@PhelypeOleinik It's been a long time since I used Pascal, but I think it's a more compact way of storing things with at the cost of possibly slower access to the individual items.
@Skillmonlikestopanswers.xyz Hm, from a fast assessment it seems alpha_file is declared as a struct thingy, which looks quite like a Matlab's struct type, which I understand. So yay, that's a start!
@Skillmonlikestopanswers.xyz Looking at that it doesn't seem directly comparable (but I don't know C). packed in Pascal can apply to elements of type array, file, record and set. But my knowledge is seriously faded.
@AlanMunn but how should Pascal further pack an array containing only elements of the same type? (or can Pascal store elements of different types into a single array?)
Hi. I revised this question yesterday. If there is no good answer, then I could close or delete it. Or someone could post an answer saying there is no good answer.
My OS is Debian GNU/Linux 10.2 (buster). TeX Live is a backport from unstable, version 2019.20190605.51237-3 (a snapshot of TeX Live 2019).
The code below is an LuaTeX form letter implementation. Without the comment environment lines, it compiles without problems. With the comment environment, i...
@UlrikeFischer @barbarabeeton @Skillmonlikestopanswers.xyz So I found the answer to the question about the calls. In short: We copy your project to the /compile folder and call basically latexmk -cd -f -jobname=output -auxdir=/compile -outdir=/compile -synctex=1 -interaction=batchmode -pdf /compile/a/b.tex when a/b.tex is the file to be compiled. During this, latexmk adds /compile to the front of TEXINPUTS (because of the -auxdir option).
If you want more details, I can provide them (though it's gonna be to lengthy for the chat I'm afraid).
@UlrikeFischer So back to your original question: It will be failing in standard linux, because TEXINPUTS won't be set in a way that would make it work.
@yo' no thank you I don't need more details, once I realized the texinputs part I also understood why the project compiled only on overleaf. Now I only have to decide how at best to get it working locally too.
The whole issue is that locally on your computer, you would either do cd /compile/a; pdflatex b.tex or cd /compile; pdflatex a/b.tex and we want to accomodate for both options. And this makes things very complicated :)
anyway, I gotta go, might be back in ~1hr.
@UlrikeFischer neither do I, and it does have some pecularities.
@yo' the project has a lot of chapters and they try to setup things (with subfiles and similar tweaks) so that you can compile both the main file and chapter files. That sort of thing is always quite challenging to setup.
@UlrikeFischer Challenge: Make an animated gif that zooms in forever. (Where “forever” means looping around to the beginning, of course.) You'll need the innermost tikzling to have a white shirt, I guess, to allow seamless looping back to the beginning.
@JosephWright @UlrikeFischer new bidi on ctan (I used the bash script, getting l3build to work would take more re-organisation than I had time for yesterday) (the docstrip arrangement is "interesting")
@yo' I think that every single word in Norwegian containing a C is a (relatively recent) loan word from a different language. I suspect German is somewhat similar in that regard.
@UlrikeFischer (sorry I missed this message) yes, subfiles is very popular at Overleaf, and yes, due to the TEXINPUTS frenzy, it behaves different at OL and offline.
@HaraldHanche-Olsen words with c are normally load words, but not not necessarly recent, we got quite a number from latin and french and quite a lot which have now a k had previously a c.
@UlrikeFischer Okay. And I also forgot about the sch combination, found in your name even – corresponding to sh in English – and to sj or skj in Norwegian
@HaraldHanche-Olsen no surprise. (Funnily, most Czech words starting in C come from German, Hungarian or Gypsy. One such example is "cálovat" -- to pay)
@HaraldHanche-Olsen that's CH tho; it has it's own sound [x] that has nothing to do with [c], [s] or [k].
@DavidCarlisle We are begin pretty strict with master so that really shouldn't happen; I went with a release/2020-02-02 branch to stage, but won't push that one
@JosephWright ah I guess I'd have merged all of devel to master then cleaned up version strings there, I don't trust that big gitkraken display where if you catch your mouse things move around
@Skillmonlikestopanswers.xyz well yes you could but I wouldn't to my mum, so it depends on the context. In a math document I'd probably write $x\in[a,b]$
@DavidCarlisle well you can use "deutsch" as an adjective (like "the German car", which would be "das deutsche Auto") or as a noun (like "David speaks perfect German", which would be "David spricht perfektes Deutsch").
@DavidCarlisle that's why translators might be tripping over it.
@barbarabeeton oh my, deepy sorry! My bad for not following the reply arrow! I found the image randomly on the internet. I believe it's fake, but still...
@Werner Sorry the delay, power went down. The problem seems to be that tocindent2 gets lost somewhere in the document. Not really sure when. Though for some (even more odd) reason, adding a \subsubsection solves the problem...
@DavidCarlisle -- Ah. Well, there came a point when I was cut out of the department information channels, so it became rather difficult to keep up with what was going on. Somebody else's problem now.
@barbarabeeton hopefully by the time that documents reach the AMS they do not have section headings with no text following at the end of the document, so you are unlikely to ever see that in production
@Werner the class does \AtEndDocument{\@writetocindents} but if there is a pending inline heading they get written before the heading does its internal labels and things are out of sync, so it should do \AtEndDocument{\if@noskipsec\mbox{}\fi\@writetocindents} to make sure things are flushed out
@DavidCarlisle -- And the \@writetocindents has never worked properly. It's basically a reasonable idea, but Michael Downes never got around to making it do what was intended. The whole TOC design was put on hold "waiting for specs", which were never provided. That was in the mid-1990s.
@DavidCarlisle -- But thanks for the diagnosis. If @Werner can pass it along, that could be very helpful.
While the issue is minor, and only relates to "incomplete sectional units", a class should take care of such things to avoid the end user feeling like something is amiss.
“MediaWiki2LaTeX can put different pages from a Wikimedia wiki into a PDF. It can now make a PDF with around 5000 pages. Previously this was 800 pages.” Cf. MediaWiki2LaTeX.
@DavidCarlisle if you compile a minimal document with lualatex, where is the font names database loaded from? (should be in the log: luaotfload | db : Font names database loaded ...)
@Sigur why would you box things so many times? Wouldn't something like \vtop to 6cm{\hsize\dimexpr\textwidth-2\fboxsep-2\fboxrule\relax #1} do the trick?
What you might use is \vbox to 6cm{#1\par\vfil}. Without the fil, you'll get underfull boxes. If you place the fil before #1, it will be bottom-flushed. If you place it in both places, it will be middled.
@yo', nothing wrong. I want the framed box with width equal to \textwidth and inside \item it will produce overfull (I think). So this is why I used figure.
@Sigur well, that doesn't really look like a list. Wouldn't a Theorem-like environment look neater? With \usepackage{amsthm}\theoremstyle{definition}\newtheorem{problem}{Problem}? It would be more prominent where the Question/Problem/Task/... starts, with the bold title, and you wouldn't have to solve this mess. Also, note that [h] floats can float away, causing a mess.
Last but not least, if you have such a thing, you probably want \raggedbottom, and actually don't put a new paragraph before the box, but rather \\* (to make a new line and avoid linebreaks at the place)
