« first day (3216 days earlier)      last day (27 days later) » 

6:12 AM
@AlanMunn I don't think there is actually any such thing as editor wars. It's imaginary. Like the weapons of mass destruction.
7:05 AM
@FaheemMitha On Mercurial and Bitbucket, I think it's a simple commercial decision: they are likely seeing hardly any usage, and the costs don't make sense
@JosephWright Sure. Except that they have been neglecting Mercurial for years.
Possibly since they purchased it from Jesper.
If you don't maintain a system, you can hardly expect people to want to use it.
TeX is 1000 times more active, and nobody gets paid to maintain TeX.
@FaheemMitha I know
@FaheemMitha I'm afraid I think they are right about the way things are going: Git for new projects, SVN for older ones
@FaheemMitha Certainly jobs for maintaining TeX are ... rare
@FaheemMitha It's all academics, though, really, so it's not 'real' work
@FaheemMitha @UlrikeFischer does offer paid-for TeX support: she's best placed to know about the marketplace
@JosephWright I'm told lots of people use Mercurial. It just isn't particularly visible.
Particularly for proprietary projects. And ironically, it's quite popular on Windows. Or so I hear.
@JosephWright I think the academics would disagree about it not being real work.
@FaheemMitha I know what my real work is: TeX's not it!
@FaheemMitha I am an academic :)
@FaheemMitha Quite possibly, but I think if one were hiring today for a commercial project, I'd be asking about Git experience
@FaheemMitha I used Mercurial for several years, so I have no issue with it (I always found the interface a bit better than Git, though I think that has likely levelled out)
@JosephWright Version control isn't really significant. Particularly specific version control. It's something you can easily learn if you need to.
If I was hiring someone, it's not something I would focus on.
7:14 AM
@FaheemMitha Fair enough
@FaheemMitha True, more likely to be a one-liner than something to probe
@FaheemMitha As the only languages I could ask about are TeX or Lua, I'd probably see if they understood expansion :)
@JosephWright Actually, these days Mercurial is somewhat ahead. Due to heroic efforts by Pierre-Yves. Though it's possible that most people wouldn't care.
... and catcodes
Pierre-Yves has spent most of this decade trying to get Evolve+Topics working.
@FaheemMitha I meant that there are a lot of good GUIs nowadays for Git, and I think also more guides on how to use it at the CLI
From this I conclude that it's not that easy. Though from the usage side, little of that effort is visible.
7:16 AM
@FaheemMitha That sounds too advanced for me :)
@JosephWright That's certainly true, yes.
@JosephWright It's not advanced from a usage perspective. Though from a development standpoint, it sounds quite hairy. Still work in progress. And I personally have encountered many bugs.
@FaheemMitha Also the general idea of DCVS is well established, which makes a difference
I just switched to it myself last year. Before that I was scared of the bugs.
@FaheemMitha Imagine I don't know what 'Evolve+Topics' means :)
@JosephWright The basic idea is quite simple. If you are rewriting a cset (which you can also do in Git), imagine the previous cset is still there, just invisible.
To a first approximation, that is what Evolve is. Of course, it get's more complicated than that. Rather like TeX.
You could describe TeX to someone by saying. Well, imagine we replace this piece of text with another piece of text. if this piece of text has a certain form.
7:19 AM
@FaheemMitha Same basic 'rewrite the history' idea as Git? In contrast to say SVN
@JosephWright Yes, rewriting history is a general idea. But specific to a DCVS. Doesn't really make sense in a centralized system, as far as I know.
But Mercurial handles it in a systematic way. And all the history of everything, the previous csets, and all the changes that have been made, are all still there.
Unless you delete it, of course.
Topics are lightweight branches that are designed to play nice with Evolve. So they come as part of the package. Technically they're separate.
Mercurial has more to offer than that, of course. But Evolve is kind of a killer app thing. Like PGF/TikZ, say.
Oh, and if you've rewritten a cset and want to undo your changes, it's fairly trivial to do so, because the original cset is still there. But bear in mind that the process is append only. Nothing is ever deleted. So undoing means creating yet another cset.
I'd been trying to switch to Evolve for years. But every time I did so, I'd run into some nasty bug. So I got put off. But Pulkit Goyal talked me into trying again last year. And the system does seem to have settled down now.
I've not encountered any showstopper bugs yet, touch wood. And more people are using it now, too.
@FaheemMitha It's actually the same in Git, at least until garbage collection: if you know the hash, you can get back an 'invisible' cset. There are only so many ideas :)
@JosephWright I doubt Git handles it systematically, though.
But I've used Git, so I don't know the details.
And I'm not aware of a comparison of that aspect of Git and Mercurial. Perhaps someone should write one.
@FaheemMitha It's non-trivial, certainly, but it's there
@JosephWright You spend a lot of time on it, for it not being your real job. Actually, I'm not even sure how that's possible.
Kind of like two different full time jobs.
@JosephWright In Mercurial, all history is available indefinitely, with all the changes. Including a description of how it was changed (there are different categories). And the rewriting happens remotely on push. And it works with the phases mechanism, which has different categories for csets. Secret, draft, and public. You can modify drafts, but public means it has been pushed, so you aren't supposed to change it any longer.
Though you can force-change it, of course.
@JosephWright There's a lot of difference between and idea, and having it working. These ideas appeared around 2010. And it's only recently that they have become usable.
Apparently there are a lot of special cases to deal with.
7:37 AM
@JosephWright I think I could have use for a GUI too, mainly to better see diffs. Do you know one which makes it easy to see differences and changes between branches and commits?
Anyway, at this point I'm rambling.
8:08 AM
@UlrikeFischer GitKraken (my GUI) lets me pick two commits and see the diff ...
8:26 AM
@JosephWright ok I will try. But I hope it has a non-black interface too. On the video of their site I can't see anything ;-(.
First commit to the evolve repos was 20 May 2011 by Pierre-Yves. Currently at 4833 commits.
@UlrikeFischer Not sure about that: I've been quite happy with the dark theme
@UlrikeFischer Yes, has themes
@UlrikeFischer ^^^
@JosephWright much better ;-) I know that quite a lot people prefer dark themes, but for me they are very tiring somehow. I saw that there is an in-built editor. How does it integrate with an external one? Does it pick up if a file is changed there?
There's also an ancient slide show from 2011, before of this stuff was actually available for real. Mostly pictures. public.octopoid.net/talk.pdf
There are probably better slide shows around.
8:47 AM
@FaheemMitha beamer :)
@JosephWright I meant the contents, not the presentation software.
@UlrikeFischer It is good for last-minute fixes; picks up changes 'live'
And I have no idea what is being used for that slide show.
Though maybe I'm misunderstanding your comment.
9:06 AM
Q: Does anyone know this math font?

makalanew to this forum. Does anyone know what math font this is and how I can use it/if it's available free? It's from "An Introduction to the Theory of Mechanism Design" by Tilman Börgers (this clip from page 12). Thanks very much.

@HenriMenke @Mensch Hi, the fonts used for this question are ArnoPro + OUP1 .ttf www-personal.umich.edu/~tborgers/Sample.pdf
Good morning Paulo :-)
@Sebastiano Hi
Never heard of the OUP1 font before.
@UlrikeFischer ^^^
4 hours later…
1:00 PM
@JaspervanLooij My MacBook Air is 5 years old and builds LaTeX fine. But I won't buy another one. New ones can't run Linux :/
@mickep The OUP1.ttf font is just used for the OUP logo on the title page I think. I remember using it ages ago when I used to work for a company that typeset OUP's Australian high school maths books :)
(In LaTeX of course! Although we used some ancient version of Scientific Workplace, which was not that impressive to my mind)
1:46 PM
@PauloCereda btw, javascript is famous for very logical implicit conversions :-)
2:09 PM
@DavidPurton Oh, interesting. This also means that the font in the question is something else.
2 hours later…
4:05 PM
@PauloCereda got my mail?
4:20 PM
@UlrikeFischer Ohh secret mail
@JosephWright ;-) I played a bit with gitkraken and getting used to it -- after I found the "view all files" button I was even able to change a file ;-)
@UlrikeFischer I think it's quite good, better than the alternatives I've tried
@JosephWright apart from the dark support site ... do you have pro version or is the free one enough?
@UlrikeFischer I know who the customer is this way :)
@UlrikeFischer Pro
@UlrikeFischer oh sorry, I completely forgot about it... I will write to you in a minute...
@JosephWright ooh
4:26 PM
I think marmot has re-spawned as Schrödinger's cat...
@JosephWright did you authentificated the access github etc? I stopped that for the first test and created a local account which seems to work.
@Werner pst.
@UlrikeFischer My guess is based on the profile description.
@Werner such a cat is a fragile thing, if you look at it ...
4:44 PM
@Werner Sometimes They Come Back
@Werner Seems likely, from the evidence.
Physics reference. Went from 1 to 1.4K overnight on mostly TikZ questions.
I smell a Marmot in Cat's clothing.
Though what he's doing, I've no idea. Waiting for the box to be opened, perhaps?
And yes, based on the style it's him. Or a really amazing coincidence.
Apparently TeX SE really is addictive. Maybe they should start a TikZ Anonymous.
5:13 PM
Q: How to recover from TeX.SE (or, in general, TeX-LaTeX) addiction?

CarLaTeXTeX.SE is addictive, we have recently had evidence. I don't think this is only due to the reputation mechanism. For example, the little games available on FB have a similar reward mechanism: you win points, and you can compete with your friends, but they don't give me any form of addiction (may...

@CarLaTeX Ha ha! Indeed.
@Werner :)
5:37 PM
@DavidPurton Thanks.
6:24 PM
This question involved quite a bit of squirrely debugging; I made suggestions, and the author did the testing: Auto index complicated formulas with catcode Question: All suggestions and testing were done via comments. Now that the results are successful, is it worthwhile to reiterate everything in a formal answer, or is the original problem insufficiently general that it will help someone in the future?
6:39 PM
@UlrikeFischer Yes, I use the connected approach: it's handy
@UlrikeFischer Also has option for GitLab, BitBucket
@UlrikeFischer I really like GitKraken: it's pretty good, particularly since they added rebasing
@JosephWright what does it do additionally to simple push/pull?
@UlrikeFischer Means it can post PRs on your behalf, check for repos (I think), etc.
@UlrikeFischer We don't use a formal 'always PR' approach, so it's perhaps not so useful, but I do like being able to make one from a branch in the editor
@JosephWright I was thinking about this lately. So you don't have to push a new branch first online before a PT?
@UlrikeFischer Does my message about the normalisation issue make sense? I can't find a change in l3build or luaotfload that causes it
@UlrikeFischer No, you can create a local branch then push-and-PR in one go
@JosephWright well I don't really see how the binary can be the cause. But I couldn't really make sensible tests due to the broken luaotfload loading. Imho this should be sorted first.
6:47 PM
@JosephWright -- Maybe you could address the question I asked a few items above?
I just got the message that unpacking of esint.dtx is broken because "my automatic corrector modified 'mffile' to 'muffle'!" ;-)
7:05 PM
@UlrikeFischer Yes, sounds about right
@JosephWright I'm not sure about latex2e with the curious fontloader in test2e but for latex3 either disabling luaotfload or simply allowing a search in texmf-dist should imho work.
@UlrikeFischer -- Aaaaarrrrggghhh! I hate "automatic correctors"!
For some reason my Emacs doesn't seem be honoring the TEXINPUTS env variable right now. I have export TEXINPUTS=/home/faheem/personal/texmf// in my ~/.bashrc.
It does work on the command line. Any idea why this might be the case? In any case, I'm not sure I should be relying on the environmental variable.
@FaheemMitha Perhaps emacs is using a non-login shell. What happens if you execute TEXINPUTS=/home/faheem/personal/texmf// emacs and try to run from within?
Perhaps that particular Emacs wasn't spawned from a shell with that environment variable set, for some reason.
@PauloCereda Yes, I think that (or something similar was true). Sometimes when I log in there is an instance of Emacs already running. If I used one of those it could explain it.
Is there a good way to force this automatically?
7:19 PM
@PauloCereda is it okay if I pass the address?
@FaheemMitha What happens when you try M-x getenv?
And supply TEXINPUTS
@UlrikeFischer I'm a bit afraid, but fire away.
@PauloCereda you can ask me if there are problems (and perhaps nothing happen anyway) ;-)
@UlrikeFischer :)
@PauloCereda It includes that path, but I killed that Emacs instance.
@FaheemMitha you monster. :)
7:23 PM
I think relying on env variables might not be a good idea, anyway.
@FaheemMitha subshells probably use noninteractive configurations...
@PauloCereda I'm not sure what that means.
@FaheemMitha vvv
paulo@cambridge ~] $ which -a youtube-dl
alias youtube-dl='/opt/paulo/softwares/binaries/youtube-dl'
[paulo@cambridge ~] $ bash --noprofile --norc
bash-5.0$ which -a youtube-dl
It depends on which files are being read.
@PauloCereda Ok, but not sure if noninteractive is the right word.
@FaheemMitha sorry, it's not mine. :)
7:30 PM
I think login vs non-login is conventional.
2 hours later…
9:40 PM
Hey guys!
I have a picture which I imported from drawing software in tex
now I am trying t oplace an arrow at a particular location
in the figure
but unable to do so as i donot exactly know the coordinate where I am trying to put the arrow
any help?
10:27 PM
@BAYMAX Trial and error? This? tex.stackexchange.com/a/9562 It's not really clear to me what you've done so far, so hard to give more specific advice.
sounds cool! @TorbjørnT.
will get back to u soon

« first day (3216 days earlier)      last day (27 days later) »