^^^ Can you please help me reopen this? I actually can answer because I had correspondence with the arXiv team about it.

@HenriMenke I've had correspondence with them as well but that question is still opinion based and rude. Personally I wouldn't vote to reopen unless it was edited

@DavidCarlisle It's rude, but I don't think it is opinion-based.

@DavidCarlisle it is rude, but I also think that the arXiv policy is a real problem. They are forcing (and training) scientist to create pdfs which are much less accessible than the xelatex/lualatex version would be -- even without any tagging -- when other scripts than latin are involved. I wonder why the all the universities mentioned as members do support this.

@UlrikeFischer yes sure there are questions that could be asked, just not like that, and if I was running an archive the size of arxiv I think I'd want some assurance of the long term maintenance of any system that I was using.

@DavidCarlisle yes I know this. But they should also think about long term usability. And if everyone writes in word because there scripts and unicode fonts can be used they can't say either that they support only pdfs created with a 10 years old version.

The "relentlessly ignorant" part isn't ideal phrasing. It's not even clear what that means.

But the main content seems reasonable, imo.

Perhaps someone could work on rephrasing?

@FaheemMitha if you rephrase it so it is polite, it is a question addressed to the arxiv maintainers, and they are not here, so at best you will get an answer from people who have had email correspondence with them in the past, or are prepared to guess. Neither kind of answer makes it suitable for asking here.

@DavidCarlisle Oh yes, I didn't think of that.

Nobody reading this forum can actually do anything about it. Unless they happen to work for arXiv, I suppose.

@HenriMenke even though I think the question was better closed I gave you a +1 for your measured response:-)

quack

@PauloCereda breakfast

@DavidCarlisle oh no

@DavidCarlisle I hope if they ever introduce LuaTeX, first thing they delete all the fonts and ban uploading them. I really don't want to read papers in Comic Sans or Zapfino.

@HenriMenke you should write "mapping to unicode" not utf-8.

@UlrikeFischer Done.

@AlanMunn translate.google.com/…

Catalan is so easy

@HenriMenke I think LuaTeX has a lot of uses in enhancing "traditional" TeX. Some of them scientific.

It's possible that asking a site like arXiv to support this may not be realistic. I don't know.

@MarcelKrüger there is a question on the texlive list about font searching in emacs (I didn't know that this exists ...). It seems to be affected by the new interfaces of luaotfload-tool.

@MarcelKrüger luaotfload-tool --list=basename breaks with a lua-error, while luaotfload-tool --list=basename:* works.

@UlrikeFischer I noticed you were answering emacs questions

@DavidCarlisle oh no

@UlrikeFischer Oops Fixed in dev with 32b5616d.

@DavidCarlisle seems to be easy - one not even need an installation ;-)

@MarcelKrüger which leads me to the question if the luaotfload-tool docu is actually still correct. Are some extension needed here?

@UlrikeFischer I think it should still be correct: * globbing is documented. The other recent change was matching non-sanitized names, but the existing documentation there is sufficiently vague that it can be interpreted either way.

@PauloCereda did you notice the github avatar of the latexonline author?, not sure if its an arara or a Norwegian blue though:-)

@MarcelKrüger I just tried the one example in the docu luaotfload-tool.lua --list="format" --fields="file->base,version" and get as file name lots of "none".

@MarcelKrüger I just understood the description of "no-strip". With --fields="basename,version" it works.

@DavidCarlisle I even can answer for the WSL - which I haven't either ;-) tex.stackexchange.com/questions/542383/…

@yo' Another Overleaf suggestion: make the clear cache files icon stick to the top of the error display. When you have tons of errors, it's a pain to have to scroll all the way down to the bottom to delete the aux files.

@AlanMunn you shouldn't make errors?

@AlanMunn yeah; well, I know about this, but it's another tricky one...

@yo' There's also something quite interesting about how using Overleaf seems to lull new users into thinking their documents work despite dozens of errors. I've been helping one of my students, and they didn't seem to notice that their document had literally 100 errors, because a PDF still showed up.

@DavidCarlisle I don't, of course.

You know, I would prefer halt-on-errors as OLv1 had, then there would be only 1 error and possibly some warnings, but now we have the batchmode and ...

@AlanMunn that's a big topic; yeah. People don't know what the red balloon is :-)

@AlanMunn naturally there is no point having students if you can't blame them for things, or take credit for anything good they do.

@yo' And some things are damn hard to find. Like ḥ in a .bib file entry. Solution: use LuaTeX. :)

@AlanMunn yeah. I would like to be able to see aux files in the left pain (as readonly) and let the errors in bbl point out there. That would help debugging bib+utf8 issues a lot.

But all these things mean quite a large change to the editor, so it's easy to propose them, easy to agree they'd be nice, but more difficult to realize.

@yo' don't you think 600% works better than 60% as a font size?

@DavidCarlisle :D

I would suggest adding:

::after {
  content: " ERRORS";
}

@yo' I was thinking of perhaps a scrolling FIX ME animated text

@DavidCarlisle they could use their ducks carrying a sign "FIX ME" ;-)

@UlrikeFischer That would just encourage people to make errors so they had an easy way to set up target practice.

@yo' Yes, I understand, and I'm certainly not expecting instant changes. But is it helpful or just annoying to hear these suggestions?

@AlanMunn Is your email address available? Or can you write me to tom.hejda at company? I have something to share with you I think...

There's a link in my bio (or texdoc msu-thesis)

@CarLaTeX nice tugboat article!

Not sure if this is a texdoc question or a general Un*x question: If my current working directory is on another volume, not the startup volume, texdoc fails. with the following error: shell-init: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory

@DavidCarlisle Right now I blame them for the final papers I have to grade. :)

@UlrikeFischer Is it already published?

@AlanMunn if students lack knowledge, blame the lecturer

@CarLaTeX yes I got the mail this morning.

@AlanMunn volumes are normally more or less hidden from file system operations, do you have symbolic links? tracing parent directories through symlinks is a bit variable

@UlrikeFischer Good, it's a bit long, I told Karl to let me know if it was too long, no news, so it wasn't :)

@DavidCarlisle I'm not sure what you mean by the question. Links from where to where?

@AlanMunn ah thanks

@AlanMunn mail sent

@AlanMunn from one directory to another: if a directory is linked then it appears to be in two places in the tree so where .. references depends on exactly what you do and can lead to errors along the lines of cannot access parent directories

@AlanMunn google shows one person at least resolved that error thus: We resolved the problem. It was stale NFS file handles on the workstations, apparently left when the server first crashed. We cleared the NFS cache on the server, then did "umount -l" on the workstations and rebooted them. When the workstations came back up, no more error

@DavidCarlisle Hmm. There are no links in the volume I was in itself. My local texmf is symbolically linked to a Dropbox folder though. Other than that, I don't really know.

@DavidCarlisle Well that actually worked. Unmounted the disk, mounted it again and now no error. Strange.

@AlanMunn basic hardware support technique, turn it off turn it on, pretend it never happened.

@DavidCarlisle :) It also coincided with another odd behaviour: when I tried to unmount the disk, it wouldn't let me until I quit the Terminal app. That usually doesn't happen either.

@AlanMunn I certainly can't reproduce this. The closest I get, if I create a new directory, cd into it, remove the directory (from another terminal window), and then try to run texdoc. It prints the complaint twice, but still succeeds in finding and showing the help item I ask for. What do you get if you run /bin/pwd in the problematic directory?

Hi. Has anyone experience with rendering inference rules in KaTeX/MathJax or anything else that can be used for the web besides a pre-processed PNG?

@HaraldHanche-Olsen Yeah, it's quite weird, and now I can't reproduce it myself, since unmounting and remounting the disk solved the problem.

@AlanMunn Yeah, I noticed that just now. Some strange glitch. In any case, ignore what @DavidCarlisle said about symlinks. They can't interfere with a program finding the path to its home directory.

@HaraldHanche-Olsen yes that's what was worrying me, they can make the parent be not what you expect but shouldn't give that error

@DavidCarlisle Some shells get a little weird when you cd into a symlink. Run the shell's internal pwd command, and it will print the path to the symlink. But run an external command, such as /bin/pwd, and it will not be confused by this since any confusion is internal to the shell.

E.g. in bash, run mkdir -p a/b; ln -s a/b c; cd c; pwd and it will report whatever/c instead of whatever/a/b.

This is intentional, but I think it is just confusing.

@HaraldHanche-Olsen This is how I expect it to be. My cmdline prompt displays the same, and not what realpath would say. Also, if I enter cd .. it will take me back to whatever. Would be weird if I landed in "a" instead.

@AlexG The big problem here is the overloading of ... Most of the time it means up (in the directory hierarchy), other times it means back. If you think external commands and the current shell interpret .. the same, you are in for a world of hurt. But yes, I agree that it is good to have a way to say back.

@CarLaTeX -- I received my advance "editor's copy" last week, and as soon as that happens, the issue is opened up to members. (I've already received a complaint about an opinion expressed in one article -- not mine -- that I'd better not ignore for too long.) I suppose I should consider announcing that here as well as to my "bosses" on the TUG board. For this issue, there will be a delay for recipients outside the U.S.

@barbarabeeton An issue? Was there anything controversial?

@barbarabeeton I hope the complaint is not about van Duck's article

@JosephWright -- It was about a disagreement on a point of mathematics notation. A disagreement of very long standing, Not about LaTeX or the duck pond.

@barbarabeeton Ah, right, I see

@barbarabeeton So nothing really to worry about? (I can guess the article you mean ...)

@DavidCarlisle -- Well established technique. My first job out of college was as a "gofer" and machine operator for an IBM 7070. Every Tuesday morning, three IBM engineers would come in, turn the beast off, turn it on, check if anything failed, then repeat for most of an hour. If nothing to change out, off they went until the next Tuesday or service call.

@AlanMunn I was ... surprised

@AlanMunn A true likeness of @egreg

@JosephWright :)

@JosephWright I thought he was greener than that

@barbarabeeton @UlrikeFischer the typesetting of LuaLaTeX in tugboat section headings is ... interesting

@DavidCarlisle ???

@JosephWright From all the ticks?

@DavidCarlisle where?

@UlrikeFischer @JosephWright well that's weird (probably a firefox nightly thing) it was not showing the small caps A in LaTeX so I zoomed in to take s screenshot then it showed it, even if I zoomed out again, but at the same time harfbuzz flipped to look like

@UlrikeFischer typesetting Bangla is apparently much easier than typesetting Latin script:-)

@DavidCarlisle MANFEOTT??

@UlrikeFischer it cuts and pastes as HarfBuzz ...

@DavidCarlisle is this with a chapter or with the full tugboat?

@UlrikeFischer full

@UlrikeFischer but I am using this morning's build of firefox nightly (pdf rendering is normally OK though)

@DavidCarlisle -- Oh, page 69. Curious. It looks okay on paper. (I've mislaid my member password.) What does the section heading on page 34 look like? (LaTeX News 31.)

@barbarabeeton all the headings there look OK to me

@barbarabeeton @UlrikeFischer as I page around it's gradually getting worse the top of page 71 was OK but I went back to check page 31 for Barbara and now page 71 looks like

@DavidCarlisle buh, couldn't it have the decency to scramble my name too?

@UlrikeFischer actually I should use a more stable pdf reader probably, Or at least kill off that session it looks like something has corrupted its font mapping

@DavidCarlisle -- I've written to Karl, cc to you.

@barbarabeeton Hmm, there aren't many articles on math notation. ;-)

@egreg :)

@DavidCarlisle yes, and there are really many fonts in the full version.

@egreg -- It's the age-old question about how to present the differential "d". I'm researching the history, and if I can get the library access i need, intend to write an essay on the matter.

@barbarabeeton doesn't everyone use ⅆ ?

@DavidCarlisle ooh

@DavidCarlisle -- Not engineers or physicists. Don't have enough information about astronomers, and I know at least one computer scientist who insists on italic.

@barbarabeeton Well, I didn't take a position: the macro I suggest can be modified according to the author's/publisher's preferences.

@barbarabeeton nobody who doesn't get paid by Wolfram:-)

@egreg -- I will take that into consideration when replying.

@barbarabeeton @UlrikeFischer as far as I can see the complete pdf looks fine in Chrome

@DavidCarlisle -- Whew! Chrome is what I'm using when proofing. (Albeit not a current version.)

@barbarabeeton OK, I call the upright d an abomination, but it's clearly a joke.

@barbarabeeton it may be OK in a real firefox as well, I use nightly which is by definition unstable

@DavidCarlisle -- I could make a comment about (in)stability, but will refrain.

@DavidCarlisle ooh what does the fox say

Stay strong, friends! ♥

@PauloCereda <3

@yo' <3

So Elon Musk will need \AE...

@PauloCereda My newest French singer obsessions if you're interested: Pomme and Leïla Huissoud.

@PauloCereda ?

@AlanMunn ooh thank you!

@HaraldHanche-Olsen he named his child a wacky name...

@HaraldHanche-Olsen X Æ A-12

No kidding