@DavidCarlisle Not sure I follow: in DVI mode doesn't the usual dvips.def get used?

@PauloCereda You up so early again? :) (I mean: Good morning, pal!)

@yo' yep. :) The morning duck gets the early donut. :)

@yo' Good morning, pal! :)

@JosephWright it gets used, it just doesn't work:-) EPS files can have binary stuff and as a fallback it tries the eps bounding box reader on unknown extensions like .png. In either case, in classic tex you can step past arbitrary bytes but luatex dies on malformed utf-8 sequences. You could push into lua and use a different file reader cf luainputenc but if you are going in to lua you may as well read the image using the lua image library and so it ought to be completely different....

@JosephWright that accounts for several of the cases why .luatex.tlg end in a fatal error and memory stats which fail every time anything changes rather than cleanly reaching \END so that the tlg log is pruned.

@DavidCarlisle Ah right: I was thinking for 'reasonable' input (actual EPS files without a preview or at least with the bbox before that)

@DavidCarlisle Makes sense: add some 'If LuaTeX then give up' switches?

@JosephWright currently I'm doing that just in the tests yes (not in dvips.def) as as you say it works in many reasonable cases. Of course if l3build had a reasonable way to say use this test with etex, pdftex and xetex, use this test with luatex and xetex , ... then I could configure the tests there rather than having fake non-result test results:-)

@DavidCarlisle Javier's asked for that too: I'm going to work on it over the next week or so (once I finish a few siunitx v3 things: really want to get that sorted)

@DavidCarlisle I think I know how to make it workable, it's just a question of implementation

@JosephWright I've had some package improvements in that state for 30 years:-)

@DavidCarlisle :)

@JosephWright something else which occurred to me just now is "fatal error and memory stats which fail every time anything changes rather than cleanly reaching \END so that the tlg log is pruned" the memory stats are never particularly reproducible across versions l3build could treat Here is how much of *TeX's memory you used: as an implicit END-TEST-LOG and prune the tlg at that point.

@DavidCarlisle Good plan: I'll do that

Good news: soon arara will speak Chinese!

@PauloCereda 不是每個人？

@DavidCarlisle not everyone? :)

@PauloCereda "doesn't everyone?" but near enough:-)

@DavidCarlisle :D

@DavidCarlisle Coming up shortly

@DavidCarlisle Did you release for the luatex.def issue?

@JosephWright not yet, plan to do that this morning (was just sorting things so build ctan actually made the zip file rather than moaning that png files were not utf-8 (I can't remember I may have checked in the tests since the last release)

@JosephWright: I am preparing a document to talk with the kids later on (cathecism). Did you know pope Francis is a chemist?

He even has a master.

@PauloCereda so that's why he's so good!

@yo' :)

@DavidCarlisle I have an l3build update sorted that normalised out all of the rubbish: I'll send that to CTAN now

@DavidCarlisle If you do check in updated test files, they'll fail until that hits release TL

@JosephWright thanks just saw the message from travis (presumably commit message comes at the next cron)

@JosephWright hmm what does this mean (in l3build clone)

$ git pull
remote: Counting objects: 151, done.
remote: Compressing objects: 100% (13/13), done.
remote: Total 151 (delta 100), reused 105 (delta 96), pack-reused 42
Receiving objects: 100% (151/151), 56.91 KiB | 2.11 MiB/s, done.
Resolving deltas: 100% (100/100), completed with 6 local objects.
From github.com/latex3/l3build
 * [new branch]      manifest   -> origin/manifest
   46a6a0d..725c0b1  master     -> origin/master
 * [new tag]         2017-11-25 -> 2017-11-25
 * [new tag]         2017-09-12 -> 2017-09-12

@DavidCarlisle I deleted a few dead branches

@JosephWright so how do I clear locally (git pull always gives the "your config..." message now)

@DavidCarlisle git branch -D config

@JosephWright delete the same branches here?

@DavidCarlisle Yup

@JosephWright ah thanks

oh it was a branch named config? I thought it meant my git configuration was saying to do do something..

hmm

$ git branch -r
  origin/HEAD -> origin/master
  origin/config
  origin/manifest
  origin/master
  origin/pdftests

$ git branch -D config
error: Cannot delete branch 'config' checked out at '/home/l3build'

$ git branch -D origin/config
error: branch 'origin/config' not found.

$ git pull
Your configuration specifies to merge with the ref 'refs/heads/config'
from the remote, but no such ref was fetched.

@JosephWright ah I'm in config, I guess that's the problem, I'll get back to master...

@JosephWright that fixed it..

@DavidCarlisle Cutting off the branch you're sitting on is said to be a bad idea.

@HaraldHanche-Olsen I blame @JosephWright I don't recall switching to that branch, perhaps I was just having a look round at some point:-)

@DavidCarlisle Hi Sir,

@Saravanan hi

You are a real TeXnician

@Saravanan but I'm not @TeXnician:-)

@DavidCarlisle :)

Yes i know that, when you come India surly call me my mobile number 8098149169

@JosephWright something missing in my mental model.... build ctan built a zip with no errors so I did git tag -a -m "Release 2017-11-22" 2017-11-22 and that shows locally if I list the tags but git push says there is nothing to do, and (as far as I can see) the tag isn't on github also how do I get things into that releases list on github, I can see how to do it from the github web interface but can it be done via a magic commit tag at this end?

@DavidCarlisle git push --tags

@DavidCarlisle GitHub releases can be done in the web interface but tags are easier (certainly for us, as real releases go to CTAN)

@JosephWright thanks I was navigating the doc to there, then once it's there I associate that tag to the release via the github gui?

Now let's see if Petra likes my zip file...

@DavidCarlisle Once you send the tag it will appear on GitHub as a release: no further work needed (we don't do complex stuff like release notes)

@JosephWright oh good so done then I think

@DavidCarlisle Yup looks good

@DavidCarlisle Once the l3build update propagates I'll probably update those test files to remove the engine test ...

@JosephWright yes I left them as they were for now (I think the update arrived to save me doing one more test from my original list of failures)

gone to ctan

@DavidCarlisle Well I'd rather wait for l3build to be updated on CTAN so Travis-CI will be able to 'cope'

@JosephWright are we running travis on the graphics-def repo?

@DavidCarlisle Hopefully within a week or so I'll sort the multiple configs too

@DavidCarlisle Oh, we are not: I guess I can just update now then (if you are happy)

@JosephWright on my extensive testing of one test run, it worked for me:-)

@DavidCarlisle you are my hero

@PauloCereda A boss of mine usually said: "The actual test is production"

@CarLaTeX Sounds like Windows update policy...

@TeXnician LOL

VM proving useful: still need to sort out beamer so it can be built cross-platform, so for the present the Linux VM is essential

Random question... What's with bidi, polyglossia, and \if@nonlatin? I sense there is some tension between the authors, but it does seem to break some things.

do we like bread in this room, or only ducks?

@AGoldMan Is polenta good?

@DavidPurton The author of bidi claims primogeniture rights, so everybody should do like he wants.

@egreg it can always be improved with salt and vinegar and a splash of tomato ketchup.

@DavidCarlisle Wrong, sorry: beans or cheese.

@egreg The American version of polenta is called grits (a tradition in the South), and is very good. A classic version is shrimp and grits. While I doubt people would put ketchup on grits, they certainly would put hot sauce (chili sauce).

@egreg it sounds good. They just don't really sell corn products where i live

Very cool. The British Film Institute has put up thousands of archived films online, many of them free to watch. Here's a link for the cricket fans: player.bfi.org.uk/free/collection/cricket-on-film

@AlanMunn It is a pity one has to be in UK to watch.

And for those without the discerning taste for cricket, here's the link for football. player.bfi.org.uk/free/collection/football-on-film

@mickep Just noticed that. :(

@mickep I guess it's time for a UK VPN connection.

@AlanMunn Oh, that is perhaps a good idea. The dangerous thing with these kind of archives is that they tend to steal time from you :)

@mickep True.

@AlanMunn Ooh

@JosephWright You'll have to give those of us without access some reports. :)

@AlanMunn I'm currently in Switzerland, but luckily work has a VPN :)

@JosephWright Conference?

@AlanMunn Beam time: psi.ch

@JosephWright Wow. That looks cool. What are you doing with it?

@AlanMunn How technical can I be? :)

@AlanMunn We are using GPD: psi.ch/smus/gpd

@JosephWright Well straight formulas probably won't mean anything, but I probably can remember my basic physics and chemistry.

@JosephWright Also, as I tell my students, we should all be able to explain our research to someone we meet at a bar. :)

@AlanMunn OK. We are using 'muons', which are unstable particles. From a chemistry point-of-view, they act like a hydrogen atom. We are using that to see where hydrogen reacts with some metal-containing compounds: they are interesting for hydrogen production.

@JosephWright So the muons allow you to do experiments that you couldn't actually do with hydrogen itself?

@AlanMunn Yes: as the muons are unstable we can detect them (or at least what they decay into), and using a magnetic field can work out where they were in the molecules. You can't do that with hydrogen itself.

@JosephWright I'm fairly certain 'muons' have to do with cows, not ducks

@JosephWright Cool. It must require a great deal of preplanning of experiments to use your time effectively.

are you sure you have the right room?

@AlanMunn The technique is mainly used for physics (magnetism, etc.): our experiments are slow, so we are in some ways better off, but have to make best use of the beam time

@AGoldMan Yes, it's quarks that have to do with ducks. It's all very elementary. :)

@JosephWright Would you have any use for this (when it's finished?) frib.msu.edu/about/index.html

@JosephWright "Proton therapy" made me think that the protons needed therapy, but then I understood that it was something much more important!

@AlanMunn Perhaps: would have to think a bit

@mickep :)

@mickep The medical aspect is quite new: it's odd having a facility largely for research with the occasional patient/family about

@JosephWright Interesting. The ones fighting cancer are the big heroes in my opinion.

@JosephWright Well if it does turn out to be useful let me know. It still has a few years I think to be operational. My neighbour is one of the many physicists working on the project.

@JosephWright That sounded a bit weird. The neighbour fact is incidental. :) But if you find yourself coming here to use in a few years it let me know.

@AlanMunn interesting thesis package you could use it at your institution

I blame @yo' :-)

@DavidCarlisle I'm thankful that most of our theses (at least the ones by people who know any LaTeX) are written in English. :)

@DavidCarlisle close but not me

@yo' I blame you for not walking down the road and sorting them out:-)