@Skillmon I was thinking about etl/\tl_act:Nnnnn stuff - I wonder if we could find a name that would work to bring this back (I was thinking \tl_traverse:Nnnnn) - thoughts?

*idea for an package.... it's still early here.

@JosephWright What's the first N-type supposed to be, the tl-var to be traversed? I'd say most of the time you want to traverse an n-type list, and (I know this is not according to usual argument order) currying that one is the more likely use case to build a highlevel function atop of the act-loop. Also from my experience using etl's loop, at least one of either a (changeable) status or functions to change the code-tokens are very handy to implement anything more complicated.

@MaestroGlanz I don't get what the code should do. Note that goto isn't as bad as many programmers want to make it (if used sparsely for the right stuff at least).

@cfr It seems to be the week of buggy fonts...

@JosephWright But etl has also useless stuff, I'd argue. The pre-filled output argument is useless I'd say, and if I'd redo the package today I would drop that one (I simply can't think of a use case for that).

@Skillmon I wasn't suggesting adding everything :)

@Skillmon No issue with n rather than N type, though normally we do both - and also no issue currying as long at the top level is in the std pattern

@Skillmon status I agree useful

@Skillmon It is defined to itself. Like a recursion. This redefinition will make it survive any edef on the way. And as long as it is not typeset within begin-document, it doesn't matter, what the current definition is. It is of course better to not built fragile stuff in the first place.

@JosephWright if the top level is in the standard form you can't curry :P What you'd usuall do would be something like \cs_new:Npn \my_replace_thingy:n { \tl_traverse:nnnn { \__my_replace_thingy:nN {abc} } { ~ } { \__my_replace_thingy_group:nn {abc} } } and curry the last argument.

@MaestroGlanz If that's your goal simply do \let\mystorage\mymacro and \let\mymacro\relax, no need for the strange \unexpanded{<whatever>} construct. Alternatively, take a look at the definition of \ekvpProtect in expkv-pop.tex.

@mickep indeed, thought I forget what the other one was, even though I remember seeing it. that's why I don't want to use luatex. well, that and speed.

@Skillmon For one expansion step, I think consistency is more important - for cases where absolute maximum speed is needed, it's almost always going to come down to coding by hand anyway

@Skillmon Anyway, jusst a thought

@mickep I wonder if the paid version has the same bug or if they keep it for the free ones.

@mickep or did you mean Cambria?

I would like this room to offer a glossary ...

@MaestroGlanz fragile commands in the preamble are rarely a problem. Who should do an edef or a write there? And I don't understand your remark about package clashes. If two packages define the same command there is always a clash as only one can win, and you will to decide which one. Such a decision can not be done automatically.

@MaestroGlanz \relax isn't expandable by any expansion whatsoever, when it reaches TeX's stomach it simply does nothing.

@Skillmon I assume, I understood it correctly then.

Another day, another siunitx bug fix

Lots of edge case bugs coming up

@cfr I had the Cambria issue in mind, yes.

@MaestroGlanz anyway the \ekvpProtect mechanism is different. It also doesn't expand inside e or x expansion, but does in f expansion and doesn't change the behaviour when it reaches TeX's stomach.

@Skillmon Sounds like \protected to me :)

@JosephWright yes, but it has the syntax \ekvpProtect{<stuff>}

@JosephWright Internally it uses two macros, one is \protected and expands to #2, the other expands to \unexpanded{\self{#1}}.

@Skillmon Ah, so a hybrid - Frank would probably like that :)

(If we were starting from scratch ...)

@JosephWright If we were starting from scratch, there'd be much that would be done differently nowadays. Maybe we should think about releasing a LaTeX3 :P

@Skillmon Indeed :)

I have a feeling we're answering the same questions of the same user over and over, just always in a slightly different context. It doesn't look like the user is interested in learning anything, just in having one random problem after the other fixed.

@Skillmon Yes - they seem to be flailing around, not really explaining what they are actually up to (seems to be something I likely would not do in TeX)

@JosephWright exactly, but as I said, having the same issues over and over, just in slightly different contexts of code. I find this rather exhausting.

@Skillmon You are free to vote to close, that's what it's there for

@JosephWright but it's not an absolute duplicate it's more an "we explained this aspect of the L3 language thrice, maybe pick up what we said and apply this on this new situation". I don't feel confident voting to close.

@Skillmon Are you talking about the parser thing? I did not follow closely, but at least they seem tro try to break this down into simple (different) questions ...

@JasperHabicht But not the example ...

@cfr I was thinking of just saving thr otfs as .pfbs from fontforge, and generating tfms from otftotfm. I should try autoinst

@DavidCarlisle oh, so it's Adobe's fault. I once opened a postscript file on okular, and the bitmapped CM was quite thick and beautiful in it.

Btw sorry to reply late

@Skillmon Yeah, I see the point

@JosephWright Yes, that's true ... =)

Anyone know of CLI tools (linux) for removing/overwriting the Author field in a PDF file? I was playing with mat2, but that messes with the text in the PDF so it cannot be retrieved with pdftotext. exiftool -all= file.pdf seems to work, but clears all fields. Your like to only remove Author (for now).

@daleif Can you overwrite the author with something like exiftool -author= document.pdf?

@yo' i did find (and tweek) an emacs mode for expl3 syntax highlighting, this way it is barely bearable.

@samcarter according to the documentation exiftool.org/TagNames/PDF.html this attach new values but the old one are still in the file, to get rid of them you need then to linearize the pdf.

@UlrikeFischer Oh! Something new learnt today! Thanks!

Is it possible to install a day-specific "version" of texlive with tlmgr?

@samcarter Ahh, that seems to work. Its options are a bit much to read.

@Lupino don't think "day specific" is a defined term

@daleif i noted some differences between my render and one on a customer's server which differ by about three months

so i wonder if it is possible to say "install texlive 2024 from May 2024"

rather than the "newest" texlive2024

@Lupino No.

@Lupino yes :) islandoftex.gitlab.io/community/projects/docker

but you can probably get close.

@daleif consider the comment by @UlrikeFischer if you want the information to be completely gone

@Lupino you can also just copy from the installed version, modulo differences in binaries. that way you don't get differences due to e.g. mirrors syncing at different times etc.

@cfr that's what i'm doing right now, but i wonder if there is a more convenient way

@Lupino I don't think there's a more convenient fully accurate way. but, then, I don't find that way so inconvenient ;). but if you are unlucky and e.g. latex was patched that day, you could get significant differences even if you get a daily snapshot.

@cfr which is exactly the reason why we keep older versions and don't do daily updates…

@DavidCarlisle btw, I found out that the book weight of NewCM has been created by a feature in fontographer. But some some glyphs have recieved some modifications by hand

@Lupino if only you used a decent editor like VIM :)

@Skillmon weiche, Dämon!

@Lupino :P

@ApoorvPotnis autoinst is a wrapper around otftotfm et al. I don't like the wrapper, personally, but it is useful if you don't want to spend a lot of time tangling with the intricacies of fonts.

how do you spell "TikZ"? "tik-zed" or "tikts"?

@Lupino tiks

(or maybe ticks)

In other news: Today I was faster merging a 183 files changed Refactor PR during dayjob than my colleague who changed some of the same files. So I don't have to resolve the conflicts but he does :)

@mickep by the way, my username is cfr and not crf. I mention this only because I noticed by accident you tried to ping me and did not succeed.

@samcarter thanks. I'm this case I think we'll be fine with "gone unless you start to dig"

@cfr oh, how clumsy of me

@Skillmon speedy rabbit!

@samcarter 'twas his own fault, he was my second approver :)

@Skillmon haha!

@ApoorvPotnis oh yes, early acrobat were simplly awful with bitmap fonts.

@MaestroGlanz if you want something to survive an edef, can't you just make it \protected ? and did you mean \definition not definition here? \def\definition{\unexpanded{unexpanded{definition}}}

@Lupino possibly unfair comment but you spell it TikZ I pronouce it as tics (no zed (as that's american) and no t because there's no t

@DavidCarlisle Yes, you are right. One backslash missing.

@MaestroGlanz then the definition is really weird, you are defining it so it survives exactly two expansions? why?

@DavidCarlisle But you pronounce the Z as voiceless s I assume ... applying Auslautverhärtung (terminal devoicing)

(But it is very difficult to pronounce k and then voiced s ... at least for me)

(As difficult as it is to pronounce [kts] ... so [ks] it is)

@JasperHabicht I pronounce it the same as tics or ticks

@DavidCarlisle Yes

Is it possible for someone other than the original poster to choose the "Best answer"? I'm asking because I think someone undeservedly gave me that mark on a question where there was a way better answer. I thought that it's (probably) quite common that those who ask questions don't have the best judgement as to what answer really is the best, and thus I thought that it's likely that someone with sufficient points has the ability to also re-allocate the "Best answer" mark?

@Atex no, but the assignment of votes and the assignment of accepted (not best) answer ticks is entirely arbitrary. The tick does not mark the best (that is more or less supposed to be shown by the highest vote so a community decision) the tick is "accepted answer" so a personal decision of the person who asked the question.

@DavidCarlisle As a consequence of the OP choosing my answer as the "accepted answer", I also think that's why it has become the most upvoted, however (because you have to scroll more to get to the others). Can I post a link to the answer in here which deserves the most upvotes so that we can make it known that my solution is not the best?

@Atex just let it go, it happens all the time. If @egreg worried about all the times he got a tick when I'd posted the best answer he'd never sleep at night.

@DavidCarlisle I guess it's something you get used to by experience. I just think about the effort the guy who actually deserves the most upvotes put, just for it to be overshadowed by a less good solution. New users who look at the answers would be mislead and deceived into thinking mine was the best, unfortunately. At least experts can appreciate the effort of the good solutions, even if they don't get their deserved recognition, however:)

@Atex write a text at the begin of your answer that you think the other answer is better. But do not worry about ticks or votes.

@UlrikeFischer Excellent suggestion, I'll do that right away! Good habit to build up when you can't guarantee that the best answer gets its credit

@Atex I agree with @UlrikeFischer. But it's also possible to add a bounty to an answer after the fact to reward it. This also will raise its general visibility, since the bounty remains visible on the answer after it's awarded.

@Atex after a while you realise that the assignment of votes is inversely proportional to the effort put into the answer look at tex.stackexchange.com/users/1090/… the highest ranking posts have answers that took less than a minute to write. I have some answers that took a week or more and never got more than a couple of votes

@DavidCarlisle Picture mode is so unappreciated.

@AlanMunn true

@AlanMunn Nicee, so that's like a permanent marker that won't get lost, right? For the solution in question, I followed @UlrikeFischer's suggestion, but I'll also keep this alternative in mind if this happens again. That way, you can also put a kind of "emphasis" depending on how big of a bounty you award:)

@DavidCarlisle one of the highest ranking answers on the network is how to exit VIM, even though that's super simple because VIM is so intuitive and everybody knows how.

@DavidCarlisle oh wow, I looked at the contrast between your highest voted answers (1st page) and your lowest voted answers (last page, or 491st page), and it's astounding - Someone even gave an answer of yours a downvote! I guess it makes sense in a way that the "least effort" (typically) gets more votes than the complex solutions which requires hard solutions. It's because mostt people (myself included) can't comprehend the brilliancy of the complex solutions, because most of us simply don't...

... have the knowledge to fathom and appreciate the good, complex answers which required a lot of effort to come up with. On the contrary, a simple solution is usually intuitive and easy to understand the logic behind it, thus being more popular, typically speaking

@Atex There's also the "If it comes with a picture, it gets more votes" effect.

@AlanMunn Good point, that I can personally relate to. Sometimes, you're just lazy and don't want to bother yourself with running the code to see the solution - Especially when you're scrolling on your phone. Therefore, you'll be more inclined to upvote an answer with a picture, because then you can see the code's results immediatedly and assess it

@AlanMunn and if it's TikZ it comes with an additional factor 2 of votes.

@Skillmon It's interesting how many factors actually play into upvoting answers. People like colorful, visually beautiful solutions, and they're typically drawn using TikZ. If the code to that is simple and comprehensible along with a picture to display the code, it's a recipe for upvotes. On the contrary, there may be a solution to something which is really impressive, but because it looks "boring" it doesn't gather as much attention.

@Lupino which mode? And why is it only "barely bearable"? I'm only asking because I'd like to know what people are after for L3 modes

@Atex look at my most upvoted answer. As I said, TikZ is a factor 2 (at least).

@Atex you need to get (or give) some downvotes as the main aim of the site is to gather palindromes. I have hit 111111, 222222, 333333, 444444, 555555, 666666, and 777777. Without asking a question and without downvotes I'd have been limited to 0 mod 5 so only one of those would have beeen available,

@AlanMunn It is much nicer if there is also an image of the output, so that one does not have to copy and run it to see.

@mickep Yes, I was more thinking of the TikZ picture effect, rather than just showing the output.

@AlanMunn Ah ok, I think it is general. And should be.

@DavidCarlisle don't new users start of with 1 reputation?

@Skillmon Not if they come from another site with rep. Then they start at 101.

@AlanMunn doesn't change the fact it's a 1-6-cycle, so 2 of those would've been possible :)

@Skillmon Yes, but you start off with a palindrome. :)

@Skillmon Oh wow, that ratio is insane! But that animation is also very enticing, and so I can see why people were drawn to it. I also looked through your whole first page, and TikZ seems to be the most recurring topic in your 1st page of top answers as well. I saw some interesting things, too - Perhaps I should bookmark your profile;) (I don't know if that's actually possible to do)

@Atex the funniest thing is that I don't even know that much TikZ and (comparatively) rarely answer those questions

@Skillmon I stand corrected

@Skillmon you could always ask me, I have a gold badge for tikz-pgf

@DavidCarlisle with a single picture-mode answer?

@Skillmon well a few: planes, donkeys, pigs, ...

@DavidCarlisle :) Of course (after all you need more than a single answer for a gold badge)

@Skillmon really goes to show the power of the TikZ effect, doesn't it?;) Your 2nd most upvoted TikZ answer was just adding the \textcolor command to a node in order to draw 2-colored text in a node. Simple, yet it leaves a stunning, visual effect:)

@Atex exactly. That's a combination of the "it took only a minute" and the TikZ effect. Somehow strange that it's not my most upvoted answer.

@Skillmon Yup. On the contrary, I've seen you post some super advanced and amazing answers in your area of expertise (which is defining commands from scratch, based on my observations), and they, unfortunately, don't come close with upvotes in comparison. A really funny phenomenon

@Atex that's to be expected. Though they (often) require much time they're only immediately useful for a very narrow public (those who need such a macro -- that's often just a single person -- and everyone who happens to be on the site when I publish/edit the answer) and often don't get any "late" votes. Answers like the multi coloured TikZ-text one still get votes from time to time because new LaTeX users discovering TikZ hit the same problem.

@Atex I have no explanation for the aperture answer's popularity, except it looking fancy. I doubt it's of use to many later visitors (how often can somebody need to draw an aperture?).

@Skillmon Yeah, it makes sense, though. Simple answers are often easily accessible/understood and are typically related to a more common problem, thus gathering more attention/upvotes, especially if they're TikZ, like we discussed earlier. Even if people don't need that colorful/fancy looking code themselves, it's safe to say that people still admire the visual output of TikZ and its potential. I know I have so many answers saved that I'll most likely never use, but I've just been amazed by!