@egreg could be powered by NAG!

@DavidCarlisle ^^

@barbarabeeton: got your poking, will reply soon. I am a hurry, even here in Portugal!

I want to use the London underground just to hear watch your step and mind the gap.

@PauloCereda You Europeans should relax and learn to live the... hm... south american style.

@mickep ooh Brazilians

wait a minute...

syntax error

@PauloCereda ohh, perhaps too much chart junk.

@mickep ooh

Guys, is the iTeX logo in TL?

I might need it for... things...

@PauloCereda it's not that exciting. What do other countries do? Just let the commuters fall through and get squashed?

@DavidCarlisle Yes, that's what happens. :) Only the English seem to care about the commuters. :)

@DavidCarlisle: I will try to not lose @JosephWright when in a metro in Rio.

@DavidCarlisle I think I'll start again with the timer/file primitives: I need to sort the ordering of xetex.web to be similar-ish to pdftex.web before I add them ...

@DavidCarlisle So I will have to do the 'tidy up', at least to some extent

@DavidCarlisle Unless you can work out what the current errors mean :)

Anyone else speak WEB?

@JosephWright where is your pgf git version? I wanted to look if I can find the reason for the move of the calc-code (tex.stackexchange.com/a/430406/2388).

@UlrikeFischer sourceforge.net/u/josephwright/pgf/ci/master/tree

@UlrikeFischer Parent at sourceforge.net/p/pgf/git/ci/master/tree

@JosephWright I found the commit (7338a) and it sounds as if the code wasn't moved for some specific reason. It was done while implementing a "fast lane". Probably Christian didn't saw that it would make a difference if \widthof etc are set later.

sigh Time to install TL2018 and see which of my packages broke.

@kan Sorry for bothering you with this here. (This was the only room that I was able to find quickly and where I was able to ping you.) If/when you have time, could you check whether you still have scan of the paper by Ljunggren? The one you mentioned in your answer on MO: chat.stackexchange.com/transcript/12070/2018/5/7

@MartinSleziak Is it hard to find? It is not impossible that we have it in our basement here in Lund (Sweden).

Maybe I should have said that it is hard to find online?

In any case, in this answer (by kan) link to a scan is given - but it no longer works.

\o @MartinSleziak

Are you drowning?

@MartinSleziak Check here. I hope the quality is good enough...

@mickep The requested URL was not found on this server. If you entered the URL manually please check your spelling and try again.

Try again, I wrote the wrong password while scp:ing...

Thanks!

Do you mind if I update the post on MO (which has dead link now) with this link @mickep?

So that it can help others searching for the paper?

@MartinSleziak I have rather limited disk space on that account. Could it be arranged somewhere else, perhaps?

@mickep Ok, so if I upload it elsewhere and add that as a link, that would be ok with you?

@MartinSleziak Perfectly fine, and I just assume no one else will care about copyright issues in this case.

Well, if somebody does than probably I am the one in trouble if I upload it somewhere.

Although I am not sure whether there can be copyright issues with a paper from 1943.

@MartinSleziak Hehe. Well, if you have downloaded it I will remove it from the server.

Thanks for taking your time to do this!

Yes, I have downloaded it.

@MartinSleziak No problem, I had another paper to copy from the basement :)

@egreg On e-TeX, I'm at the advantage of being in the middle of reading all of the sources :)

@egreg How would you typeset the completion of the of the local ring at x? Both \hat{ \mathcal{O}_{X,x} } and \widehat{ \mathcal{O}_{X,x} } look very odd.

@JohnDorian \hat{\mathcal{O}}_{X,x}

@egreg Ok, in this case it is unambigous^^ But what about the completion of the local ring A_{ \mathfrak{m} }. This should not be confused with completing the ring A and then localizing at \mathfrak{m} .

@JohnDorian \mathfrak{m} is not an ideal in \hat{A}

@egreg But \hat{A} is an A-algebra.

@JohnDorian So long as your notation is consistent, I see no issue. If you feel there is a problem, say that B is the completion of A_\mathfrak{m}

@JohnDorian Don’t be overconfident on notation

@egreg What do you mean with this?

@egreg Wait... Do I see sans headings here: profs.scienze.univr.it/~gregorio/categorie.pdf

@JohnDorian I like to experiment

@JohnDorian Also this one: profs.scienze.univr.it/~gregorio/fondamenti-primaria.pdf

@JohnDorian All notation can be ambiguous.

@DavidCarlisle @JosephWright @egreg When the luatex documentation says (regarding the \hyphenation command) "An = indicates a (non-discretionary) hyphen in the document input." what does this means in your opinion for the break points of \hyphenation{mul-ti=boun-da-ry}? (I'm trying to decide if luatex 1.04 or 1.08 is wrong here ...).

@UlrikeFischer It should mean that with \hyphenation{multi-word=bound-aries} the word multiword-boundaries can be split after multi, at the explicit hyphen and after bound.

@UlrikeFischer But it doesn't. :-O

@egreg then what is the difference to \hyphenation{mul-ti{-}{}{-}boun-da-ry} which is described in the documentation as a mean "to specify extra hyphenation points in compound words"?

@egreg it doesn't in luatex 1.07. There it is broken. Works again in 1.08 (but there the {}{}{} syntax is broken). This hyphenation business makes me feel like Sisyphus.

@UlrikeFischer As far as I can see, = has never matched the description. Or we're not understanding what that description means.

@egreg accurate description of the situation (and of quite a number of luatex related questions).

@egreg, @UlrikeFischer It's fun, isn't it :)

@JosephWright @UlrikeFischer Maybe they should implement LuaTeX in Scheme, just like MlBibTeX.

@egreg :)

@JosephWright Meanwhile, I see l3experimental in the TeX Live updates. :-)

@egreg Same here

@JosephWright LaTeX2e <2018-04-01> patch level 4

@JosephWright When did the change affecting \g__prg_map_int come into force? I've just installed TL2018 and found no issues (ref: tex.stackexchange.com/q/429859/86).

@LoopSpace The first set of post-freeze changes (went a few days ago)

@JosephWright Oh hang on. I have a local version of l3 installed. I need to update that ... sorry!

@JosephWright Ack. Remind me what I need to do to my clone of the git repository to put the right files in the right places. Do I need to build anything, or are the style files all in there somewhere?

@JosephWright Hmm, there's a date mismatch between l3int.dtx in l3kernel and expl3-code.tex in build/local. Former says 2018-04-30 latter says 2018-02-21.

@JosephWright Got it ... needed to use texlua rather than lua to build it

Okay, on to the more serious questions. I keep getting into trouble for using "internals" from L3. But much of what I'm doing is writing similar "low-level" stuff to the core stuff and so it makes sense to copy what the core does. For example (in this case) I defined an iterator over an object. So it made sense to look at things like \tl_map_inline:XYZ and essentially copy the code.

Should I instead copy-and-rename-everything? But then I end up essentially duplicating a lot of code rather than just a little bit.

@LoopSpace 'Yes', more or less

@LoopSpace We've been tidying up the kernel itself to avoid 'cheating' between modules: in the end, what we want is not to rely on internals from 'other places'

Any BibDesk users around? Do you know if it's possible to associate an autofile folder on a per .bib file basis? It seems to be a global preference.

@JosephWright Except that in this case, the purpose of \g__kernal_prg_map_int appears to be that it keeps track of nested loops, so I ought to use it in my loops.

@DavidCarlisle That sounds slightly messy. Does it still work correctly? (Sorry, just saw this message.)

For programs like autotools, it's a really good idea to keep everything as high level as possible. Because generated code can quickly become really hard to understand. That's especially true of autotools.

Enormous quantities of shell script, most of it redundant.

Checking for things and conditions that are out of date and don't matter.

@FaheemMitha got it working. as I suspected Karl confirmed that texlive builds on some platforms without automake/configure so he checks in a "generic" Makefile.in that he generates, so actually I just needed the .am file

@DavidCarlisle Ok. Actually, I like autotools. Though it gets mixed press at best.

People say good things about Cmake, but I didn't particularly care for it.

So TeX Live runs a configure script when building?

@FaheemMitha yep (several:-)

Ok.

@FaheemMitha @DavidCarlisle It seems to be very common to include Makefile.in in source distributions. After all, it is supposed to be totally architecture independent, so it doesn't hurt to include it. And also, it is good for users that they don't need to build it themselves, and they don't even need to install autotools.

@HaraldHanche-Olsen Machine generated files should not be put under version control. It's ok to include files like Makefile.in in source tarballs, though. In fact that is quite common practice.

The alternative is to have the user invoke some esoteric autotools commands.

@FaheemMitha Ah, I wasn't thinking of version control so much. Good point.

@HaraldHanche-Olsen But of course people do things one shouldn't do all the time. With computers, and otherwise.

does anyone understand why I get "height" at the end of my table?

oh, apparently because of \hline

if I use \hrulefill the problem is solved, but I don't get a neat horizontal line at the end of my table

@ShaVuklia put \\ before \hline. And also, check booktabs and stop using \hline and | altogether :-)

[ 8/21, 01:15/07:10] update: latex [13383k] (47500 -> 47612) ... TLUtils::check_file: removing C:\Users\carla\AppData\Local\Temp\42Skq3ow6c\mZjVjIs6Xl\latex.doc.tar.xz, checksums differ:
TLUtils::check_file:   TL=c24ec065fef0f23c264877927be6b7cfd5396e2cb7fcf03ecee8740d966e553049e8271d7b10605ff783adbc106e04e0ae267faa1d6f4d47a77805a3354dbdb0, arg=3dbf6c9d93cdfd02b27e41b299dfe1904b8a011fad460198bd0d5e14673ebf4e3ff52bf20dc108d71226733979548c0917ca0475f50115177e9bc27dacb5569d
TLPDB::_install_package: downloading did not succeed

My TeX Live 2018 Manager has been stopped there for 10 minutes, what should I do? ^^^

@CarLaTeX Switch to Word? ;-)

@marmot LOL Only if forced to do it

@CarLaTeX Try another mirror?

@AlanMunn Yes, but can I close the TeX Live Manager without causing damage?

@CarLaTeX I should think so, but I don't know for sure.

I also once restarted my TeXLive manager and it was OK (but of course that's not a guarantee that it is always OK)

@AlanMunn @marmot I'll try. I think I had a German mirror in TL2017 and it was always working... German efficiency!

@AlanMunn @marmot Restarted with the German mirror, but the update began with make4ht...

@CarLaTeX That's probably a good thing? I think it's pretty clever.

@AlanMunn good news, thanks!

@CarLaTeX Just hope your TeX installation is not going to put Sauerkraut on your pizza.

@marmot Oh no!

@yo' ah thanks! I'll check it out

@CarLaTeX Does your TeX work now?

@CarLaTeX You might need to do tlmgr update --all --reinstall-forcibly-removed if something has gone wrong.