7:41 AM
@CarLaTeX And such a delicious one!

7:57 AM
@HaraldHanche-Olsen :)

8:39 AM
Someone said on #lua (freenode), that he/she thought that LuaTeX support for 5.3 in TeX Live 2019 was "experimental". Comments?
Bottom line. I was thinking of upgrading to the TeX Live 2019 binaries in Debian experimental, in part because of Lua 5.3, but I'm not sure it's worth it if the support is experimental.

@FaheemMitha No, they moved from 5.2 to 5.3 for TL'19
@FaheemMitha Experiments were in TL'18

@JosephWright So support is "stable"? Not sure if that is the right word.

@FaheemMitha Yes

@JosephWright Ok.
The experimental packages are 2019.20190507.51032-1
So I guess that's quite recent.

8:58 AM
How can you change the color cycle of bar graphs using pgfplots? I know you can set the cycle to different names, i.e. exotic or black white, but I tried this and only the border, not the area inside, changes color, so I suspect the cycle list only applies to XY plots; or is there something I can add to the code? Or should define the colors manually?

9:22 AM
@MaxR fill = <whatever>

It seems that Lua doesn't have a problem breaking compatability. I'm not sure what the implications of this are for use with LuaTeX.
One nice thing about TeX is that it's very stable. I was looking at a LaTeX file from 1999, and it almost compiled on my machine. It didn't, not because of language issues, but because of missing files.
It was 2.09, though. But it seems that TeX Live in 2018/2019 still supports 2.09.

@FaheemMitha I would not say 'no problem': they made a change for good reason

@JosephWright A change? I didn't have anything specific in mind.
It seems they change things at every (point) release.

@FaheemMitha LuaTeX is more-or-less frozen at 1.10: Hans et al. have a new engine purely for ConTeXt (LuaMetaTeX)

I don't doubt they had their reasons.

9:32 AM
@FaheemMitha That's what a release means .. change
@FaheemMitha Handling of integers

But doesn't that mean that LuaTeX today may not work in 10 years?

@JosephWright Is that the only major change?

@FaheemMitha Well, there is a question of who continues to support the engine, but that is also true of pdfTeX and XeTeX, really: (u)pTeX does have a core set of active devs
@FaheemMitha I think that as LuaTeX is written in C(WEB) it's actually much easier to find 'help' than for XeTeX/pdfTeX (speaking as someone who's added primitives to both)
@FaheemMitha Like I say, likelihood is that there will be few major changes in LuaTeX - more like pdfTeX in that sense

@JosephWright It's written in CWEB? You mean Knuth's literate programming thing?

9:42 AM
@FaheemMitha Something like that, anyway: I think the source files may be 'normal' C but with WEB-style comments ...
@FaheemMitha It's certainly C not Pascal

@JosephWright I see.

@JosephWright hm

@PauloCereda Have you read pdftex.web?

@JosephWright I glanced it. :)
@Skillmon hi mr. rabbit!
Today is a very cold day: 10C!
Brrrrrrrrrrrr

10:05 AM
@PauloCereda LOL, we are having a May which seems an October

@CarLaTeX is this a good thing? :)

@PauloCereda No, too cold

@CarLaTeX oh no

@PauloCereda On the other hand, @DavidCarlisle won't have a May anymore...

@CarLaTeX I see what you did there. :)
@CarLaTeX don't make fun of them, they might end up getting Bamm-Bamm from the Flintstones as prime minister...

10:08 AM
@PauloCereda hi mr. duck!

@PauloCereda And she said what Barlusconi said in 1994 "this is the Country I love". They can't make fun of Italians anymore :)

@CarLaTeX sadly, pineapple pizza is still a thing. :)

@PauloCereda We're working on it, too :)

@PauloCereda pineapple pizza? Is that a problem?

10:11 AM
@PauloCereda We need the army :)

@FaheemMitha for me? not at all, I have Nutella pizza once in a while. :)
@CarLaTeX ^^

@CarLaTeX LOL this is mean
@CarLaTeX we could celebrate Claudio's 80th anniversary with some pineapple pizza. :)

I think the blokes from Torino are not as explosive as the ones from Naples. :)
@CarLaTeX ^^

10:14 AM
@PauloCereda I think Claudio wouldn't appreciate!

@CarLaTeX nope :)

@PauloCereda Many people from Naples lives in Torino
@PauloCereda Due to the FIAT (now FCA)
@FaheemMitha Just look the scenes, they are self-explanatory

@CarLaTeX ooh

@CarLaTeX I see a lot of yelling and gesticulating. And a certain amount of pizza throwing. But I thought that's just Italy.

@FaheemMitha An Italian throws a pizza only if there's pineapple on it :)

10:19 AM
@CarLaTeX I see. :-)

2 hours later…
11:58 AM
@JosephWright yes, I'm aware of that command, but is there a possibility of using a color cycle?

12:30 PM
I'm wondering if a macro like this could be rewritten in Lua. This is an actual TeX thing, so possibly not that trivial to do.
\newcounter{tabenum}\setcounter{tabenum}{0}
\newcommand{\nextnuml}[1]{\refstepcounter{tabenum}\thetabenum.\label{#1}}

1:05 PM
@JosephWright I was under the impression that votes from a removed user were not cancelled if the user had cast “many votes”: I lost 510, David lost 300, Werner 330 and Mico 275. That seems a lot of votes.

1:25 PM

1:43 PM
@egreg Down to how the removal was done
@FaheemMitha It could, but would be a bit painful

@JosephWright Ok. Well, it wasn't on my todo list.

@AlexG beamer inserts a \clearpage in \AtEndDocument and then the special is too late. That's quite obviously a problem ...

@UlrikeFischer Yup
@UlrikeFischer I could remove the \clearpage I guess, but doubtless Till had his reasons

@JosephWright lastpage package: \AtEndDocument{...\clearpage\lastpage@putl@bel ...}. All struggling to get the last page right ;-)

@UlrikeFischer Yup: clearly one for kernel-level support

1:58 PM
@JosephWright something that works with harftex: github.com/u-fischer/luaotfload/issues/60

1 hour later…
3:01 PM
@egreg Raise with the Powers

1 hour later…
4:01 PM
@egreg Just saw that this morning as well.

1 hour later…
5:02 PM
@egreg I -245
@egreg I think they aren't cancelled if the user have high reputation, but they are if a low-rep user casted many votes.

5:30 PM
Who needs \usetikzlibrary{decorations.fractals} nowadays...

6:00 PM
@CarLaTeX Yes, but the user had high rep (if I guessed right who).

@egreg Care to drop me a line?
@CarLaTeX Removal for anyone over a trivial rep does has a safeguard for this

@JosephWright We could just talk here.

@FaheemMitha Sure, but then it looses the 'link' to the question you've asked

@FaheemMitha To be honest, I'm not fussed: it's really not a big issue. Here is fine, perhaps @DavidCarlisle or someone might also help out
@FaheemMitha One issue is that, whilst luacode is in some sense convenient, it muddies things as much as illuminates them: I strongly favour sticking to \directlua

6:07 PM
@JosephWright Fair enough.

@FaheemMitha I also would not use luacode, but I honestly do not understand your question at all.
@egreg I seem to have had a lot of user removed rep loss recently
@CarLaTeX do we know who left?

@DavidCarlisle gave you a nice round number ^^

@JosephWright Thank you for the answer. But my point was, that for the simplest input possible, I still get an error. So what goes wrong in the transition from TeX to Lua? Or is just uninteresting technicalities?
Is it that they represent strings differently?

@FaheemMitha I don't know what you mean: the example I've posted in my answer works just fine. Your question doesn't make sense.

And you mentioned \unexpanded, but I don't see where expansion enters into it.
@JosephWright In my original question, if I use \directlua{joinstring(#1, #2)}, I get an error. That's the error I was talking about.

6:19 PM
@FaheemMitha You've not quoted the strings there
@FaheemMitha Lua is a 'normal' language: strings have to be explicitly marked as such

@JosephWright Uh, quoted the strings?

@FaheemMitha you get an error in the corrected example because you didn't used the "" lua needs to see a string: \directlua{joinstring("#1","#2")}.

@FaheemMitha Because \luaescapestring expands it's argument, like I said in my answer

You mean the #1 and #2 arguments?

@FaheemMitha Yes, they are strings

6:20 PM
@UlrikeFischer Oh, is that all?

@FaheemMitha It's no different to if you'd put joinstrings("foo","bar")

But #1 and #2 are already strings.

@FaheemMitha No

Oh, I see, it just appears as ascii text directly?

@FaheemMitha why do you say that?

6:21 PM
@DavidCarlisle I was about to say exactly that
@FaheemMitha Yes, no different from anywhere else in TeX

@JosephWright Ok.

@FaheemMitha as I said here the other day you are just generating lua source code, \directlua{...} is basically like \edef\tmp{...} with the requirement that \tmp gets defined to be syntactically correct lua

@DavidCarlisle No, I don't

Actually, I normally use the [[ and ]] syntax. The " and " give me errors.

@FaheemMitha well either but you need to add them. " will not give errors in in teh normal cases

6:24 PM
@FaheemMitha The only time you'll get an error is if " is in the input, but that can't happen with \luaescapestring in use

@DavidCarlisle Ok. I guess it's simpler than I thought. So \luastringN is just to handle corner cases?

@JosephWright Yes, but not if the user is a low-rep frequent-upvoter

@JosephWright I'll try it with "" right now. One sec.

@CarLaTeX It's complicated ;)

@FaheemMitha as I said I wouldn't use luacode package myself

6:26 PM
@FaheemMitha Well yes, but please see my answer, where I've just written everything out

@DavidCarlisle So what should I use then?
@JosephWright Yes, I see it.

@FaheemMitha I'd just use the primitives

So write out in full:
"\luaescapestring{\unexpanded{#1}}"
?

@FaheemMitha Yes, exactly: this is what \luastringN does

6:27 PM
If that \luastringN is equivalent, why not just use that?

@FaheemMitha perhaps, or if you have control over what goes in #1 (the usual case) just use "#1"

@DavidCarlisle Ok, I will.

@FaheemMitha I'm not keen on 'hiding' the ", as they make the nature of the string unclear

@JosephWright Hmm. Yes, I see. Good point.

@FaheemMitha but given that the code in your original question worked I still have no idea what your actual question was.

6:30 PM
@DavidCarlisle Why I was getting an error when I just did \directlua{joinstring(#1, #2)}.
Ulrike answered that - I just needed to add quotes. Obvious in hindsight.
So that get_macro thing isn't relevant, then?

@FaheemMitha get_macro is for accessing the content of a tex macro from Lua but it doesn't seem very related to your question. (As I said your question was not clear at all)

@DavidCarlisle Yes, you already said my question was unclear.
Someone mentioned in an answer that it was better to use that instead.

@FaheemMitha which makes it hard to say if get_macro is relevant

Also in hindsight, I think I was misunderstanding Marcel's answer. I think he was talking about something else.

0

I wanted to address what was raised in chat today where some reputation changes were made and many of the highest reputation people decreased substantial amounts of reputation (and by substantial - for our site I classify that as between 500 and 1000 reputation points). If you haven't seen this ...

@DavidCarlisle, @egreg, @CarLaTeX ^^^ May be relevant
-1

The help on all sites doesn't cover three things well enough IMO. https://meta.stackexchange.com/help/user-was-removed It doesn't acknowledge that in some cases - community managers will intentionally delete large sums of reputation. It doesn't define what large is. It doesn't point peopl...

6:43 PM
@JosephWright It's a bit odd, but to be honest I don't care, rep comes and goes:)

@DavidCarlisle Oh, neither do I

7:03 PM
\documentclass{article}
\usepackage{filecontents}
\begin{filecontents*}{luafunction2.lua}
function joinstring (s, t)
tex.sprint(tostring(s) .. tostring(t))
end
\end{filecontents*}
\directlua{require "luafunction2.lua"}
\begin{document}

\newcommand\joinstring[2]
{%
\directlua{joinstring(
\luaescapestring{\unexpanded{#1}},
\luaescapestring{\unexpanded{#2}})
}%
}

\joinstring{foo"}{bar}

\end{document}
@Mico ^^^ Still an error :(

@HaraldHanche-Olsen It's nice when things aren't behind a paywall.

@HaraldHanche-Olsen article looks quite good as articles go (after a quick skim over it) the discussion thread goes the way of all these things, tex v roff emacs v vi not very enlightening:-)

Though enormous left margins are distracting and annoying.

@egreg I am surprised you can lose so much reputation. Aren't you reputation capped pretty much all times? Or is there a flaw? I mean a user upvotes some posts here and there, and they increase your reputation. You would have gotten more points on those days, but there is the reputation cap. Then the user gets removed. I would expect that then the points that you lose get charged against those which you didn't get because of the cap, so you shouldn't lose any. But your losses suggest otherwise.

7:13 PM
And ironical, considering the article is about TeX.

@DavidCarlisle Yes, also quite a lot of plain people plus 'DVI is best' ...
@DavidCarlisle I see they talked to FMi

@DavidCarlisle Have you written any articles on TeX?

@marmot It's a big hit

@JosephWright I only very briefly skipped over the discussion (or do you mean the quote from Frank at the end of the article)

@DavidCarlisle Yes, that's what I mean: I only skimmed but did see Frank mentioned

7:15 PM
@JosephWright Well, but given that @egreg seems to exceed the cap usually by a huge margin, this means that the votes must have been substantial at a given day. Isn't that serial voting?

@marmot No: vote loss from user removal all hits on the day they are removed

@marmot I have lost 300, 30, 10 and 50 in the last few days.

@JosephWright Indeed I lost reputation also on Italian.SE and Graphic Design

@JosephWright So let's say one gets 20 points from that user plus 280 from others. After the cap I am at 200. Later the user gets removed. So I lose 20 regardless of the fact that at that day I got 280 points without this user?

7:19 PM
@marmot Yes
@marmot It would otherwise require recalculating all rep for all affected users, which would be heavier on the back-end, but also might reveal more about who the removed user was

@JosephWright I see. That's what I suspected. So the impact of the a user removal will depend on their timezone. It is less annoying if you remove a user from the US. (Makes sense...)

@marmot That will also depend on the site: for TeX-sx, we have mainly European users, but say the main site (StackOverflow) will be US-heavy

@JosephWright Are you saying this because this is the official statement? It seems that the powers have left more than enough loopholes that allow us to figure out who the user was, who just downvoted me and so on.

@DavidCarlisle i lost 200 and 15 because @egreg stole my tick but I'm most miffed up that someone downvoted an old answer and now I'm no longer in the 2/7 circle ;-(.

@marmot We are a small community when you look at the 'active' users. The rules are driven by StackOverflow, and there it's more-or-less impossible to track these things
@marmot Serial behaviour does tend to show up, I agree: partly it depends how much you want to know

7:25 PM
@JosephWright What I meant to say: consider user A who votes in the first 8 hours of a day (with the time zone being the one in which our reputation gets binned) and user B voting in the last 8. So if user A upvotes, say, @egreg a lot and gets removed, this will decrease his score, but not user B because their votes were cast when @egreg already hit the limit.

@marmot Yes, that's fair (I think ... the details of the back-end are not public)
@marmot I've been on the site every day from day one, and I can't say I know all of this stuff!

@JosephWright I guess you have just revealed a means to "ruin" users. Just create a ton of dummies who do nothing but upvoting your "friends" early in the day, and let them get removed after a while. ;-)

@marmot That would be sock voting, and I hope would get picked up and handled manually (we've had a few manual interventions by staff)

7:44 PM
@FaheemMitha The enormous left margin only happens on a fairly wide screen. Without it, the text lines would be too long anyhow. On the ipad, the left margin is pretty much gone, and the pull quotes (in the margin) wind up between paragraphs instead.

@DavidCarlisle Closer to \message ...

@JosephWright details

@DavidCarlisle :)
@DavidCarlisle e-type :)

@JosephWright jaguar

7:54 PM
@HaraldHanche-Olsen Oh. My screen isn't that wide. But whatever.

@JosephWright probably not:)

Anyway, thanks for the link. The HN discussion is actually quite interesting.
BTW, some years ago I tried to add line numbers to math text. It didn't work terribly well. Has that improved any?

@FaheemMitha ???

@DavidCarlisle Sorry. Sleepy. Edited.

@FaheemMitha doesn't lineno do something useful? (I can't say I have ever wanted line numbers personally so haven't tried recently)

7:58 PM
@DavidCarlisle Yes, I think that was the one I used many years ago. And no, it didn't work particularly well.
Just wondered if the technology has improved.

@FaheemMitha possibly although I think that's the usual package to use,although the specialist packages like reledmac can probably do line numbering too, Not something I have ever used.

@DavidCarlisle Ok.

@DavidCarlisle Any thoughts on my mail earlier today

@DavidCarlisle Text is (relatively) easy. It gets harder when there is multiline math involved.

8:39 PM
Hello. In TeXLive I would like to lock the horizontal scroll of the internal pdf viewer. It is locked automatically, when the document is fit to page width, but when I change it to fit to text width (to slightly zoom in), then scrolling through the document is a nightmare. This is almost never an issue on a desktop monitor, but I am using a 13" laptop. Is there a workaround to this?

@Matsmath Do you mean in TeXworks?

Ahm, I am sorry, I am using TeXStudio.

9:01 PM
@marmot I think that the system lists the votes lost because of user removal; the rep is recomputed later, so the actual loss cannot really be determined. Once I got -4360 Serial upvoting reversed, when the real loss was not even in the hundreds.