12:32 AM
NFD is decomposed unicode (http://unicode.org/reports/tr15/)
I never thought about it before, does Latex (pdflatex or lualatex) care about the normalization of its input?

12:50 AM
@dedded Except for (optionally) XeTeX, none of the engines normalize the input but they also do not complain if the input isn't normalized. So e.g. in LuaTeX you can have different macros whose names are canonically equivalent etc. (DO NOT DO THIS!!!) and rendering is done in the form you input. (Especially under pdfTeX this means that accents will work much better if your input is in NFC, under LuaTeX it depends on the font and other factors)
In general I would recommend NFC input under most circumstances.

1:26 AM
Hmm, after a couple of experiments (and some fruitless googling) I believe that Emacs doesn't normalize its output, at least by default.

4 hours later…
5:43 AM
A user with a prohibited display name: tex.stackexchange.com/users/205326/stackexchange

1 hour later…
6:44 AM
@UlrikeFischer Normal Form Decomposition: breaking down a codepoint to separate out the accent(s) fromthe base char

@JosephWright yes, got it now, was a bit late yesterday ...

@UlrikeFischer :)
@MarcelKrüger Usually yes. For my specific cse, though, life is a little different. Greek requires that accents are removed when uppercasing. That means doing NFD first to get the list of codepoints manageable: you just have to look for the accents themselves after NFD.
@UlrikeFischer I could do a limited version in pdfTeX, but I really don't fancy implementing the full algorithm; lots of look-ahead for multi-byte sequences
@samcarter_is_at_topanswers.xyz, @UlrikeFischer On the transparency issue, is there a fix needed in the beamer core?

@JosephWright I think we should ask for comments and tests first. (I have e.g. no idea how well e.g. \MakeUppercase if at all works with greek)
@JosephWright which transparency issue?

@UlrikeFischer The pgf business with mask
@UlrikeFischer Yes, would be good: I'll release the update today so all 'basic' stuff works

@JosephWright I don't know if beamer relies somewhere on the pgf bug. But probably not: xetex always required unique shading/fading names, so if something were wrong you would probably have got some reports.

7:46 AM
Since when gets expl3 preloaded into the kernel?

@JosephWright thanks, I already looked it up in ltnews :)

8:51 AM
ooh vlistdirection
@MarcelKrüger ooh let's do this

9:04 AM
@UlrikeFischer Could you give me the email of the Greek maintainer?

9:18 AM
@JosephWright I'm on a train now. I will forward you a mail, but unedited. Simply ignore the content.

1 hour later…
10:25 AM
hi all

@Anush ello

@PauloCereda nearly lunchtime

@DavidCarlisle oh no

@JosephWright I said you should ignore it ;-) (I'm not sure if it makes all sense without the context of the previous mail, but I'm trying to avoid to get into puenc.def stuff that imho don't belong there)

10:42 AM
@UlrikeFischer local formats again

@DavidCarlisle :(

@JosephWright hard to believe all these people are running fmtutil-user, something must be triggering a local build.

@DavidCarlisle I certainly never run it and I got them too.

@DavidCarlisle Yes

@JosephWright maybe they use l3build install in a checked out base, in which case we can blame you

10:48 AM
@DavidCarlisle Why should that create a format in the texmfvar?

11:04 AM
@UlrikeFischer it creates in the texmf in my home dir, but that comes to the same thing, multiple formats in the path

@DavidCarlisle hm, I need to check this, but I'm installing quite often from base and imho I don't get local formats from it every time.

11:20 AM
@UlrikeFischer Exactly: it seems to be random
@UlrikeFischer I wonder if it's linked to having some job running that's using the global .fmt file: only obvious systematic cause

@JosephWright you mean two parallel pdflatex calls? That could happen if I run a test suite in the background.

@UlrikeFischer Yes, that sort of thing
@UlrikeFischer I often have TeXworks sitting with a run stopped at say a \show, which might do it

11:38 AM
@JosephWright it could explain it. And in the case of users it could be the problem of non stopped processes @PauloCereda wrote about in the texlive list.

@UlrikeFischer ooh

2 hours later…
2:04 PM
@UlrikeFischer ah no sorry I forget: if you install it only puts the files there but then if you do fmtutil-sys it builds from those files not the main tree.
@JosephWright but surely even on windows the format file shouldn't be locked for reading when being used?

2:36 PM
@DavidCarlisle Probably not, although I've never tested explicitly :)

3:01 PM
wow eso-pic starts with \NeedsTeXFormat{LaTeX2e}[2020-10-01]. That is quite daring ... @DavidCarlisle @JosephWright

3:17 PM

1 hour later…
4:22 PM
Boy the new syntax highlighting is terrible. tex.stackexchange.com/a/387741/2693

@JosephWright greek is complicated ...

@AlanMunn That's not even its worse. I like how it makes some chunks plain green: tex.stackexchange.com/q/565369/134574

@PhelypeOleinik Yeah that's pretty bad too.

@AlanMunn Though it seems we'll have a much better one soon thanks to @schtandard: github.com/highlightjs/highlight.js/pull/2726

@AlanMunn Tomorrow is Wear Something Gaudy Day - Maybe the syntax highlighting is preparing for the celebration?

4:32 PM

@AlanMunn while it is sometimes odd, I actually have more problems with code on golatex: at least on my laptop it is in some light blue on white, and I often can barely read it. golatex.de/viewtopic.php?f=21&t=23291

@UlrikeFischer medium blue links on only slightly paler blue background is even worse :)

We should ban syntax highlighting all along!

@UlrikeFischer Yes, that's not very readable indeed.
@PauloCereda When it works, it's not a bad thing.

4:47 PM
@AlanMunn :)

@UlrikeFischer I guess I didn't re-install the syntax highlighting after the new forum installation. Or there was an incompatibility, I don't rmember right now. This sample would be better, right? latex.org/forum/viewtopic.php?f=47&t=33778
@PauloCereda ban syntax

@StefanKottwitz the colors are much better, only the font is a bit too small for me.

@StefanKottwitz or tokenize everything as the same class. :)

@UlrikeFischer I will check what I can do about the code presentation

@StefanKottwitz thanks ;-) (I do manage even with bad color, I can always copy & paste). But it is nice if one can see more directly).

5:36 PM
Greetings all. Should a package file contain documenting comments or just code, with the documentation relegated to a separate file?

@LaTeXereXeTaL It depends on the author, if you choose to work in .dtx format it is better to have a single file that generates everything

6:13 PM
@PabloGonzálezL That is indeed what I currently do.

Sorry UK

Can someone remind me what the term "balanced text" means?

@LaTeXereXeTaL Taste thing, in general I like that there is the implementation part, it makes you understand (more less) what you can and can't do.

I used to know, but forgot, and I've been able to find a clear definition in the places I've looked.

It depends on the context...in perl it's something like {bal(lan)cea{a}do}
The number of braces/parentheses should be matched even if they are nested without crossing each other

6:27 PM
@FaheemMitha syntactic nesting
You have an open/close pair thingy. :)

@PabloGonzálezL The context is TeX. :-)

@FaheemMitha It is the same :)

Any specific kind of braces/parens? Does {} and () cover it?

@PauloCereda Hola don 🦆, como está el día de hoy?

@PabloGonzálezL rainy, but good. :) And yours?

6:32 PM
@PabloGonzálezL And if they cross each other?

It depends on the macro, usually {balanced mandatory argument }...but in TeX there are a lot of ways to define a macro and not all of them follow this pattern
@FaheemMitha You can cross it but escaped \{ (as a symbol) not as a delimiter of macro
@PauloCereda Bored at a national election training :(

@PabloGonzálezL I don't follow.

@PabloGonzálezL ouch... we have elections this year too...

I just meant - if there is crossing, is it no longer a "balanced text"?

@FaheemMitha It all depends, if it is some command that uses {...} the content within these must be balanced for these {...}, otherwise receive an error. In the xparse documentation there is a good example.

6:42 PM
@PabloGonzálezL I just want to understand the definition as it's normally used.

@FaheemMitha And in simple terms, if you cross it is no longer balanced

For example, would {ab(c}d) be considered a balanced text?

Noooo
Missing }

It's still balanced, for some definition of balanced. An {} pair, and a () pair.

Yep...but not cross
But depend of context
\verb+{...(})+
+ are in pair
but \foo{....(...}..) }

6:48 PM
So plusses have to balance too? I thought it was just brackets/braces.

usal show error
TeX is special and always depends on how the macro is defined

@PauloCereda Oh.

@FaheemMitha It depends, do the test with the examples of xparse and you will understand better the behavior, Verbatim is more special (and complicated) than the rest

@PabloGonzálezL Examples of xparse in the documentation?

6:52 PM
@FaheemMitha you have a catcode assigned to a char in TeX, and that's what's used to define open/close thingies. But TeX allows reassignment and catcode manipulation at expansion time, so it really depends on the context, as @PabloGonzálezL says.

@PauloCereda I was just thinking of default modes. Assuming that reassignment of this kind isn't common.

@FaheemMitha Yep
@FaheemMitha When you say default mode I assume you mean text mode?

@PabloGonzálezL Yes, I suppose I do. Nothing special. Definitely no math mode.

@FaheemMitha Even if at first sight one does not notice it, TeX always scans and reassigns the argument and then returns the result, that is what my context comment was about, assuming that we are only talking about macros with arguments usually delimited by {.}

@PabloGonzálezL Do balanced pairs only appear in the context of macros?

Does the following ressign \ to be of catcode 12?
\catcode‘\ =12
@PabloGonzálezL Ok. Thank you.

@FaheemMitha If you are working on simple things (text mode) you probably won't have to reassign "catcode" unless you want to complicate your evening for a while :)

@PabloGonzálezL I'm reading an example. I wouldn't dream of reassigning a catcode myself.

@FaheemMitha Yes, you need to read it beforehand...otherwise @DavidCarlisle clasic phrase will haunt you

@PabloGonzálezL Read what beforehand? And what classic phrase?

7:18 PM
@FaheemMitha I don't think I can quote you, you are entitled to @copyright by @DavidCarlisle (And I never knew how to quote in this chat)

@PabloGonzálezL You lost me there.

8:00 PM
I don't completely follow this example from "TeX by Topic".
It looks like this.
The importance of category codes in delimited arguments is shown by the following example:

\def\a#1 #2.{ ... }
\catcode‘\ =12
\a b c
d.

which gives

\a #1 #2.-> ...
#1<- b c
#2<-d

Explanation: the delimiter between parameters 1 and 2 is a space of category 10. In between a and b there is a space of category 12; the first space of category 10 is the space that is generated by the line end.
But shouldn't it be "between b and c" in that last sentence?
Since TeX starts looking to match the arguments after \a?

8:21 PM
@AlanMunn there is a plan in action to fix that
@UlrikeFischer not only that, it seems to lose spaces and run words together in weird ways.
@LaTeXereXeTaL one of those, yes.
@FaheemMitha no just { and } or at least characters of catcode 1 and 2 so {aaa )} is balanced.
@FaheemMitha no it assigns space to be catcode 12

@DavidCarlisle Right, space. Sorry, I misspoke.
@DavidCarlisle Ok.

@FaheemMitha no there is a catcode 12 space between a and b there is another one between b and c but he's not talking about that one.

@DavidCarlisle Yes, I realise they're both cat 12, because the space was redefined to be cat 12.
But I don't see how the former is relevant here.
Oh, hang on. When TeX is searching for the argument, it skips characters of catcode 10?
So if a space is set to catcode 12, it no longer skips the space?

@FaheemMitha yes although in the normal catcode 10 case there would not be a catcode 10 token after \a it would never be tokenised the space would be dropped while tokenizing \a but a catcode 12 space isn't dropped and is tokenized

8:58 PM
@DavidCarlisle I understand. So when the author of "TeX by Topic" wrote:
> When a macro with an undelimited parameter, for instance a macro \foo with one parameter \def\foo#1{ ... #1 ...} is expanded, TEX scans ahead (without expanding) until a non-blank token is found.
by "non-blank token" he meant a token not of catcode 10.

@FaheemMitha yes

@DavidCarlisle Ok, thank you.

@DavidCarlisle and it adds curious dots over letters.

What is the way to tell someone that their answer to my question has weird and unexpected consequences?
4

Here's a couple of options: 1. Make your bibliography \raggedright Just add \AtBeginBibliography{\raggedright} to you preamble. MWE \documentclass{article} \usepackage[left= 4 cm, right=4 cm]{geometry} \begin{filecontents}[overwrite]{\jobname.bib} @inproceedings{Meshlab, booktitle = {Eurogr...

Specifically

9:13 PM

@UlrikeFischer Can I past an image of them in the comment?
I made a new MWE that shows the cases in which that doesn't work
(i.e. its not from a package interaction or something, its just ONLY worked in the specific MWE I constructed. With a slightly different line length it does odd things)

@UlrikeFischer Its more "This answer doesn't work" type of thing

9:29 PM
@Canageek well it depends on your question. If you didn't mention crucial details than better ask a new question. If you think your question was fine and the answer doesn't really answer it then add a comment mentioning what is wrong.

@UlrikeFischer I thought I had dust on my screen

So its a line breaking question. They provided code that adds a penalty for breaking the line in the middle of the DOI. Works fine with the example I provided and the two they added.

But that is pretty much the *only* case it works in. As in, every other instance in my bibliography it messes up the formatting. So I copied some representative examples to the MWE they provided, and nope, it ONLY works with the examples they provided.
So the QUESTION is fine.
But I can't put in dozens of test cases in a MWE and they just happened to make an answer that works with that ONE test case and almost no others

@DavidCarlisle ;-)

1 hour later…
10:50 PM
Welp, just opened two meta questions to find out what I did wrong
And how to do better in future
and then realized I'd spent pretty much my entire day figuring out:
a) How to change "Bibliography" to "References" (Biblatex helpfully assumes that if you are using a class with chapters you want Bibliography not References, and don't provide an option to undo this)

b) Writing up two responses to that question and figuring out how to do a better MWE in future
Oh I'm not getting this chapter done for Monday, efffff
Or at least not without working this weekend

@Canageek asking on meta is good! thanks for that

@StefanKottwitz Well, in theory, less good when it makes me forget to eat lunch and spend much of an afternoon doing that when I'm trying to write an entire PhD thesis by April...and I'm only just finishing one chapter...out of 5-6...not counting intro and conclusion....
and I've been working on it for a month...

@Canageek each and every answer by me on TeX.SE Meta since Jul 17 got downvote(s), that was since when I posted my opinion on that Topanswers name promotion in chat, so maybe wait for answers by people who are not targeted (no reputation loss here, that looks like "only" manipulating user feedback stats to the site)

@StefanKottwitz Not targeted? o.0 Is there stuff going on in this stack exchange? I've been away for a long time, since I've either been not writing much LaTeX, or only using the same templates and code that I got working a few years ago. Now that I'm trying to do more fancy stuff, I've got more questions and came back...
2

@Canageek Who knows? I got 9 downvotes on this explanation (and +13, so you see 4) so do you think nothing is going on? or is that answer strange?

11:13 PM
In other news, I wish LaTeX had a 'restart' command instead of having to hit run or stop when I find and fix an error
Or better yet, rewind one line and start from that spot

@Canageek in a PhD, the writing is a formal thing you can manage - focus on your topic, grant forgiveness to LaTeX and tools and go ahead without worrying.
LaTeX things are much easier to fix than actual PhD things.

And don't spend a bunch of time testing notes2bib to see if it actually had a bug I can report to @JosephWright >.>

@Canageek @JosephWright is a very helpful chap, so let him know