you can test. 😃
 
.... X -DLUA_USE_DLOPEN -Wimplicit -Wreturn-type -g -O2 -MT luatexdir/tex/libluatex_a-maincontrol.o -MD -MP -MF luatexdir/tex/.deps/libluatex_a-maincontrol.Tpo -c -o luatexdir/tex/libluatex_a-maincontrol.o `test -f 'luatexdir/tex/maincontrol.c' || echo '../../../texk/web2c/'`luatexdir/tex/maincontrol.c
...
 
1:02 PM
@DavidCarlisle ooh
 
1:26 PM
@mickep original test as posted here works, I'll run the 2e tests... (takes a while on this machine)
 
@DavidCarlisle why does this command do? And why gives it so curious errors if its value is 0?
 
1:44 PM
@UlrikeFischer I made casual remark to @mickp and by magic a feature got added and brok miktex:-)
 
@DavidCarlisle But what does it do (or is meant to do if you hadn't broken it)?
 
@UlrikeFischer normally in a mathchardef you specify mathclass and family but f you want \mathbf to work you need to specify mathclass=7 which works like class=0 for spacing. now if mathfamvar is set (eg to 99) and you set the fam to 99 then you can specify any mathclass and the current \fam font is used. (so you could have bold \mathpunct in \mathbf)
but the variable is initialised in the engine to 0 not -1 so (a) this feature is on by default and (b) you really don't want everything in fam 0 to do this. so the fix is to initialise to an illegal \fam
 
@DavidCarlisle came if from the question how to change the digits?
 
@UlrikeFischer a question here the other day about something, yes
Oct 25 at 20:08, by David Carlisle
@mickep in lmtx math can you have a variable alphabet (like class 7) without giving up the mathclass for spacing. Could you have a variable font \mathpunct "?" ?
 
@DavidCarlisle it sounds useful. But you probably then need open type fonts for the math alphabets?
 
1:57 PM
@UlrikeFischer I think it applies to tfm as well (which is why it breaks more or less anything that uses fam0 in the classic setup, such as digits)
 
@DavidCarlisle yes, but it wouldn't work for all symbols if \mathbf is an OT1 font wouldn't it?
 
@mickep all tests passed (latex2e repo, main branch)
@UlrikeFischer well the font switching feature would work, you'd get missing glyphs for anything not in OT1. Original example was \mathbf{abc?} not giving bold ?
 
@DavidCarlisle yes I meant the missing glyphs. But as it is luatex one would have open type fonts anyway. In other news it looks as if Marcel found out why base mode eats text ...
 
2:14 PM
@DavidCarlisle oh, great, some dodos know their luatex coding...
 
@UlrikeFischer That was a funny one. Two engine bugs, both barely noticeable on their own, causing chaos when combined...
 
@Skillmon Some sort of expandable mapping from name of original image-file to pdf-page-index via auxiliary file is indeed feasible and the external file containing the mapping-data can be created while creating the pdf-file that contains all images. ;-) But it was just an idea, as they come into my head when I busy myself with something to distract myself from pain ;-)
I think at the stage of fiddling with the document assembly I'll just leave it at loading the graphicx package with the draft option. ;-)
 
2:31 PM
@MarcelKrüger ;-)
 
the aircraft soars again
4
 
2:44 PM
@DavidCarlisle I was quite confident that the picture mode expert would kick in ;-) (Did you like the longfbox package?)
 
@UlrikeFischer I only looked at longfbox as far as the first line of the doc where it said the command used was like \fbox and single line. Someone once told me it was good to read package documentation, I thought I'd try for once.
 
@DavidCarlisle I continued to the line Fast and portable: only uses the standard LATEX picture environment for drawing Fast and portable: only uses the standard LATEX picture environment for drawing
 
3:12 PM
@UlrichDiez sorry to hear that reason! I hope it's getting better!
 
 
1 hour later…
4:19 PM
@Skillmon Thank you. As long as I can hang around here and exchange ideas with courteous people like you and learn interesting things, things are bearable. ;-) I don't know how you and all the other people who hang around here are doing - in any case I wish you well.
 
 
1 hour later…
5:42 PM
Is there a way to align LaTeX in both sides automatically? I got an error Overfull \hbox (9.3887pt too wide) so do I need to fix it manually?
 
@JaakkoSeppälä In the standard classes justified text is the default ... What is on the line causing the overfull box?
 
I just tried an example:

\setlength{\parindent}{1cm}

Suomessa suomessa Suomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suome
 
@JaakkoSeppälä use real text --- and be sure to set the language correctly, otherwise LaTeX can't find breaking points and the alignment will (greatly) suffer.
 
Re language, the following works better:
\documentclass[draft]{article}
\usepackage[finnish]{babel}

\begin{document}

\setlength{\parindent}{1cm}

Suomessa suomessa Suomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suomessaSuomessa suom
 
6:01 PM
@JaakkoSeppälä line breaking depends on the length of the line (which you have not shown), even when you set the correct language it may not be possible to find acceptable breaks in narrow lines such as table columns or \parbox unless you give tex some help eg \sloppy which allows inter-word spaces to stretch more
 
@Rmano The proper algorithm should work also for fake text.
 
@JaakkoSeppälä it works, for the TeX definition of "working" ;-) as @DavidCarlisle said, LaTeX will normally refuse to have very ugly spaced lines, although you can force it.
 
@DavidCarlisle I tried to write similar in HTML and CSS with text-align: justify; and it gave me much nicer layout.
 
@JaakkoSeppälä As said, it depends on the font, the line length, and a lot of things. Why don't you post a proper question with the comparison you mention here? With two real examples, reproducible by everybody, the answers could be much more objective.
 
@JaakkoSeppälä you probably did something wrong in the tex (which in general hs much more sophisticated line breaking than a browser) but with no example, hard to say anything
@JaakkoSeppälä no: tex has language sensitive hyphenation and by the default US English rules xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx has no breakpoints so if the line is not that wide it will overflow the margin, that is by design.
 
6:14 PM
@DavidCarlisle Oh. I see.
 
6:28 PM
@PauloCereda time for some Zelda...
 
7:04 PM
@DavidCarlisle Meanwhile, another \variablefam question.
 
@mickep blame @UlrikeFischer
 
@DavidCarlisle looks so, I think Christian misunderstood me and pumped to the next development version instead back to an older one.
 
@DavidCarlisle As you always suggest.
 
@mickep would you rather I blame you?
 
@DavidCarlisle Well, I feel no responsibility for what is shipped in miktex, so sorry... :P
 
7:14 PM
@mickep blame and reality are not necessarily related
 
@DavidCarlisle Well well, better blame someone the miktex users know of.
 

« first day (4410 days earlier)      last day (520 days later) »