@UlrikeFischer re Burmese, do you (or @MarcelKrüger) understand Hans/Luigi's experiments with a harfbuzz interface, and is that available in context yet (or ever)
@PauloCereda I'll fill you in on what happened. People talked about ducks, @marmot talked about marmots, @UlrikeFischer broke luatex, @egreg stole some ticks.
@UlrikeFischer that's odd grffile doesn't have special access it is just doing normal tex tests with \openin, how come it's not following normal TEXINPUTS search rules?
@UlrikeFischer ah there was once, in some previous life, a discussion about what should happen and what various tex systems did in that case, but I can't remember when or where or what the outcome was.
@UlrikeFischer yes I saw that at the time, I just wondered if any version had crept in to context since, I guess not
@DavidCarlisle Regarding the font.each issue Hans proposed a "solution" which would only list fonts loaded by the fontloader instead of all fonts. Would this be enough for your usecase?
@UlrikeFischer What did you do to the tail of my fellow marmot? (BTW, I think the figure does precisely what it should: it tells people "This is really complicated stuff, consider using the tikzducks package instead." ;-)
@PauloCereda Hello! How R U? I hope that your long absence has a nice explanation such as a long vacation in the Andes (and not that you started working for Microsoft ;-)
@MarcelKrüger The original issue was that in luatex even for tfm fonts that if the same font was loaded twice under different names \font\zzza=cmr10 \font\zzzzb=cmr10 then in \showbox output (ie the main tool of the latex test suite:-) luatex showed \zzza and pdftex/xetex showed \zzzzb (or the other war round) so until this luatex update I had some code in test2e.tex that just checked for duplicate fonts and switched around the ids, to be honest I forget the details (Hans helped a bit)
@MarcelKrüger without that test if I let the font code run in my 1.09 (experimental branch) luatex then luatex segfaults and the latex test suite never completes. but with it as it is now then for luatex 1.08 and later we get dozens of extra difference files as luatex reports different font names in box output whenever latex has made a font substitution
Interesting update to pdf(la)tex in texlive this morning.......
$ type pdftex
pdftex is hashed (/usr/local/texlive/2018/bin/x86_64-cygwin/pdftex)
davidc@dc-bantham /c/tmp
$ pdftex
davidc@dc-bantham /c/tmp
$
@JosephWright ^^^ the test suite runs very quickly with this setting (luatex, texlua, xetex all do the same...)
@PauloCereda “Mojave requires support for the Metal graphics API, which means dumping most Macs made before 2012 (with the exception of 2010 and 2012 Mac Pros that can be upgraded with newer Metal-compatible GPUs).” – Ars Technica
@marmot No, the ; is not required, the package ends all the paths with a ;. In fact it can even cause a complain in the .log file about missing null font characters (as helpfully discovered by the great tugboat editors!)
@samcarter Maybe one can write some insert path styles such that \fill[yellow,duck]; works, where duck/.style={insert path={...}} .... then there won't be complaints about excess ; ;-)
@boycott.se-yo' @DavidCarlisle could still be right, he did not tell you the units. He probably meant units of 100 secs, which makes his answer correct.
@DavidCarlisle why did you add this comment? It doesn't add any new information except that analogue relation of center <-> \centering and flushleft <-> \raggedright.
@Skillmon well that's odd I didn't see your answer which seemed to have been posted some hours before:-) Either it was an old tab or I didn't scroll down past zarko's deleted answer or something. I have no idea now:-)
Hello, yesterday a former teammate of my high school added me on Facebook, and today I have 2 more requests (they did not talk to each other). What's going on?!?