TeX, LaTeX and Friends

8:29 AM

@DavidCarlisle Really needs the engine people to engage with this in the sense of aiming to make it general

@JosephWright yes although first I need to work out why test suite hanging (probably a local issue but I couldn't check it in last night and it was too late to start debugging:-)

Joseph Wright

@DavidCarlisle :)

Joseph Wright

9:39 AM

@DavidCarlisle See you've checked it in

David Carlisle

@JosephWright yes was just writing mail, there are a few luatex things I need to check, but wanted something Javier could look at re babel and as it was passing in xetex and etex seemed best to check it in before I got too far ahead of svn

Joseph Wright

@DavidCarlisle So overall the plan looks solid?

David Carlisle

@JosephWright possibly but the idea of peer review is others comment not the one who just ~~broke~~ improved everything:-)

2

@JosephWright It went a bit beyond bug fixing, and changing both \document and \@outputpage makes it a bit scary, but since existing behaviour was clearly buggy and it's all guarded by \IncludeInRelease seemed changing them is the right thing.

@JosephWright why are testfiles/helpers in testfiles/helpers and not in ../support (which was cause of my local setup issues:-)

Joseph Wright

@DavidCarlisle ../support is 'general stuff to support tests', ../testfiles/helpers are for files directly tied to specific tests but not in the .lvt files themselves

David Carlisle

@JosephWright yes but that distinction is a bit subtle at 2 in the morning

Joseph Wright

9:56 AM

@DavidCarlisle :)

@egreg No email yet?

@DavidCarlisle We can look to revisit that ...

David Carlisle

@JosephWright well it's OK really but what happened was that I was experimenting with variants for setting/restoring language in verb and page heads, but only a dozen or so tests are affected by that at all so to speed things up I moved the testfiles directory to one side and just had a directory of the tests i wanted. But as I didn't copy in the helpers got a missing file error on \usepackage{something} then it hangs in a yes -loop.....

@JosephWright If you run it one at a time it's in scroll mode so it just fails normally with the missing file error shown in the log but the setup for running them all silently doesn't take well to missing files.

Joseph Wright

@DavidCarlisle Yes, that's by design: running all of the tests is largely for checking things are 'working', when you are setting up or chasing an issue you expect to do them one at a time

David Carlisle

@JosephWright I know but sometimes you want to run a few at a time, because hacking the core rather than writing a new test file. But really not doing it after midnight is probably the best plan, but it's hard to build in such rules.

egreg

@JosephWright Answered

Joseph Wright

@egreg Thanks: I've passed to Frank/Rainer (for admin)

@egreg Just heading out: will send you some info later

Paulo Cereda

10:21 AM

Saturday, yay!

Sebastiano

@DavidCarlisle Thanks for your suggest. I deleted my answer. As we can see some times I do not understand the user's request.

@PauloCereda , @egreg , @Johannes_B, @JosephWright Gooddddddd morning to you with such warmth

Paulo Cereda

@Sebastiano konichiwa!

Sebastiano

@PauloCereda :-)) I do not know what it means but it certainly is an auspicious greetings

David Carlisle

@Sebastiano no honestly it's not that, it's just that you do not properly test your answers. Your answer gave exactly the same error as the question was asking about. So even if you did not understand anything about the question why post an answer that generates a tex eror?

Paulo Cereda

@Sebastiano It's 'hello' in Japanese, do not worry. :)

Sebastiano

10:26 AM

@DavidCarlisle With papeeria I not give me no errors.

David Carlisle

@Sebastiano do not use that system for checking answers then. But it is hard to believe it does not give you an option of checking the log file.

Paulo Cereda

@DavidCarlisle: I have to use SVN, help! :)

David Carlisle

@PauloCereda alias cvs=svn then just use cvs as if it was rcs, and you'll be fine.

Paulo Cereda

@DavidCarlisle you are not solving the underlying problem, I guess. :)

David Carlisle

@PauloCereda the problem being the duck between keyboard and screen?

Paulo Cereda

10:31 AM

@DavidCarlisle not that much underlying. :)

@JosephWright: Yooka-Laylee is released today, I guess. :)

David Carlisle

@JosephWright luatex failures are OK, it's a documented luatex difference setting \language no longer makes whatsit nodes (so the \showbox output is different) it stuffs the information into the glyph nodes viewable only from lua.

Sebastiano

11:05 AM

i.stack.imgur.com/YbVcQ.png @DavidCarlisle Dear David I do not lie. You can see that I have no errors

David Carlisle

@Sebastiano I never said you'd lied:-) so what happens if you click on that "open raw logs" link in the image you show.

@Sebastiano what I said was "But it is hard to believe it does not give you an option of checking the log file." and your image shows that there is such an option.

I can't use firefox nightly this morning :(

Paulo Cereda

@DavidCarlisle nananana stable FF 52.0 here. :P

David Carlisle

@PauloCereda I started it 4 times and it sent 4 crash reports to mozilla so I took the hint and switched but now i'm on an old firefox that gmail says is too old to be supported, I suppose I could update it, or use chrome or ie or something but ...

Paulo Cereda

@DavidCarlisle lynx :)

David Carlisle

@PauloCereda I miss emacs w3-mode

Paulo Cereda

11:16 AM

@DavidCarlisle eek

David Carlisle

@PauloCereda I used to use that quite a lot

Paulo Cereda

@DavidCarlisle :)

David Carlisle

@PauloCereda over original dialup lines when paying by the second for internet access, a browser that (a) was emacs and(b) didn't download images was a good thing

Paulo Cereda

@DavidCarlisle indeed

David Carlisle

@PauloCereda I trust you are agreeing with the first part

Paulo Cereda

11:18 AM

@DavidCarlisle no :)

David Carlisle

@PauloCereda I'm back on FF 55!

Paulo Cereda

@DavidCarlisle Yay! :)

David Carlisle

@PauloCereda had to re-install from Chrome, as nightly was crashing on startup so couldn't update itself so was a bit stuck...

Paulo Cereda

@DavidCarlisle Canary is no more?

David Carlisle

@PauloCereda no Chrome and I had a disagreement over math support so I stopped following its dev channels for now (although it's not impossible that gets picked up again, there's been a lot of work in webkit's math support that could be moved to blink, and in harfbuzz that blink will get whether it wants it or not)

Paulo Cereda

11:29 AM

@DavidCarlisle oh my

@DavidCarlisle: you are one of a kind, kind sir

David Carlisle

@Sebastiano so does the log show errors or does papeeria have an old version of beamer that manages not to give an error there?

Paulo Cereda

@David: I followed a bit of discussion in the dev channels some time ago, but the email frequency was just too high, I gave up. I had several lists going on...

David Carlisle

@PauloCereda the discussion over math support wasn't on public lists (which is a good thing, really:-)

Paulo Cereda

So far, I only follow some TUG lists and the Apache Commons one.

@DavidCarlisle so you could call them names. :)

David Carlisle

@PauloCereda the discussion was "wide ranging"

Paulo Cereda

11:35 AM

@DavidCarlisle ah

David Carlisle

@PauloCereda but made a lot more sense a month later when google announced they were forking webkit to make blink and throwing out all kinds of stuff that apple had added. that hadn't been mentioned in the math discussions (not surpising that they could not do that) even though math got thrown out at that time.

Paulo Cereda

@DavidCarlisle oh I see

David Carlisle

@PauloCereda not on the latex-l listserv list?

Paulo Cereda

@DavidCarlisle Ah yes, that one too. :)

@DavidCarlisle: did you buy M a Switch?

David Carlisle

11:55 AM

@PauloCereda no

Johannes_B

1:36 PM

Does anybody remember that there is this documentation thing going on on SO? Slow movement in the LaTeX part.

Joseph Wright

@Johannes_B I think we mainly decided to avoid it ...

Johannes_B

@JosephWright I completely agree with that.

yo'

5:19 PM

@JosephWright @Johannes_B IIRC I asked once whether it's possible to remove LaTeX altogether from this stupid thing. No response.

Joseph Wright

@yo' The Powers are very keen on the plan, so I think it's unlikely

yo'

5:41 PM

@JosephWright Right. Let's propose they put a big banner THIS SECTION IS CR*P I MEAN RUBBISH to the LaTeX section.

Johannes_B

6:34 PM

@yo' LaTeX is not a programming language, at least not fitting into the documentation scheme there.

Alan Munn

7:18 PM

@JosephWright Hi are you informally maintaining csquotes? If so, here's bug (but easy to fix). tex.stackexchange.com/q/358000/2693

Joseph Wright

@AlanMunn I am maintaining it (or at least formally a fork from PL's original)

Joseph Wright

7:37 PM

@AlanMunn Fix on way to CTAN

Alan Munn

@JosephWright Great. Thanks. I've updated the answer to reflect that too.

Sebastiano

8:48 PM

\documentclass[aspectratio=1610]{beamer}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{amssymb,amsmath}

\begin{document}
\begin{frame}{Example Quadratic Formula}
\begin{columns}
\column{0.5\textwidth}
Discriminant
\begin{math}
\text{function} \left[d\right]=\text{discriminant}(a, b, c)\\
\Delta = b^2 - 4\cdot a\cdot c
\end{math}
\column{0.5\textwidth}
some text
\end{columns}
\end{frame}
\end{document}This is XeTeX, Version 3.14159265-2.6-0.99992 (TeX Live 2015) (preloaded format=xelatex 2017.3.11) 11 MAR 2017 20:47

Johannes_B

@Sebastiano No line here to end.

Sebastiano

@DavidCarlisle Dear I did not change anything in the code; I am attaching also the image

@Johannes_B hi, Johannes I have absolutely looked at the log file. I've seen now. But with Papeeria it works without giving me any error. You can do the test yourself.

Johannes_B

@Sebastiano Online editors try to be even smarter than offline editors. You are a customer, you should feel save. That means: Hide some warnings or even errors to make you feel great. It is the approach of a company.

@Sebastiano As far as i know, Microsoft Wrod does not give erros, it just works. Or it doesn't, depending on how far you are to the deadline.

Sebastiano

@Johannes_B I never thought to hear the most good. If I had installed MikTeX I would have noticed the error. The sooner I'll have to install. I am haywire for so much work at school and at my house. I simply told the truth as I always do, for better or for worse and I apologize right away when I'm wrong. I have not read so much as a second log file.

Johannes_B

c:35975975 Why apologizing? It isn't your fault. (I never used that service)

Sebastiano

9:03 PM

@Johannes_B For example look my question tex.stackexchange.com/questions/355320/…

i.stack.imgur.com/Lt9Ya.jpg I changed many times my question, I always see in a short time and how I am following despite being a tenured professor for many years a university course

@Johannes_B I always try in as possible and with such humility to always do good and do well

Johannes_B

@Sebastiano You are asking for a ransom letter.

@Sebastiano If you are alright with taking a screensot of the blackboard, you can do it.

You cannot compare handwriting with a predefined font.

Sebastiano

@Johannes_B No but jokes. I would like to make you understand that I make every day a large number of errors due to many engagements. I wanted to associate the classic calligraphic letters with those of mt2pro. Now I salute you because I'm in the car and my laptop battery is low. Good evening and good Sunday

Johannes_B

@Sebastiano I am not an aficianado of mt2pro, but on the other hand, i sense there is a communication problem. I really don't know (and i saw the edits) what your goal really is.

David Carlisle

10:01 PM

@Sebastiano if you use an editor that only shows the input source and the typeset output then of course you do not see the error by default, many desktop editors do the same, but still if there is an error the pdf is completely unreliable and should never be used. and definitely not used as an answer.

3

@Sebastiano that's nothing really special about papeeria you would see the same with miktex or texlive if you run in batchmode. But the idea of batchmode is that tex gets to the end and shows something rather than stopping at each error, but the intention is then you look at the errors at the end, not that you assume everything is OK.

David Carlisle

10:19 PM

@JosephWright thanks for the typo fixes, it probably wants a complete re-write nearer the release date but seemed worth getting some headings and words into the file.

Joseph Wright

@DavidCarlisle Sure: it's 2e so it's down to you/Frank mainly ...

David Carlisle

@JosephWright no one believes your "I don't do 2e" anymore

Joseph Wright

@DavidCarlisle Ah, but remember I'm happy to address such issues as will carry forward to a new format

@DavidCarlisle For example, siunitx is clearly a 2e package but the ideas should carry forward, particularly in terms of the ongoing v3 work

@DavidCarlisle You don't get a lot from me on most of the 2e threads!

David Carlisle

@JosephWright working on gradual corrosion of your defences.

Joseph Wright

@DavidCarlisle What you don't see is during the team meetings I tend to pay only just enough attention during the 2e parts to write some vague notes :)

@DavidCarlisle What I want is Frank to put Alice in the SVN

David Carlisle

10:26 PM

@JosephWright oi don't give potential new members a bad impression, they have to be ~~tricked~~ enthused into picking up the work.

Joseph Wright

@DavidCarlisle :)

David Carlisle

@JosephWright speaking of things going forward did you see the \scalebox bug? It's certainly right but I must check what appendix g says is supposed to happen:-)

Joseph Wright

@DavidCarlisle Have also to bear in mind position on likely outcomes and some updated ideas on format stability

@DavidCarlisle Didn't look in detail

David Carlisle

@JosephWright it defers the scaling to the back end but then just forces the box to have the right size by setting its \wd explicitly but somehow in reboxing from a math fraction atom into the horizontal list the forced box size is dropped (as if it had been unboxed) so the fix was to instead wrap it in a hbox of the requested size, with \hss padding (as I did anyway for negative sizes)

@JosephWright not sure what the l3 drivers do there "I don't do l3" :-)

Joseph Wright

@DavidCarlisle I'll take a look then ...

David Carlisle

10:36 PM

@JosephWright meanwhile based on latex-l discussion we might need nag.co.uk/numeric/fl/nagdoc_fl26/html/s/s14aaf.html#description If I could just get FFI working reliably....

Joseph Wright

@DavidCarlisle :)

David Carlisle

@JosephWright actually I may yet get that working, we have python interfaces and lua interfaces would probably not be dissimilar, and it might be interesting if pgfplots could then plot the library functions directly instead of gnuplot for eg nag.co.uk/numeric/fl/nagdoc_fl26/html/d03/…

Ulrike Fischer

@DavidCarlisle @Sebastiano Well imho this papeeria is rather special. One get notifications about error message but if the beamer class is used it seems not to recognize all of them. Their error parsing looks a bit broken ...

Joseph Wright

@DavidCarlisle Looks OK

\documentclass{article}
\usepackage{amsmath}
\usepackage{expl3}

\begin{document}

\ExplSyntaxOn
\hbox_set:Nn \l_tmpa_box { a }
\box_scale:Nnn \l_tmpa_box { 2 } { 2 }
\def\contents{\box_use:N \l_tmpa_box}
\ExplSyntaxOff

$\frac{\contents}{\contents}$
$\frac{\contents}{\contents\kern1sp}$ % correct

\end{document}

David Carlisle

@JosephWright sorry my code is rubbish!

Joseph Wright

10:43 PM

@DavidCarlisle Well I had the advantage of having named variables for everything, starting from a load of material I could read and re-implement with some idea of the end point and a test right ready to go. Also, please check I've understood the point correctly!

David Carlisle

@JosephWright yes seems Ok here

Joseph Wright

@DavidCarlisle Phew

David Carlisle

@JosephWright I had a piece of paper with a sketch of required commands from a conversation between Frank, Chris and Leslie from before I joined:-)

Joseph Wright

@DavidCarlisle Starting out with an FPU is also handy

@DavidCarlisle Goodness, that is going back some

@DavidCarlisle One thing I would like the team to do is look at the currently experimental stuff I've done in terms of box ops: they are I think all working correctly so could move to stable status

@DavidCarlisle Remember I've go a side project to implement TikZ in expl3 :)

David Carlisle

@JosephWright that would be good I think, then you can do beamer2

Joseph Wright

10:47 PM

@DavidCarlisle Need style sheets for that

David Carlisle

@JosephWright Hm I see Akira's changing the list of libraries that are linked statically to the execuatble, hmmmmm

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends