@yo' and @PauloCereda -- I can vouch for the fact that Prague does have snow in the (northern hemisphere) winter. Warm boots are a necessity. But if you really want to see something outrageous, take a look at this: nytimes.com/2019/07/01/world/americas/…

@barbarabeeton tomorrow is the total solar eclipse yaaay!

I am interested in the following images (they are docs): cens46adultos.blogspot.com/2016/09/…

For example:

I am just curious, are these pages made in LaTeX?

The font is very similar to the LaTeX one

@manooooh Unlikely since the fourth image has straight quotes, but the other parts of the image have proper quotes. This is not something that LaTeX does without a lot of intervention. But Word does it simply by having two different preference settings used by different people who worked on the document.

@AlanMunn gotcha, thanks!!

The duck's natural predator

@UlrikeFischer Updates to CTAN

@JosephWright ok. What is the date of expl3 and l3pdf in the release?

@UlrikeFischer 2019-07-01

@JosephWright is this also on github?

@UlrikeFischer Just done now

@JosephWright installing now. Yesterday evening I got two issues about the driver command, so I should be hurry ;-)

Hello!

I am using MikTeX in my laptop. I want to write vector and I use \bf for that. But the problem with that is, the next character also behaves like vector, which may not be one! E.g., \bf{f}'({\bf {x}})=A, where $A$ is a matrix, but $A$ also became bold, which is misleading.

Plese help.

@Silent and don't use \bf in general: tex.stackexchange.com/questions/516/…

@CarLaTeX Wow! \mathbf worked! I will try not to use obsolete commands. Thanks a ton.

@Silent You're welcome!

@PhelypeOleinik On debug: looks like Bruno broke something :(

@JosephWright ooh

@PauloCereda I'll fix later: first I have to do releases

@JosephWright not-my-fault tag? :)

@PauloCereda :)

And this, boys and girls, is why this community is awesome! ^^

Also, samcarter. <3

A better screenshot. :)

@JosephWright did you already upload l3kernel + l3experimental?

@UlrikeFischer Yes ... did I miss something?

@JosephWright no I was only confused because you wrote that you have to do releases. I'm ready to upload too after lunch.

@UlrikeFischer Ah, I meant 'also LaTeX2e-dev' ... but first I have stuff to fix

@JosephWright I wished we had only one date format - I always miss to tag one date as it uses / instead of - or the other way round ;-(.

@UlrikeFischer I hope we are moving to purely ISO: certainly in my own code I have

@UlrikeFischer Have spotted a few deprecations to strip out of expl3: all 'expired' commands so only one or two people needing a prod

@JosephWright the interesting question will be how many people used the \driver-commands and will complain that they are gone ;-)

@UlrikeFischer We will see soon

@UlrikeFischer Next I want to work out what Bruno broke (thanks @PhelypeOleinik for picking this up)

@JosephWright Good Morning! I just woke up and stuff is already broken? ;-)

@JosephWright I'm almost done fixing that, if you can wait a little...

@PhelypeOleinik Sure: you fix it, I'll break it again :)

@JosephWright Seems like a good plan :-)

@JosephWright grr. I wanted to upload and typed "l3build ctan" and now the whole circle is starting again ;-(.

@UlrikeFischer Oops

@PhelypeOleinik I need to rebase the debug branch at some stage ... once decisions are made

@JosephWright This means that...

How did I never know this?! I just learned that capital letters are acceptable as length units. That is, \rule{1Em}{2eM} works just fine. Who would have thunk??

@PhelypeOleinik Once we finalise the new debug approach: there are still some interface issues

@JosephWright Oh, right!

@JosephWright I should send you something very soon

Is there an easy way (or any way) to produce a table where one column has the same value for every row? It doesn't really matter; I'm just curious. But I did a search and did not find anything.

@DavidCarlisle Do we step the file version for ltvers every release (or every PL), or only when changing the logic in the file?

@JosephWright how about minor revisions?

@PauloCereda this cries for a laughing and a crying mouse which are quantum entangled with the cat.

@UlrikeFischer ooh

@PauloCereda @CarLaTeX in other news, the Bär is now doing advertiment on a bus ^^^

@UlrikeFischer ooh :)

@JosephWright uploaded ...

@PauloCereda It's a policy question, really

@UlrikeFischer Cool: now we wait for complaints!

@JosephWright we can blame Frank ;-)

@JosephWright Almost done here. Only a couple of checks are failing

@JosephWright sorry

@JosephWright I have a number of examples in the doc folder, and I normally want to typeset them with two engines, and create e.g. example1-pdftex.pdf and example1-luatex.pdf. Is this possible?

@PhelypeOleinik vvv

:)

@PauloCereda Exactly :P

@JosephWright just when logic changes

@UlrikeFischer Oooh, Bär is famous!

@CarLaTeX he agrees.

@PhelypeOleinik l3build: 13 checks failed.:just assume your code is bug free, blame Joseph for an error in l3build and publish to ctan anyway

@DavidCarlisle I was just going to do that (sort of) :P

@DavidCarlisle (@JosephWright) I changed the code a bit and the log changed as well (the general outcome is the same, but TeX shows the error in a different macro. Should I change the test file?

@JosephWright we need to update unicode-data for ㋿ (wow system fonts got updated for that a lot quicker than they did for 🦆 :-)

@DavidCarlisle I am adding a flag --override-errors

@PhelypeOleinik ^^ :)

@PauloCereda That one and --blame=<subject>

Quick question: With TL19 what is the recommended method of setting up a missing unicode char? I'm using pdflatex and one of my presenters has ọ and I'd like to write the presenter names in UTF8 (as they are easier to parse). But ọ (\d{o}) is not setup by default

@PhelypeOleinik ooh good one!

@daleif we should add it but \DeclareUnicodeCharacter{1ecd}{\d o}

@PauloCereda implemented as print("Ulrike") ?

@DavidCarlisle and I still need newunicodechar?

@PhelypeOleinik in general yes, it's hard not to make test files reveal more implementation details than you want so quite often they need manual checking then updating after any change

@daleif no, that is in the format, no package needed

@DavidCarlisle or just print(random("Ulrike","David","Joseph"))

@daleif but then it's wrong 1/3 of the time.

@DavidCarlisle Thanks, I figured it would be something like that...

Now I just have to figure out how to use git...

@daleif it's a surprise. :)

@DavidCarlisle hmm, strange, I get errors unless 1ecd is written in uppercase

\DeclareUnicodeCharacter{1ocd}{\doh}

@DavidCarlisle if print("Ulrike") is correct, then print(random("Ulrike","David","Joseph")) must more wrong 2/3 of the time.

@UlrikeFischer Not at the moment: we'll need some new functions

@JosephWright I haven't made up my mind if it is really needed here. I think I'm misusing them as additional testfiles. Probably it is better to create in such cases real test files and compile the examples all with luatex. But in the xskak documentation I removed one of the included pdf as it would have need a latex+dvips compilation and I didn't see how to do it.

@daleif that's because it has to be in uppercase, sorry:-)

@PhelypeOleinik is this for a PR to l3 ?

@DavidCarlisle Yes. To the debug branch which Bruno broke :-)

@DavidCarlisle would it be possible for it to test for that? Such that one would get a better error message than

@DavidCarlisle Why?

@PhelypeOleinik why did I ask? why not:-)

@DavidCarlisle Oh :-) I thought you was going to say something about it. Ignore me then :-)

@DavidCarlisle ooh

@daleif no point in testing, it could just uppercase but it is primitive tex syntax "1ECD is a hex number and "1ecd is the number 1 followed by typesetting ecd, so you can actually get this in lots of places

@DavidCarlisle my point is that it is not obvious that this is the error.

@daleif sure but you would get the same error from \setcounter{chapter}{"1ecd} so while it sort of feels more natural to use hex for unicode it's a question of how many places throughout the system do you want to allow lowercase hex if it is not allowed everywhere?

\let\z\DeclareUnicodeCharacter\def\DeclareUnicodeCharacter#1{\uppercase{\z{#1}} would make lowercase work but you still get a bad error from non-hex such as {zzzz}

@DavidCarlisle with the exception of ^^ notation where lowercase is probably used just to confuse people.

@UlrikeFischer yes naturally

@UlrikeFischer No uses of \driver_... in TeX Live other than the obvious

@JosephWright I couldn't find it in my stuff either ;-). now lets wait what exist in the wild.

@StevenB.Segletes -- I'm pretty sure this is documented in the TeXbook.

@barbarabeeton there is a new package on ctan (logix) with lots of (mathematical) symbols.

@barbarabeeton @StevenB.Segletes Final paragraph of Chapter 10 (p. 61).

Generally putting a % at the end of lines is harmless, but if it's at the end of a filecontents* environment, it breaks the compile. E.g.

`\begin{filecontents*}{abbreviations.csv}%`

Unless I'm making some other error, or course. But removing that % makes the error disappear.

@FaheemMitha I wouldn't say it was harmless, putting % after \ifnum#1=0 would be a bad idea for example

@FaheemMitha given that filecontents is a verbatim environment where % is not a comment, what do you want it to do, write % to the file or do nothing? (like most verbatim environments filecontents is trying to ignore a newline at the start)

@DavidCarlisle I was just trying to add a comment. Is that a problem for verbatim environments?

Well, do nothing. I didn't realise the location would be a problem.

@FaheemMitha a comment in the file you were writing? Just put that on the next line

@DavidCarlisle No, a comment on the filecontents block.

I've moved it before the block.

I mean before the initial \begin{filecontents*}.

@FaheemMitha % is not a comment marker in a verbatim environment, it just writes a % it would be weird to interpret it as a tex comment surely?

@DavidCarlisle I assumed that the verbatim/filecontents environment started on the next line.

I guess that was incorrect.

@FaheemMitha well it starts where it says \begin there is a clue in the macro name:-)

And apparently it was upset that I removed the newline.

@DavidCarlisle Ok.

@FaheemMitha we probably document that somewhere, it is expecting a newline so it can remove it...

@DavidCarlisle Yes, I figured that. But if it doesn't find a newline, it could also choose not to freak out.

I assume this is a feature of the original filecontents.

Plus the error message isn't exactly helpful.

@DavidCarlisle quack

@FaheemMitha we didn't have alot of room for error messages in 1993..

@DavidCarlisle That's a long time ago.

@FaheemMitha The whole of the latex2e design is influenced by the fact that if it was any bigger it would not have fitted on an emtex installation in msdos

@DavidCarlisle oh no

@DavidCarlisle How big is that?

@DavidCarlisle I once wrote a program for my dad. Since it would not be helpful to add a descriptive message error, all abnormal executions would display a "call me, dad." message. :)

@FaheemMitha well with latex2e+article class+amsmath and no other packages, you had around 50 command names left to define so at most 50 in total \label or \newcommand etc

@DavidCarlisle I meant the size, in MB, or whatever. But apparently you had a programmatical restriction too?

@FaheemMitha well the machines were 640K but that was not really the issue it was the size of the hash table in emtex

@DavidCarlisle Oh.

So, does anyone miss the good old days, I wonder.

Computationally speaking.

@UlrikeFischer -- Thanks. Interesting. I see some things in the documentation that look suspicious, and will consider following up.

@FaheemMitha -- I miss them in one way ... I have little patience with people who expect instantaneous output of a 500-page book. It used to be appreciated if such a job would finish overnight.

@daleif -- This one's for you: Vertical alignment of left/right subscripts

@barbarabeeton Personally, I like it when things happen fast. But I guess we live in an instant gratification world.

@FaheemMitha -- Indeed. Sigh.

@PauloCereda I bet it solved more issues than actual error messages :-)

@barbarabeeton Nowadays people don't even bother learning TeX's interactive mode to fix errors at compile time. I'm guilty of that :|

@PhelypeOleinik -- At least you recognize that there is an interactive mode! (Learning how to use it will pay for itself in the long run.)

@barbarabeeton Looking from this angle... I know the basic insertion with I. Not much more.

@PhelypeOleinik -- "I" is very powerful. With it you can add missing definitions that will allow you to continue. "X" means "quit"; "Q" means "quiet", i.e., keep going until completion or failure, but don't bother reporting errors on the screen (they still show up in the log). "Q" can be dangerous -- you may not realize that you're in a loop. A digit means skip that many input characters, including the one just read in. Try it!

@manooooh The used font is probably some type of Baskerville. In case you want to reproduce the look ...

@RalfStubner wow, didn't know about that font, thanks!! I just was curious hahaha

@barbarabeeton it is cloudy here right now. Hope we could see the eclipse. At the moment I am at work but surely I can watch it

Yes, but the interactive mode doesn't allow one to back up.

@FaheemMitha -- That's true, no backing up. If the situation is fatal, that's when you need to use "X". But often, you can continue and get rid of (i.e., identify) a few more glitches before canceling. The trick is learning what's fatal, and what's just annoying.

@barbarabeeton " I'm pretty sure this is documented in the TeXbook." Well, that is a very nice and polite way to tell me to RTFM! Thank you for your kindness.

@HaraldHanche-Olsen Much obliged!

@barbarabeeton Oh, yes, I use “X” quite often too. The other ones are shortcuts to \<something>stopmode, right?

@barbarabeeton Some months ago I tried to use the digits thingy to figure out what “special codes” Donald Knuth meant in “The TeX tuneup of 2014”, but didn't succeed.

@PhelypeOleinik -- I wouldn't look at the other responses as "shortcuts" but they do all result in a stopcode, except for "Q", which enters batchmode, and can be unpredictable, as mentioned above.

@PhelypeOleinik -- The "special codes" are meant to get around the loop in the file texbook.tex that prevents that file being compiled. (The reason for that is to try to keep people wishing to post a pdf copy of the TeXbook on the web from doing so; read the comments at the top of the file to learn why.) I do know how to do this, but I can't tell you; otherwise I would have to delete and expunge myself.

@barbarabeeton Yes, I tried to do it just for fun to find out what those codes are, not to keep the pdf file. I didn't expect you to tell the codes: that would spoil the fun :-)

@PhelypeOleinik -- Yes, you're definitely correct that it shouldn't be there. I've written to the webmaster and officers of the group, several times over many years, but never get a response. Sadly, China doesn't seem to have the same recognition of copyright as most other countries. I guess it's time to try again.

@barbarabeeton Oh, too bad :/

@barbarabeeton I think I should probably delete my message above... I flagged it.

@UlrikeFischer All in the PUA. :-(

@JosephWright I did a pull request for the debug thingy, but I'm not sure I filed it to the debug branch or the master... Reading xii.tex is easier than this git thing :P

@DavidCarlisle I think you and @PauloCereda missed this:

@barbarabeeton Since you know all the history, did Knuth ever consider the possibility of a Lisp-based typesetting system?

@AlanMunn -- All I see is "403 Forbidden" on a pretty blue background.

@barbarabeeton Weird. It plays for me in Firefox.

@FaheemMitha -- I don't know for sure, but I don't believe so. Even in the first cut, he wanted it to be platform independent, so he chose a well-defined subset of Pascal. The actual situation should be recorded somewhere in Digital Typography. (But my copy of that is currently in a box.)

Here's link to the spot in the whole video:

@AlanMunn -- Thank you! Wonderful!

@barbarabeeton Ok.

@barbarabeeton Did a platform-independent Lisp not exist at the time, then?

@StevenB.Segletes I’m pretty sure you read too much into that comment. @barbarabeeton knows the texbook quite well for sure, so no doubt she knew she had seen the info there. Pointing that out is useful: It made me sufficiently curious to go look for it. (I had a hunch or two, and the second one panned out.)

Furthermore, this is chat. It is by nature informal here, so there is much less obligation here than on the main site to give complete answers.

@AlanMunn I got 403 no permission

@DavidCarlisle Yeah, that seems to be the problem. Try the youtube link above ^^^

@AlanMunn will do ..

@FaheemMitha you haven't looked at the tex source code I take it:-)

@DavidCarlisle I might have at some point. I probably have. Of course, I didn't understand anything I read. Why?

As I recall, there's a book with the entire printout of the TeX source code (the original Pascal version), with commentary. I think I've looked at it. Probably at UNC.

@FaheemMitha because the coding style is all global variables and gotos, it's as far from the functional list paradigm as you could imagine

@DavidCarlisle Actually, Lisp is multi-paradigm.

@FaheemMitha yes tex-the-program the source code for that is in texlive (although the printed version has indices and other things)

@PhelypeOleinik Don't worry, I'll sort it

I don't see why people insist on calling it functional. It can do functional. It doesn't insist on functional. Unlike some languages.

@FaheemMitha true but still it's probably not your favourite programming language if you want to do everything with explicit global state and program jumps:-)

@DavidCarlisle Is there some necessity to do things with explicit global state and program jumps?

@FaheemMitha also I doubt it would have been fast enough in 79-80 when the decision to go to pascal was made

@FaheemMitha not really, I mean authors choose programming languages that match their programming style

@DavidCarlisle I know that the current Lisp flavors didn't exist then, but weren't there fast Lisps at the time?

@JosephWright Okay. Thanks!

@FaheemMitha I doubt it, TeX is really blindingly amazingly fast. DEK is above everything else the master of fine tuning algorithms and TeX does what it does with almost no overhead

@DavidCarlisle I'm sure that's all true. I was just wondering whether fast Lisps existed at the time.

@DavidCarlisle Have you noticed that DEK and Bruno have never been seen at the same time ...

I'm not very familiar with the history. SBCL is quite fast, but it wasn't around 30 years ago. Though its forebear CMUCL was sort of around at the time.

The history isn't very clear on that.

@JosephWright Bruno?

@FaheemMitha define fast ther were lisp machines etc but no one could afford them. To get tex to run on a PC much of the core inner loop had to be hand compiled in the early implementations if I recall correctly, which is possible as it's all just moving indices round in integer arrays, there really is no space to run things over a lisp virtual machine in a 640K msdos box

@FaheemMitha LeFloch

@JosephWright Ok, got it.

@JosephWright although I have not seen either of them with underwear over their clothes

@DavidCarlisle Is it a bird, is it a plane ... no it's macro man

It's funny the way people describe becoming part of TeX development. Like it just sort of happened by accident.

Oops, I stumbled and fell into TeX.

@FaheemMitha it's not the sort of thing you would plan or volunteer for.

@DavidCarlisle You mean there weren't long sleepless nights considering whether to do TeX development? Or not?

@FaheemMitha How do you think most of us ended up on the team (even FMi)

@FaheemMitha No I was just offered a free weekend trip to Hamburg and 30 years later I'm still working for nothing:-)

@DavidCarlisle Sounds about right

@JosephWright No idea. I've always assumed that it must be buried masochism, along with perhaps a saviour complex, or something. Plus, really, really liking typesetting, presumably.

@DavidCarlisle Hamburg is a nice city, or so I hear. I had some guests from there, recently.

@DavidCarlisle Should have taken the trip to Brazil ... that was fun! (@UlrikeFischer, @egreg)

@JosephWright timing was optimally bad

@DavidCarlisle Yes, true

@DavidCarlisle We should get Will to arrange a meeting in Australia!

@HaraldHanche-Olsen -- Plus at the moment all the books in my typesetting library are still in boxes since my retirement, so if something isn't online, I can't refer to it. (I also have available only an old, outdated laptop, without a working TeX system and with flaky browsers, so sometimes even looking something up on CTAN is fraught.) It's a good thing that I do still seem to have a reasonably well working memory, at least for things that happened sufficiently long ago.

@JosephWright -- If you succeed, while he's doing that, how about arranging a TUG meeting?

@barbarabeeton :)

@HaraldHanche-Olsen I was merely having fun. I took no offense (nor meant any to @barbarabeeton) by my comment. In fact, such an RTFM comment directed at me might have even been justified. I am just happy by the collegiality of this community.

@StevenB.Segletes Good!

@DavidCarlisle I got a free T-shirt. ;-)

@DavidCarlisle, @UlrikeFischer, @egreg First dev release sent to CTAN

@JosephWright Let's wait for the issue reports!

@egreg This first one is strictly a doc-only change: mainly have to see how it goes with CTAN/TL/MiKTeX. Next week ... UTF8-and-space ... more interesting

@JosephWright will be interesting to see how it is included in texlive - and if it clashes with all my local settings ;-)

@JosephWright no latex-dev.ini ?

@DavidCarlisle No, don't think that was the plan: should use standard latex.ini as the file names are unchanged

@UlrikeFischer Indeed

@DavidCarlisle We still have the tex-ini-files to think about

@JosephWright so how will users switch between the main release and dev? don't we need a latex-dev.fmt in the end?

@DavidCarlisle Ask @UlrikeFischer ;)

@JosephWright oh aux-tree switching? hmm

@DavidCarlisle when I tested this in march (see mail 4.3, "testing "latex-next" formats") it worked simply through the progname settings in texmf.cnf.

@UlrikeFischer yes that;s what I thought but didn't that mean you had a different progname?

Heyo chat!!

any ideas how can i draw this using latex?

seems surprising

@DavidCarlisle yes I have a pdflatex-dev.exe, but this is simply a renamed copy of pdflatex.exe

any of the usual ways `\includegraphics, tikz, pstricks, .... do you just have a picture or do you have a parametrised curve to plot?

@UlrikeFischer yes but I suspect for fmtutil-sys it's easier to have a separate ini file (a copy of the standard one more or less) rather than have a non standard copy "at the end"

@UlrikeFischer you made the format as pdflatex.fmt and then renamed it to pdflatex-dev.fmt by hand as well I assume?

@DavidCarlisle no I added pdflatex-next pdftex language.dat -translate-file=cp227.tcx *pdflatex.ini to a local fmtutil.cnf and then called fmtutil-sys -byfmt=pdflatex-next.

@UlrikeFischer oh you probably read some fmtutil documentation as well, that's cheating:-)

@UlrikeFischer but yes that's better it would want a line adding to that cnf anyway so using the same ini file but specifying a format name there is better, thanks

@DavidCarlisle We tipped Karl off ...

@DavidCarlisle ;-). I got a bit practice with formats with all this luatex variant testing. But it seems to work fine. They only need to setup the texmf.cnf and the fmtutil.cnf correctly. I tested in miktex too, it should work similar.

@JosephWright yes I know I was just wondering exactly how but it seems Ok, I see ctan don't like our back to the future dating system

@DavidCarlisle Yeah ... I was not really sure what to do

@UlrikeFischer Yes, me too: hopefully this will all 'just work;

@JosephWright we could put a real upload date in README and document there that the format date is an arbitrary future date so latexrelease package doesn't get confused, then use teh readme date in the upload?

@DavidCarlisle Yes, could do that: probably a quick team chat first

@UlrikeFischer Pass!

@AlanMunn ooh

@JosephWright <3

@JosephWright hmmmmmm

@JosephWright quack team chat? :)

@yo' ooh

@PauloCereda I'm sure @DavidCarlisle manages; I would be pretty surprised if emacs didn't have a duck extension.

@yo' it probably does, built-in. :)

@yo' C-c C-d dinner

@AlanMunn :set alan mean-mode

@PauloCereda emacs responds: Trailing garbage following expression

@AlanMunn oh