TeX, LaTeX and Friends

SBM

8:47 AM

Hello

Joseph Wright

9:01 AM

@SBM Hi

SBM

Good afternoon, though I just got some bad news

egreg

@SBM Sorry to know it.

Joseph Wright

@SBM Oh dear: nothing serious I hope

karlkoeller

9:22 AM

Hi, all. I hope this could be helpful for someone here

2

A: Double backslashes disappear from code

Based on \\ corruption, I've created a new query TeX.SX \\ corruption (based on user ID) which I hope will help anyone to find their own posts affected by this bug!

egreg

@karlkoeller Surely useful! Thank you!

CarLaTeX

9:48 AM

@egreg @karlkoeller Already upvoted, but didn't they tell us they'd have automatically fixed it?

@DavidCarlisle In this case: tex.stackexchange.com/q/371718/101651 the \label within the caption is not good, showkeys overwrites the caption

David Carlisle

10:11 AM

@CarLaTeX a feature.

CarLaTeX

@DavidCarlisle LOL

David Carlisle

@CarLaTeX showkeys tries harder not to change the spacing in the original document than it does to put the string somewhere sensible, so if it overprints in some cases, that's not a bug:-)

CarLaTeX

@DavidCarlisle but in cases like that shouldn't be better to shift up a bit the key shown? I imagine perhaps I'm asking too much

egreg

@CarLaTeX No, they proposed a special review queue.

CarLaTeX

@egreg OMG I think we'll fix all by 2030!

egreg

10:19 AM

@CarLaTeX Maybe. But the idea was that edits in the queue don't bump the question to the main page.

CarLaTeX

@egreg That's good, of course!

egreg

@CarLaTeX As you can see, the main page is now full of old questions, most of them with accepted answers.

CarLaTeX

@egreg Yes, I saw @karlkoeller has worked hard this morning

David Carlisle

@CarLaTeX I might look but I wrote that a long time ago the hard part is deciding you need to do anything different, you don't really "know" you are in a caption, you just know you are in some inner box and can't easily access the page margin so anywhere that you put the string will be a guess

@CarLaTeX you don't think it's a brilliantly effective plan to ask volunteers to edit 7000 questions in their spare time, just offering a plain textbox web interface?

CarLaTeX

@DavidCarlisle However, there is a solution: put the label within the caption but at the end, not at the beginning

David Carlisle

10:31 AM

@CarLaTeX in that case, oddly enough the package is optimised for the place where I habitually put \label.

CarLaTeX

@DavidCarlisle They didn't put a big effort offering their solution...

@DavidCarlisle LOL, I'll do the same, then!

Paulo Cereda

12:01 PM

quack!

TeXnician

@PauloCereda Did I already say you're contagious? See my question yesterday.

Paulo Cereda

@TeXnician ooh upvoted

@TeXnician: ducks are good for MWE's :)

TeXnician

@PauloCereda Ask @DavidCarlisle. He would give you zzzzzzzzzzz instead (infinite z pool).

Paulo Cereda

@TeXnician :)

TeXnician

@PauloCereda And it's a pity that % is not allowed in command names, else egreg could use his favourite character in MWEs ;)

2

Paulo Cereda

12:09 PM

@TeXnician Don't give him ideas, or he will deliberately change the catcodes in order to achieve this! :)

2

TeXnician

@PauloCereda Or he would present us with a funny \csname combination that extends etex to egregtex :D

2

Paulo Cereda

@TeXnician :D

CarLaTeX

@TeXnician You should give your contribution here: tex.stackexchange.com/q/346695/101651

TeXnician

@CarLaTeX I already upvoted (nearly everything), but answering is currently out of my league. I'm no graphic designer (yet) :(

SBM

hello

CarLaTeX

12:18 PM

@TeXnician Why not? Do you think the other answerers are graphic designet? :D

SBM

Oh

so cool

TeXnician

@CarLaTeX Probably they are better than me. I could give you a duck hiding behind a bath curtain ;)

SBM

You are so good at TikZ everyone

TeXnician

@SBM It is easier to follow the chat if you use the reply button (right-most button on a chat post).

@SBM And no, by far not everyone. Only our creative ducks here...

SBM

oh

Ulrike Fischer

1:08 PM

@DavidCarlisle: I think my combo-size function works now ;-) \DeclareFontShape{TU} {combotest}{m}{n}{<->combo*combotest-regular}{}
\DeclareFontShape{TU} {combotest}{bx}{n}{<->combo*combotest-bold}{}

David Carlisle

@UlrikeFischer ooh well done (where did you see this combo thing, is it new?) you could self answer and pick up a few points:-)

SBM

:)

Paulo Cereda

@DavidCarlisle she mostly certainly reads manuals. :)

Joseph Wright

@DavidCarlisle We forgot about the shell-escape pull request at the team meeting. I'm thinking of going with 'accept and see who objects' ;)

Paulo Cereda

Ostrich algorithm

In computer science, the ostrich algorithm is a strategy of ignoring potential problems on the basis that they may be exceedingly rare. It is named for the ostrich effect which is defined as "to stick one's head in the sand and pretend there is no problem." It is used when it is more cost-effective to allow the problem to occur than to attempt its prevention. == Use with deadlocks == This approach may be used in dealing with deadlocks in concurrent programming if they are believed to be very rare and the cost of detection or prevention is high. For example, if each PC deadlocks once per 10 years...

:)

Ulrike Fischer

1:12 PM

@DavidCarlisle Not very new, sometimes last year. And it is in the luaotfload manual. I put the code in a sty along with some code for a better interface, I think I will have to upload it before I can answer. But the core trick was idea with the size-function ...

David Carlisle

@JosephWright ok by me, sent a mail just to alert people?

Joseph Wright

@DavidCarlisle I guess ...

@DavidCarlisle Long weekend is a chance to do several of these things then update CTAN

David Carlisle

@UlrikeFischer ah @PauloCereda suspected that you had seen a manual, must admit I don't think I've ever looked at that:-)

Paulo Cereda

@DavidCarlisle ooh :)

David Carlisle

@JosephWright we're out tomorrow

Paulo Cereda

1:19 PM

@DavidCarlisle, @JosephWright: I thought Cambridge was a bit different. :)

SBM

oh

I'm still at school

Paulo Cereda

@SBM me too, I am a poor duck student. :)

SBM

oh

David Carlisle

1:33 PM

Paulo Cereda

@DavidCarlisle oh not that duck. :)

David Carlisle

@PauloCereda You don't look so poor here ^

Paulo Cereda

@DavidCarlisle not related. :)

David Carlisle

@PauloCereda closer ^ ?

Paulo Cereda

@DavidCarlisle it's more accurate. :)

ooh @MartinScharrer is here! Hello!

Martin Scharrer

1:40 PM

@PauloCereda Hi

David Carlisle

@MartinScharrer have to ask, any chance of updating standalone to work with luatex?

Martin Scharrer

@DavidCarlisle I'm just working on the bug list right now.

David Carlisle

@MartinScharrer thanks, it's missed the freeze for the DVD unfortunately but if it could get into the first round of updates it would be good:-)

@egreg @JosephWright looks like hans and luigi fixed that ifincsname bug, or at least the log shows that, even though I never got a reply on the list. I suppose I should build the development version and try... `| 2017-05-17 21:52:01 +0100 (Wed, 17 May 2017) | 1 line

ifcsname fix (HH)`

Ulrike Fischer

2:02 PM

@DavidCarlisle I just tried with the w32tex luatex-dev and it prints now "AAA" like pdftex.

David Carlisle

@UlrikeFischer ah good thanks I'm running make but it's still building stuff

David Carlisle

2:19 PM

@UlrikeFischer same here now, no more zzz

SBM

oh

Martin Scharrer

@DavidCarlisle, @UlrikeFischer: About the issue with dvips and standalone. I saw your bug ticket. Thanks for reporting it. For a workaround for now use the preview option of the standalone class.

David Carlisle

@MartinScharrer dvips? must have been Ulrike's issue:-) For me the really critical one is luatex (especially as the changes are trivial just a few command name changes)

Martin Scharrer

@DavidCarlisle Ok, I mixed you up here maybe.

You mean the issue reported as bitbucket.org/martin_scharrer/standalone/issues/23/…

Right?

David Carlisle

@MartinScharrer yep

@MartinScharrer that one comes up on this site all the time.

Martin Scharrer

2:36 PM

@DavidCarlisle And the code posted by Falk Hanisch in the ticket fixes this for you as well?

David Carlisle

@MartinScharrer standalone and geometry are really now the only two high profile packages that I know that haven't been updated for luatex 0.85.

@MartinScharrer haven't tried, I could if you want:-)

David Carlisle

2:49 PM

@MartinScharrer just looked over the pull request here bitbucket.org/martin_scharrer/standalone/pull-requests/1/… and it looks more or less sensible but I don't really know the standalone codebase to know if there are other places you use \pdf... commands that are not defined in luatex. a quick grep suggests that there are not many such places.

Martin Scharrer

@DavidCarlisle I'm looking over it right now.

It is anoying that LuaTex changes its behaviour like this.

David Carlisle

@MartinScharrer oh sure it cost Joseph and me days of work to keep latex working, but it was marked as beta software so really you can't complain (although we did:-)

Joseph Wright

@DavidCarlisle Same here

@DavidCarlisle Cool

David Carlisle

@JosephWright flushed with success on that test file, I'm running build check in base...

Joseph Wright

@MartinScharrer Hans is very clear about the fact that LuaTeX is not back-compatible

Martin Scharrer

3:05 PM

@JosephWright Well, backwards compatibility is a curse I know.

egreg

@DavidCarlisle Of course not in the binary for TL 2017 :(

@PauloCereda I've been watching the Giro d'Italia: this was the last climb

google.it/maps/dir/Valstagna,+VI/Foza,+VI/…

David Carlisle

@JosephWright @MartinScharrer and speaking of compatibility, running the latex test suite with today's lautex build gives

  Check failed with difference files
  - ../build/test/tlb-extraprimitives-001.luatex.diff
  - ../build/test/tlb0056.luatex.diff
  - ../build/test/tlb0320.luatex.diff
  - ../build/test/tlb0380.luatex.diff
  - ../build/test/tlb1638.luatex.diff
  - ../build/test/tlb2727.luatex.diff
  - ../build/test/tlb3480.luatex.diff
  - ../build/test/tlcs01b.luatex.diff
  - ../build/test/vlatex01.luatex.diff
  - ../build/test/vlatex02.luatex.diff
  - ../build/test/vlatex03-2015.luatex.diff
  - ../build/test/vlatex03.luatex.diff

@JosephWright actually just looked at all the diffs, it's not so bad: all same issue change in discretionaries (again)

Joseph Wright

4:13 PM

@DavidCarlisle Oh goody

@DavidCarlisle Those were passing with TL'17 pre-release?

Alan Munn

Do we have a question that addresses 'package avoidance' (i.e., I want to do X but I don't want to use a package)? It comes up every so often and although there are sometimes reasons to avoid using a package, for end users (as opposed to package/class writers) it's usually preferred to use a package.

Joseph Wright

@AlanMunn Probably not a general question

Alan Munn

@JosephWright We do have one about whether you should use packages in classes or other packages. (With one answer saying 'yes'.).

Joseph Wright

@AlanMunn Again, probably said lots of time in comments

Alan Munn

@JosephWright Yes, that's where I'm coming from. As with many 'best practices' questions, it would be a bit opinion based, but it might be good to have something to point people to in such comments.

Joseph Wright

4:23 PM

@AlanMunn We can give some objective advice, I hope

David Carlisle

@JosephWright I think so (but I have a feeling at least a couple of them changed when we reverted back to 1.04 from 1.05 there is a new discretionary related primitive and all the diffs are discretionary related, I haven't checked in detail but as they were only discretionary lines in the diff I think it's a change in logging rather than a change in line breaking

Joseph Wright

@DavidCarlisle Yes, sounds right

David Carlisle

@JosephWright typical example, full diff from first failure is

Joseph Wright

@DavidCarlisle Fro L3 it's clearer: always release TL

David Carlisle

./build/test/tlb0056.luatex.diff
::::::::::::::
*** ../build/test/tlb0056.luatex.tlg    2017-05-27 15:35:22.622794200 +0100
--- ../build/test/tlb0056.luatex.log    2017-05-27 15:35:27.700919600 +0100
***************
*** 207,213 ****
  ....\hbox(0.0+0.0)x0.0, direction TLT
  ....\OT1/cmr/m/n/10 D
  ....\OT1/cmr/m/n/10 K
! ....\OT1/cmr/m/n/10 -
  ....\OT1/cmr/m/n/10 2
  ....\OT1/cmr/m/n/10 1
  ....\OT1/cmr/m/n/10 0
--- 207,215 ----
  ....\hbox(0.0+0.0)x0.0, direction TLT
  ....\OT1/cmr/m/n/10 D
  ....\OT1/cmr/m/n/10 K

Joseph Wright

4:26 PM

@DavidCarlisle Ah yes, seems familiar

Alan Munn

@JosephWright Yes, I think so. Maybe I'll write up a question then.

Joseph Wright

@AlanMunn Cool

@AlanMunn Gives me a chance to promote expl3 ;)

Alan Munn

@JosephWright :D

David Carlisle

@JosephWright all of them are like that or just a - for a line lost where an untaken discretionary is now not logget at all

TeXnician

@AlanMunn Sound like a very good idea.

Werner

4:37 PM

3

...somewhat interesting...

TeXnician

@Werner You would have needed 365 gold badges, sorry.

Werner

@TeXnician I guess. I'll try harder next time.

TeXnician

@Werner Then you'd have more gold badges than egreg...

David Carlisle

@Werner much more interesting than the totally uninteresting number of 568,577

TeXnician

@DavidCarlisle And that's even better than 390,758 ;)

egreg

4:52 PM

@DavidCarlisle Which is prime.

David Carlisle

@egreg and undeserved

barbara beeton

5:39 PM

@CarLaTeX -- actually, they haven't offered a solution. we suggested it, but i don't think we've heard anything back from them since.

CarLaTeX

@barbarabeeton Even worst :(

David Carlisle

5:57 PM

@egreg been following the cricket?

egreg

@DavidCarlisle No, Bologna was playing some team and lost 1:2. :)

Joseph Wright

8:37 PM

@DavidCarlisle, @egreg As always, fix one issue, find another: github.com/latex3/latex3/issues/371

egreg

@JosephWright Short blanket. ;-)

Joseph Wright

@egreg I have a feeling this makes more sense in Italian!

David Carlisle

@JosephWright I don't understand the issue description:-)

Joseph Wright

@DavidCarlisle Better title welcome: basically I've messed up and you can't escape strings properly in \lua_now:n

egreg

@JosephWright Yes, it's common for saying that when you fix something, something else will break: if you cover the shoulders with a short blanket, the feet will be uncovered.

David Carlisle

8:42 PM

@JosephWright but why can't you do the _-ised versions of \directlua{...\luaescape{...}}?

Joseph Wright

@DavidCarlisle That's more or less what I'm saying we should do, but that will alter \lua_now_x:n so I'd like to take some soundings first

@DavidCarlisle I'll add what I think I should change it to

David Carlisle

@JosephWright I think I'd better look at what the current definition is (given that \directlua expands like \write anyway, what does the _x version of lua_now do?

Joseph Wright

@DavidCarlisle The _x version is \directlua, the one without wraps the input in \unexpanded

@DavidCarlisle See updated issue :)

David Carlisle

@JosephWright ah now it makes more sense, I'll look at the update:-)

Joseph Wright

@DavidCarlisle :)

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends