3:18 AM
@JasperHabicht From time to time, it usually comes up because a user gets angry when you ask them to add an MWE. I used to close just-do-it-for-me questions in the past, but then I change my mind, it's a useless battle: tex.meta.stackexchange.com/a/8121/101651

Hi! Anyone here can suggest the best latex IDE? thanks! :3

@DialFrost That's a matter of opinion: tex.stackexchange.com/questions/339/latex-editors-ides

@CarLaTeX Ye I'm on the link already, just not sure which to use :/

@DialFrost If you're using Windows, I recommend you TeXstudio. I use Notepad++ with a Nppexec to compile with arara, but I am a bit nerd. Real nerds use Emacs or Vim.

4:01 AM
hmm ok thx

4:26 AM
Is there a syntax xdocumentation of latex?
I couldnt find any online or here

4:48 AM
@DialFrost Try learnlatex.org

Ah thanks!
Hiya @egreg! :3

2 hours later…
7:07 AM
@DialFrost LaTeX: A Document Preparation System by Leslie Lamport

8:05 AM
@JasperHabicht it suddenly comes up because one user is serial editing everything - no matter how old the question is and how many contemporary problems won't be seen because they are pushed off the front page :(
^^^ I doubt people had trouble to find the "Problem description" without this giant header ...

Yay I posted my first answer :3 (even though I have insanely little experience in latex)
@JosephWright Interesting I think that's a good book! Thanks

@samcarter Yes, I also saw that it was the same user several times in a row. I was wondering whether such an edit was considered OK. But, since it does no harm to the original question, I think it is fine. They popped up in my "suggested edit" review list and after the third one, I was wondering what's goint on

@DialFrost whether or not it is "good", it is the definition of LaTeX

@JasperHabicht It does not harm the edited question, but it takes attention away from other questions which deserve their time in the spotlight.

@samcarter Yes, this is true. On the other hand, editing and improving is part of this site. So, the problem is rather that the user edited four or five questions in a row and not that they edited them at all, I think.

8:19 AM
@JasperHabicht more like 20 something questions ....
@JasperHabicht on a fast flowing site like SO this wouldn't matter, but on smaller sites, thats' annoying

@DialFrost Hi!

@samcarter Oh, really? I only saw a few and it already made me wonder =D

@samcarter do such edits still pop up on the active list? (as I never use it, I don't know)

@UlrikeFischer yes, SO still didn't manage to implement non-bumping edits. I exclusively use the active list to see when OPs might add additional information, but such edits make it really hard to find anything :(

@samcarter I know about this discussion and I know that because of this happening one shoudl "edit with care". However, I really think that this should actually not prevent users from edititng and improving posts. I have to admit that I also sometimes edit only a few things in my own answers to clarify or improve the wording. I find it difficult to decide in such a situation what is better: reject the suggested edit or let the entry pop up in the active list

8:31 AM
@JasperHabicht I think if an edit adds additional information or fixes serve formatting problems (e.g. missing code block) then it is ok, but just adding markup as in the screenshot above should be rejected

8:49 AM
@samcarter I answered a question and when they comment I will suggest to slow down a bit.

@UlrikeFischer Thanks!

2 hours later…
10:57 AM
Quick question, why does latex still provide output even if there are errors in the code?
E.g. after a few tweaks to someone's code, I achieved the desired table (i think) but there are 174 error messages in the output log :/

@DialFrost it does not by default

@DavidCarlisle Uh but it does for me! On overleaf

@DialFrost Overleaf likes to torture the people who try to help their users :P

@DialFrost overleaf runs in scrollmode so it does not stop on errors

@samcarter :(
@DavidCarlisle Oh dear

11:02 AM
@DialFrost not just overleaf, most IDE do, so thy can parse the log and step through errors, tex will make a pdf if you want any log trace to the end

"if you want any log trace to the end"? Wdym? Sorry :P
I'm assuming it's not a good idea to post code that does the job but has errors?

@DialFrost eg overleaf puts a red circle saying 100 errors, and a marker on each source line with an error. It could not do that if it stopped at the first error.

@DialFrost don't

@DialFrost not as an answr no, an answr with errors is wrong

@DavidCarlisle Hmm, I understand, but where are you getting? Sorry I'm quite unfamiliar with tis cuz I'm used to simply right or wrong code

11:06 AM
@DialFrost There is a broad spectrum of code between right and wong

@samcarter starting with me, and ending with egreg?

@samcarter I just can't understand how something can be produced when there are errors (I always thought errors cause things to stop?)
@DavidCarlisle Hehe, good representation :D

@DavidCarlisle of course, but which end you are at will depend on your lunch plans :P

@JosephWright Interesting book, 1 problem, it's like impossible to find anyting using the find command since the pdf is so loooooong ;P
@DavidCarlisle But how can the code work if there is an error? E.g. if the error is someone using the wrong syntax that should stop the code right?

@DialFrost If the error is serious enough, it won't continue. For not that serious errors, latex tries to recover from it to syntax check the rest of the document.

11:14 AM
Hmm, so it "ignores" the part with the error?

@DialFrost Ignores or tries to guess what it should have been, e.g. if math mode is missing
\documentclass{article}
\begin{document}

1_w

\end{document}
@DialFrost ^^^ look for example at the error message from this
! Missing $inserted. <inserted text>$
l.4 1_
w
I've inserted a begin-math/end-math symbol since I think
you left one out. Proceed, with fingers crossed.

Ah! I see now, thanks so much for clarifying!!

11:42 AM
@samcarter Oh, crossing fingers. I do that every time I compile. Didn't know TeX could do it for me. :P

@mickep If tex can do the coffee stains for you, it can certainly also cross fingers :)

@samcarter 🤞

@mickep 🦆

@samcarter Oh, a crossing duck.

@mickep ducks are good at crossings i.pinimg.com/736x/fd/0c/49/…

11:56 AM
@samcarter Hahaha

@UlrikeFischer WHAT DID YOU FEED TO THE BÄR?
4

@samcarter ^^^ the Bär sneaked out of the house and makes a visit in Bonn

@UlrikeFischer :) they look cute together
This could be the inverse plot of en.wikipedia.org/wiki/Honey,_I_Shrunk_the_Kids

2 hours later…
1:59 PM

@DavidCarlisle Oh, fighting the headings. The question here is also, why did he/she put the headings in bold when OP did not command it?

@DavidCarlisle Thank you!

@DavidCarlisle Thanks for notifying this person. Maybe this styling is used on other SE sites but it really does not contribute much to most posts

2:15 PM
@DialFrost I had to be out for a while but I meant this, Overleaf (or an offline editor) can only mark all error lines like this if it does not stop on first error.

@DavidCarlisle all these troubles for counting the errors and then show the number so small :P

@samcarter I could sell @yo' my blue arrow error design

@DavidCarlisle I'm not sure they can afford such elaborate artwork!

@samcarter they could arrange a whip-round, \$1 per registered user should cover it
2

2:21 PM

@samcarter longtable in the kernel by default

@PauloCereda together with sillypages? Great idea :)

2:58 PM
TIL that this is invalid syntax: \usepackage[margin = {1cm, 2cm}]{geometry} (unless I do something wrong...)

3:08 PM
@yo' hm, more braces help. @DavidCarlisle do we need to change the keyval handler?

@UlrikeFischer well, we haven't faced this in our testing I think, which makes my finding even more surprising. (Cf. our "users do all sorts of stuff" discussion before.)

@UlrikeFischer we havent changed geometry have we?
@UlrikeFischer just remove the space after = @yo'

@DavidCarlisle I know how to solve it, I was just surprised, nothing else :) btw I think it's been failing in TL2021 as well.

@yo' failing forever, blame @Skillmon for \zap@space we could make geometry use the new kv option handler

@DavidCarlisle blame is already too long so you shorten it to "blm"? :)

3:33 PM
@DavidCarlisle ah right, then it is not new ;-)

3:59 PM
@samcarter Did anything interesting happen out of this?

@DavidCarlisle -- I've forgotten (because I never run in scrollmode): If a job is canceled on the first error, is the (incomplete) .aux file written out, or not? Incomplete .aux files invariably cause confusing errors in the next run. (I think the answer is no, but would like to confirm.)

@user202729 You can find an archive of the puzzles at topanswers.xyz/tex?q=2059 It was fun + I learnt a lot from the helpful code reviews
(and I still hate opening { in the next line :) )

@DialFrost -- Where did you find a pdf file of Lamport. As far as I know (and believe), the book is supposed to be available only on paper, so this sounds like a copyright violation.

@barbarabeeton scrollmode you can't stop on first error, In the default error stop mode if you exit cleanly with x the the aux is written, but people just kill the job eg kill the operating system window, in which case the aux is usually corrupted

4:16 PM
@DavidCarlisle -- That's why I never run in scrollmode. I know how to correct (some) errors on the fly, and make heavy use of that facility.

Does anyone know of any alternative documentation/tutorials on using make4ht? The main documentation is not particularly beginner friendly.

@DavidCarlisle wait, that macro was your idea! I'd never used something like it! :)

4:33 PM
@AlanMunn yes
@Skillmon any reason not to blame you?

@UlrikeFischer That's the one that IMO isn't terribly beginner friendly... :(

@AlanMunn ah I was trying to find something, but I think that was it:-)

@DavidCarlisle Oh I thought you were being your usual "I don't do pragmatics" self. :)

@AlanMunn simple solution, first read the latexml manual, then come back to that with a new appreciation of how friendly it is
@AlanMunn it did occur to me that you would think that when my firefox history failed to suggest the place immediately

4:43 PM
@DavidCarlisle ooh

5:14 PM
@PauloCereda dinner

@DavidCarlisle oh no

5:39 PM
@UlrikeFischer May I ask with which software you check how the PDF ist constructed? More concretely: How did you find that large negative kerning in the context of the "mysteriously disappearing text" question? I opened the PDF using Illustrator and there, you can "see" that there are elements far to the left by selecting everything (but they are too far left to be selected individually)

@JasperHabicht I uncompressed the pdf (with qpdf) and then looked into the file with an editor.

@UlrikeFischer Thank you! I wonder, someone posted a preable ... so maybe it could be possible to generate random PDFs using it and test for such bad things. In the end, what we need to get this error is an MWE, but it seems that nobody is able to construct one.
Just an idée fixe =)

@DavidCarlisle yes, I'm the one responsible for the well behaving fast key=value solution, you're the one to blame for the strange features of the reference implementation!

@Skillmon allowances that you can't spell "impressive"

@DavidCarlisle also you completely mispredicted the future when you thought options will only be used for single keyword-like things.

5:55 PM
@Skillmon blame Eberhard Mattes

2 hours later…
7:59 PM
hi, folks. how does the "e" and "E" arguments function in xparse/ltcmd syntax work? I've read xparse's docs, the ltcmd docs, and like 5 or 6 Q&A here, but still have 0 idea of what they do or how they're usually used (what use case would be an "aha, that's a job for e/E")
(interestingly, in of those Q&A's, someone asks something similar and somebody comments "duh, just read the docs"
so sometimes people "forget" to consider the possibility that the asker HAS read the docs, it's just that the docs... don't doc very well)

@DanielDiniz easiest way to understand them: They can replace TeX-typical super- and subscript notation. So an embellishment defined with e{_^} will read do you and understand? in both of the following: \foo_{do you}^{understand?} and \foo^{understand?}_{do you}.
(also that's maybe the one fine usage example)

And even that is somewhat questionable since the syntax is not quite the same as TeX's normal subscript/superscript.

8:16 PM
mhm, I see. yes, from what I've read it seems they're always used with _^... thank you, your example helped
have you ever used them, or have ever seen them used, with things other than _^?

@DanielDiniz not that I remember, but my brain's memory compression algorithm tends to be lossy unfortunately.
@MarcelKrüger yep, especially with weird font-switch macros in the way :P
(I'm not sure whether this still works, but I think there was a point in time in which x_\mathrm{y} was doing the same as x_{\mathrm{y}}, though the former is absolutely by chance and not supported LaTeX syntax)

I was thinking about a macro like "enclose the argument in (...) , or in some other delimiter you give me"
\NewDocumentCommand{\bracketit}{ E{l}{\lparen} m E{r}{\rparen} }{%
#1{#2}#3%
}
so it "works", I suppose: \bracketit l! foobar r! ==gives==> !foobar!

(the initial goal wasn't that, was to make a macro to keep me from forgetting closing )'s in logic formulas. because they usually use ()'s, but might also use []'s and {}'s, I'd rather have some way to manually specify the delimiters)
oops, actually i'd be \bracketit l! {foobar} r!

8:35 PM
@DanielDiniz It works, but interfaces like that really aren't a great idea. They don't fit normal expectations of LaTeX users, lead to code which is impossible to read if you don't know exactly how the local macros are defined and lead to odd side effects (imagine explaining someone why their code breaks if they refer to a variable r directly after parentheses...)
For many of the less common xparse specifiers the rather minimal documentation should indicate that if you don't already know that you have to implement exactly this interface (e.g. when replicating legacy interfaces) you should probably never use them.

yeah, it does look a bit silly/bad in hindsight... oh well. well, if anything i think i now understand a bit what e/E does

@Skillmon well "chance" defined as "deliberate decision to avoid annoying too many users" but yes.
@DanielDiniz why not \DeclaredPairedDelimiter from mathtools ?

@DavidCarlisle yeah it's probably a better idea, I already use something similar for writing sets, to keep me from forgetting the closing brace

8:54 PM
@DanielDiniz Reminds me of a previous life with non well founded set theory:-)

1 hour later…
10:07 PM
In unicode-math, is it correcct that \divslash and \fracslash are of class relation but \slash is not?

10:23 PM
@mickep \mathslash hmm unicode and mathml have U+002F as binrel, but I think TeX has always had it as mathord so 1/2 and H/G get tight space
mml3 had spacing as 1mu. mml4 draft changed it to 4 I wonder if that's intended, might be difficult to change at this point.

@DavidCarlisle ok, I think Lucida's fracslash has too thin bounding box. So there it is good that it is a binrel. But one could ask which class makes most sense for them in general.

@mickep I'm not sure it's a real thing Murray uses it as a magic character in Word's "linear syntax" to trigger \over a built-up fraction, so it doesn't actually typeset it at all (I think, I know nothing about Word really)
@mickep oh / makes \over fraction / makes \xfrac a/b unicode.org/notes/tn28/UTN28-PlainTextMath-v3.1.pdf

10:40 PM
@DavidCarlisle Ok, thanks. It is a bit funny that there are several different slashes and that their usage is not clear.

@mickep "standard Unicode weirdness" chat.stackexchange.com/…

@DavidCarlisle It is strange sometimes. Anyways, thanks for the input!

11:20 PM
@mickep -- Is there really a just plain \slash in unicode-math? I think of \slash as a text slash after which it is permissible to break a line.

11:36 PM
@barbarabeeton \mathslash (\slash just has its plain tex def from latex.ltx)

@DavidCarlisle -- Thanks. That's kind of what I thought.

@barbarabeeton I hope you appreciated my informative link a couple of comments up

@DavidCarlisle -- But of course. (I did notice that you're the only one blaming me.)