« first day (3537 days earlier)      last day (36 days later) » 

12:02 AM
@barbarabeeton That's what I'd do... These checks aren't valid for cashing, are they?
@PhelypeOleinik -- The checks on the "Bank of San Seriffe" aren't, but for a major reward, Don will write a real check. (He has written somewhere that in the "early days" of TeX, there were a few students who kept body and soul nourished from the proceeds of their rewards. I would never cash one. I once waved one of mine in front of one of my nephews, an electrical engineer. He grabbed it, and kept it to show off to his geeky friends. He gave it back a year later, quite smugly.)
12:18 AM
@barbarabeeton Oh, I thought he ended the real checks to avoid too large expenses, but it's the other way around. I don't think anyone who knows what these checks are nowadays would actually cash them. I bet your nephew had some nice stories to tell :-)
@PhelypeOleinik seems that way, you really can't trust editable git histories
@DavidCarlisle :-)
@PhelypeOleinik -- He ended the real checks because some uncouth malicious marauder took over the checking account he set aside for the rewards. He has never begrudged paying a legitimately earned reward. (Admittedly, the account was never able to be balanced.)
1:17 AM
I am trying to cite the following
> @article{ghahramani2015probabilistic,
title={Probabilistic machine learning and artificial intelligence},
author={Ghahramani, Zoubin},
publisher={Nature Publishing Group}
But I keep getting the error
> ! Package natbib Error: Bibliography not compatible with author-year citations.
But that citation contains a year...
I am using \usepackage{natbib} and \bibliographystyle{alpha}
Any idea?
I already removed all files and compiled everything again, but same error
never used natbib
I have no idea of how it works
but it's already quite annoying
I am compiling with
"pdflatex" -synctex=1 -interaction=nonstopmode %.tex|"bibtex" %.aux|"pdflatex" -synctex=1 -interaction=nonstopmode %.tex|"pdflatex" -synctex=1 -interaction=nonstopmode %.tex
1:42 AM
Q: Package natbib Error: Bibliography not compatible with author-year citations (again!)

nbroI am trying to cite the following @article{ghahramani2015probabilistic, title={Probabilistic machine learning and artificial intelligence}, author={Ghahramani, Zoubin}, journal={Nature}, volume={521}, number={7553}, pages={452--459}, year={2015}, publisher={Nature Publishing Group} } with \citet...

5 hours later…
7:02 AM
@cis the symbols are slightly off (this is most recognizable in ex. 17). And seeing the real output with the colours I thought of another option which might be useful: When showing the steps, it would be nice if there was an option to only show the most recent step in another colour and the steps already done in another (is this understandable?)
@PhelypeOleinik I'm so envious of this guy!
3 hours later…
9:55 AM
@UlrikeFischer, @PhelypeOleinik, @DavidCarlisle Expect updated cctabs soon
1 hour later…
11:23 AM
@Skillmonlikestopanswers.xyz I think you mean something like that.
@UlrikeFischer Hopefully I'll be back on color today/tomorrow
11:36 AM
@Skillmonlikestopanswers.xyz That the gray sometimes shines through is probably due to the fact that the editing in FontForge is not a highly professional editing; the TikZ positioning should be correct.
@JosephWright I tried in the morning to understand what pgf is doing (and why colorspace fails there) but this is quite a mess ...
11:52 AM
@UlrikeFischer You asked what a new model needs: currently, \__color_backend_<model>:n, \__color_parse_model_<model>:w and \__color_parse_mix:nw
@UlrikeFischer I've not yet added any code that allows for any of those being missing: the original models are always available, so there was no need
@UlrikeFischer Yes: I'm hoping to start unpicking the backend set up soon
@JosephWright the problem is not so much the backend but that pgf bypasses the xcolor code and so some variables simply don't reflect the right color.
@UlrikeFischer Oh, that would do it, yes: I had wondered why things were not working, as colorspace uses xcolor, and the latter builds on color
@UlrikeFischer So don't worry at present, finish l3color, get pgf to use that?
@JosephWright yes. There must be better interfaces so that this awfull patching stops.
12:11 PM
It errors in 2017 and 2016 too. Also even if it not errorred, did it actually used spot colors? And why are you using spot colors for a presentation? — Ulrike Fischer 3 mins ago
@UlrikeFischer ^^^ It seems to be a poster and not a presentation
@samcarter_is_at_topanswers.xyz yes the example, but in the comment he wrote "presentation".
@UlrikeFischer Ah, right. Maybe they weren't using tikz in the old presentations. colorspace with beamer (and the \def\spc@getref stuff) does not throw errors (no idea if the colours are right or not)
@UlrikeFischer Nevermind, I just saw that you already know all that from the earlier question tex.stackexchange.com/a/552487 (sorry, I did only see the new question before)
@samcarter_is_at_topanswers.xyz only in part, see tex.stackexchange.com/questions/552478/…. The problem is that the "current color" can have a variety of forms (and is sometimes even stored in different commands, depending who did set the color) and so colorspace fails to correctly retrieve the settings. I hope that the new l3color will do it better.
12:27 PM
@UlrikeFischer Yeah, there are some "surprising" interactions between tcolorbox and beamer because they can't agree on the current colour....
@UlrikeFischer We have addressed that part I think: once I have creation-of-colorspaces sorted, it should be ready to go
@JosephWright yes, then we will have to convince pgf to use it which probably will mean that they have to rework their color code. But imho using spot colors will only work properly in pgf/beamer if it passes through the l3 interfaces.
12:48 PM
@cis I meant something like that, but with the symbol not being grey but white (which can easily be adapted via your colour keys).
1:02 PM
Is there a principle way to animate such a filling?
Say filling the red from the left to the right.
@cis in TeX you'd have to produce small steps and then would use ImageMagick's convert to turn it into an animated gif.
Can the construction of a character be done in steps?

(This is not a primary goal of mine now, I just had it in mind.)
@Skillmonlikestopanswers.xyz I believe that you can do PDF animations, but I'm not sure it's easy or robust
@Skillmonlikestopanswers.xyz Because I am interested in the topic, I will ask a general question on the main page, e.g. animated filling of the character ' / '.
@cis not that I know of, but that is nothing I was ever interested in.
@yo' might be possible. Maybe something like a stepwise blend with a moving blending axis is possible (think StarWars scene changes), but I don't know whether something like this would be supported in PDF.
1:11 PM
@Skillmonlikestopanswers.xyz supported likely yes. Welcome? probably no :)
@Skillmonlikestopanswers.xyz Can't say I'm not... just a bit ;-)
1:28 PM
Maybe there is no answer, but
Q: Animated filling of a font charachter

cisIs there a way to fill a sign, say the sign ' / ', with red in steps? Because, I ask myself if I could create sth. like that, using the corresponding character from a matching font (in the picture U+3042 (12354)): No MWE: \documentclass[margin=5mm]{standalone} \usepackage{tikz} \begin{document} ...

@barbarabeeton Well, and also things like {\Large This is my Project Report and the Title is Longer than One Line} (I consider it fine to manually format the title page) will most likely give you wrong baselineskip :-)
@yo' when I first designed a title page this really bugged me :)
@Skillmonlikestopanswers.xyz \let\endLarge\par and \begin{Large}This is my Project Report and the Title is Longer than One Line\end{Large} #dirtyhacks #sinfulcoding #annoyDavidCarlisle
@yo' stop hurting rabbit eyes! #animalcruelty
1:45 PM
@yo' well that's more or less what center does with \centering, but uses a trivlist to annoy @barbarabeeton
@DavidCarlisle :-)
@DavidCarlisle \let\Large@cmd\Large\def\Large{\def\@tempa{Large}\ifx\@tempa\@currenvir\expandafter\Large@env\else\expandafter\Large@cmd\fi}\def\endLarge{\endtrivlist}\def\Large@env{\trivlist\Large@cmd\item\relax}?
2:38 PM
People who have several TeXLive installed: do you have any trick to reduce disk usage, or just really large disk capacity?
@PhelypeOleinik You could drop the source and doc trees
@PhelypeOleinik I just accept I need a big disk: current laptop has a 256 Gb HD, I will make sure the next one is 512 Gb
@JosephWright Hm... Yes, that should help. Thanks!
@PhelypeOleinik Now, who wants to buy me a new MacBook Pro 13" Dell XPS 13"
@PhelypeOleinik It's what I do for the Travis-CI set ups, which have size considerations; somehow I doubt I need the docs for TL'09 any more ;)
@JosephWright Oh, wow. Mine is 512 and I'm already running out of space :-) I need to declutter...
@PhelypeOleinik Depends what else you have installed, of course
2:42 PM
@JosephWright According do @DavidCarlisle, you don't need the TL'20 docs either :-)
@JosephWright 30 GB simulation results :-)
Several of them
@PhelypeOleinik Simulations?
@PhelypeOleinik I don't really know what you do as a day job ;)
@PhelypeOleinik TL'09-TL'20 just over 60 Gb including everything, then I have MiKTeX 2.9 and a Linux VM for TL'20 there
@JosephWright 48 years of wave hindcast results in different parts of the Brazilian coastline.
@JosephWright Student doesn't count as a job. Hopefully I'll get to delete these results soon and get a proper job :-)
@PhelypeOleinik Ah, ENV (environmental science: the department next to mine)
ubuntu@ip-172-31-12-203:~$ df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/root       7.7G  7.2G  496M  94% /
2:49 PM
@DavidCarlisle :)
@DavidCarlisle How many hits today?
@JosephWright I have 15, and 18-20. Last three years or so they started to grow considerably in size
Tue Jul  7 14:49:18 UTC 2020
  Number of distinct users:     2
  Number of document requests:  6
  Number of documents returned: 2
  Number of unique user agents: 2
Mon Jul  6 23:58:36 UTC 2020
  Number of distinct users:     8
  Number of document requests:  20
  Number of documents returned: 12
  Number of unique user agents: 6
Sun Jul  5 23:56:04 UTC 2020
  Number of distinct users:     5
  Number of document requests:  68
  Number of documents returned: 66
  Number of unique user agents: 7
Sat Jul  4 23:57:43 UTC 2020
@JosephWright I'll apply to work next to you then ;-)
@PhelypeOleinik Lots of hardcore simulation stuff?
@PhelypeOleinik which OS do you use?
2:52 PM
@JosephWright You mean what I'm doing for my thesis?
@DavidCarlisle Well something has improved with tex.ac.uk, as now it at least gives a 404 reliably for all combinations :)
@PhelypeOleinik Yes
@Skillmonlikestopanswers.xyz Ubuntu 16.04
@PhelypeOleinik don't know about apt, but maybe remove some cached packages (on my Arch, with pacman caching quite aggressively, this clears roughly 20 GB).
@JosephWright I can't remember what's supposed to happen, does it all go to github pages as "normal" or does it go to Stefan's server and just redirect the texfaq urls to texfaq.org?
@DavidCarlisle Just meant to go to GitHub now: @StefanKottwitz no longer involved as I have DNS control
2:55 PM
@JosephWright ah OK
@DavidCarlisle I've looked up a few people trying similar stuff, and they say it should work, or they say one needs some CloudFlare voodoo
@JosephWright I wrote a program to transform large-scale sea state data like that (time scale of days to years) into small-scale data (fractions of a second to a few minutes) to be used in CFD simulations, plus some case study for wave energy conversion
@PhelypeOleinik yep, all those templates only useful to a handful of students, but of course they have to be in TeX Live :(
@Skillmonlikestopanswers.xyz How do you do that on pacman? I'll try to find an equivalent for apt (though the disk usage of package files seem low)
@Skillmonlikestopanswers.xyz I bet I'll be the only one to actually use the template I made. It seems people here would rather have an arm cut off than drop Word
@PhelypeOleinik Ah, right, so big data then: makes sense (a lot of environmental science seems to be that or going to exotic holiday destinations 'to collect data')
2:59 PM
@DavidCarlisle Yes, also on mine the space is scarce
> df -h /System/Volumes/Data
Filesystem     Size   Used  Avail Capacity iused      ifree %iused  Mounted on
/dev/disk1s2  932Gi  620Gi  291Gi    69% 4887252 9763090908    0%   /System/Volumes/Data
@egreg Stumbled across arxiv.org/abs/2003.13758 - bonkers (no idea if the maths is solid)
@egreg Show-off
@egreg That's the MacoBook Pro 16"?
@PhelypeOleinik pacman -Sc
@PhelypeOleinik My template caught on, I know three other students using it! (but it's not on TeX Live, because of reasons)
@JosephWright so A to gh-page and C to www.texfaq.org ?
@JosephWright :-D
% df -h /
Filesystem             Size  Used Avail Use% Mounted on
/dev/mapper/main-root  139G  131G  1.4G  99% /
Living on the edge ^^^^
3:03 PM
@DavidCarlisle Indeed
$ORIGIN tex.ac.uk.
*                       300 MX    10     mail.texfaq.org.
@                       300 A  
@                       300 A  
@                       300 A  
www                     300 CNAME        www.texfaq.org.
@JosephWright Yes, with the online teaching business it has filled with nonsense
@JosephWright Yes. The purpose of the program I wrote is to make this large-to-small scale conversion as needed to avoid storing terabytes of probably-won't-ever-be-used data
@Skillmonlikestopanswers.xyz Thanks!
@Skillmonlikestopanswers.xyz Good luck solving their problems whenever anything unrelated goes wrong ;P
@egreg Wow: we are at present still working out details. My hardcore recordings are on my work PC, which is quite fast (i9 workstation)
@PhelypeOleinik Ah, right: what do you use for that?
@JosephWright I think the CNAME file at gh pages should have www to match?
@PhelypeOleinik two of them already finished. And I also solve the problems of other friends because somehow they found out I know stuff.
3:05 PM
@DavidCarlisle So you think in the above I just want texfaq.org?
@JosephWright An Inverse Fourier transform to transform wave spectra into time series of sea surface elevation, plus some heuristics to fix too large waves that don't actually exist
well the two sites I have that are working have woring (including learnlatex) have www in the CNAME file
@PhelypeOleinik Aj right
@JosephWright Written in good'ol Fortran :-)
@PhelypeOleinik isn't all software written in fortran
3:07 PM
@DavidCarlisle We can change that, but shouldn't changing the CNAME record for tex.ac.uk to drop the www do the same?
@DavidCarlisle People like abstractions that eventually are translated into Fortran :-)
@DavidCarlisle And TeX
@JosephWright possibly, my knowledge of dns is sketchy in some details. I think they should be the same anyway
@PhelypeOleinik I generate a lot of fortran from xml, you would like it here
@DavidCarlisle Seems like an interesting job :-)
@DavidCarlisle /Applying at NAG
@PhelypeOleinik although currently generating around ~2000 C files for matlab interfaces so people can use matlab not fortran (once they all work) From the same XML that generated the fortran interface and teh public html documentation
@DavidCarlisle The first language I learned was matlab (it's easy: you just hit a button to evaluate some bit of code), but it started to be too slow for what I needed it for (and I never got the hang of writing C code with matlab interfaces), so I learned fortran for the heavy stuff. I use matlab mostly for making pretty pictures of my results now
3:54 PM
@DavidCarlisle I've tried changing the CNAME for tex.ac.uk to texfaq.github.io: doesn't seem to be working :(
@JosephWright www on both?
@DavidCarlisle No, as texfaq.github.io doesn't have www
@yo' -- That will certainly give the wrong \baselineskip. The concept that \baselineskip (and some other things) aren't "frozen" until a paragraph break is somewhere between elementary and intermediate. If it's not in the tutorial, it should be. (I guess I should check.)
@JosephWright gh pages only say you can give cname to a subdomain and all their examples use www. texfaq.org works (redirects to texfaq.org) https:www gives a certificate error but at least at the dns level www.texfaq.org exists#
4:10 PM
@PhelypeOleinik -- Where in the "student" stream? Are you in the "tend to your thesis" cohort? If so, I'm willing to volunteer for that duty.
@DavidCarlisle OK, so we should adjust that first; do you remember who registered/controls texfaq.org DNS?
@JosephWright you?
@DavidCarlisle Don't think so ... can't track it down
@JosephWright wasn't it @StefanKottwitz?
@DavidCarlisle I've adjusted the CNAME file on GitHub, but that leads to a warning as I can't alter the DNS. Was it @StefanKottwitz?
@UlrikeFischer Exactly
4:19 PM
@JosephWright someone in Germany: whois.com/whois/texfaq.org
@barbarabeeton Yes, I'm finishing some of the coding part then I only need to finish writing the results chapter and some tidying of the structure of the text. It's in Portuguese though...
@DavidCarlisle Changing the CNAME doesn't seem to be wroking ...
can you not have A/C records for tex.ac.uk an just have an ALIAS record making it texfaq.org?
@DavidCarlisle I can try that
@DavidCarlisle Nope, Mythic Beasts don't let me add arbitrary records
4:38 PM
@PhelypeOleinik -- Nagging doesn't involve understanding the language it's in, only being encouraging as the due date approaches. If you'd like that "service", let us know when it's due (or when you expect to finish it). It does sound like you're well along.
@barbarabeeton Oh, like “finished your thesis”. I saw that before :-)
@barbarabeeton With the whole covid thing, I have until almost christmas to finish it, but I'd rather not. I'd like to have it done in a month or so
4:54 PM
@PhelypeOleinik -- Yup. That's one of them (though not by me). I'll start encouraging now. The most important part seems to be making sure the overall structure is logical. Onward!
@barbarabeeton Thanks :D
@CarLaTeX -- Since it seems to be thesis-encouragement time, be encouraged. How's it coming, or has the lockdown been a (thesis) slowdown?
@barbarabeeton No, it has been a fastup, not very fast, though :)
5:11 PM
@JosephWright what DNS server do you want to use? I can set them for the domains and then you control the actual DNS entries
@DavidCarlisle texfaq.org and www.texfaq.org both point to You may need a web server level redirection and a proper certificate that covers both names-
@DavidCarlisle -- I've just finished reading *The Dictionary Wars", by Peter Martin. In it, I have learned that there's very little actual Webster in a "Webster" other than the name. I have also confirmed that even a literate author can misuse "reign in" instead of "rein in". Comparing the n-grams for the two would lead one to believe that the former, being the one found most frequently, must be correct, but that's wrong; the meanings are entirely different, being proper in different contexts.
@JosephWright I could handle www.texfaq.org separately as just a redirect service to texfaq.org including an own certificate
@AlanMunn -- I have a gripe against n-grams. (See my previous comment.) Without context being shown, mere frequency can be dreadfully misleading. How to improve the situation?
@barbarabeeton There's not much to be done about the limitations. You need to be very aware of them. The corpora do have some part of speech tagging, which can sometimes be helpful, but it's still a very rough tool easy to be misused, especially by people who don't know what they're doing.
Hello all, this is an opinion-based question maybe :). If you experts would start a new document with listings, would you go for minted or still use the classic listings package? (related tex.stackexchange.com/questions/389191)
5:22 PM
@AlanMunn -- Sigh. Could "rein in [something]" vs "reign in [somewhere]" (and the alternatives) be made workable? I.e., where are available tags to be found?
@barbarabeeton Probably not. Since 'something' and 'somewhere' will have the same syntactic category, there's no way to get this.
@barbarabeeton You'd need to actually get the examples and then code them for the semantics, I suspect.
@DavidCarlisle and you could make www to an A record instead of a CNAME (extra complexity, not faster)
@AlanMunn -- Double sigh. How is one to learn "proper" usage then? I'm amused by the current kerfuffle over "irregardless". My high school English teacher gave this as an example for how/why not to use it (although it is indubitably still a word): "Irregardless of no matter what!" Well, that's definitive, but not something that goes into some formal context.
@CarLaTeX -- Well, faster then, girl! You haven't that much further to go! And you know you'll be delighted to have that piece of parchment in your hand.
@barbarabeeton You mean proper usage of ngrams or English?
@AlanMunn -- English, actually. So that I wouldn't get quibbles from native English speakers when I change some bit of their prose.
5:34 PM
@barbarabeeton I'll try :)
@Dr.ManuelKuehner If you are happy with how things look like, listings is easier to use. If you want to customise, the possibilities with minted are nearly infinite ... So just a matter of taste really
@barbarabeeton Hmm, I guess I wouldn't use ngrams for sure. But I don't know really. And language does change, so prescriptivism is always a tricky issue.
@Dr.ManuelKuehner In case you speak French, a similar questions was recently asked at texnique.fr/osqa/questions/8766/…
@AlanMunn -- I do try to avoid prescriptivism. If there's obvious logic involved, that's what I'll use for an argument. (But I've failed to convince DEK to not follow Chicago on the question of in what order should final quotes appear vs. final punctuation. If there's an obvious interpretation problem, he'll rewrite to remove the problem in another way.)
@JosephWright texfaq.org:
5:47 PM
@barbarabeeton Is Chicago quotes outside? I always thought this was crazy.
@StefanKottwitz I can't do anything:-) @JosephWright has the domain
@DavidCarlisle yes, I just used the chat answer function to stay within the thread and it was your message in that thread :-)
@AlanMunn -- Yes, it's crazy; Fowler says use logic, which I find logical. I finally got AMS to change after showing the editorial folk pleas for help from confused users trying to follow instructions in a user manual telling them to 'Type this "xxx ."' where if one types the period, insanity ensues.
@barbarabeeton Yes, Fowler has always been a voice of reason. Much better than any of the American style guides.
@barbarabeeton Not to mention having a much better sense of humour. :)
5:53 PM
@AlanMunn -- I generally ignore Strunk & White, but there's one (rather late) edition that has delightful illustrations, and I have a copy of that just for the drawings.
@PauloCereda -- Quack! (Finished your thesis? ... see earlier messages today.)
@barbarabeeton oh no
@barbarabeeton <3
More people with theses. :)
@PauloCereda -- And needing encouragement.
@barbarabeeton and poking!
@PauloCereda but squares are not so good for eating
6:07 PM
@DavidCarlisle oh no
6:59 PM
See the points 19, 20 and 21.
PS: And there is a wrong startpoint in (19), (20) to fix at FontFrorge.
7:14 PM
PPS: All this colored stuff is not that important to me.
But it could become important with the Kanjis.
7:31 PM
1 hour later…
9:00 PM
@JosephWright google doesn't like my Japanese:(
@DavidCarlisle ?
@UlrikeFischer automatic mobile accessibility report from google search console, says elements too close together
@UlrikeFischer we can now use %!TEX log \showoutput for all examples and don't need to bother with pdf at all
9:17 PM
@DavidCarlisle Not quite sure if suitable for beginners but nice ;-)
Anyone here know anything about .htaccess files?
@AlanMunn I knew more 20 years ago, but yes
@DavidCarlisle MSU is eliminating our afs server which hosts my personal web page. The url is www.msu.edu/~amunn When I add the following redirect to the .htaccess Redirect 301 "/" "amunn.github.io/index.html" it ends up redirecting to "amunn.github.io/~amunn` which of course doesn't exist. I've tried various versions and none work.
@DavidCarlisle Somehow the ~ is messing things up.
don't you want Redirect 301 /~amunn https://amunn.github.io/index.html
@DavidCarlisle Nope, that does the same thing.
9:35 PM
@AlanMunn that's odd, you could use mod rewrite instead of the redirect
@DavidCarlisle I just got word from IT that I can't do it that way (even though other htaccess stuff works). So I need to do a meta tag instead. Thanks.
@AlanMunn ah
9:55 PM
@DavidCarlisle I ended up with a javascript redirect instead.
@AlanMunn so the university is not hosting personal pages at all and you all have to move to github or similar?
@DavidCarlisle Well they're providing some Wordpress crap I think (lowest common denominator) but no useful space in the way the old afs space was.
@DavidCarlisle And jekyll makes maintaining the site on github far easier anyway.
@DavidCarlisle The only downside is that it removes the msu domain from the picture and is therefore less 'official', but I think more and more people are doing this, so the officialness is not much of an issue these days.
@AlanMunn we've been having some success with that setup (latex-project.org, texfaq.org, learnlatex.org Joseph's texdev.net are all gh-pages)
@samcarter_is_at_topanswers.xyz Thanks for the replies!
@DavidCarlisle Yes it seems really nice. I rebuilt my site using jekyll in a day, basically, not having ever used it before. The hardest part was getting it installed properly on the Mac, because of the new OS restrictions (well new to me, since I've just upgraded the OS).
@DavidCarlisle But building the site itself took very little time.
10:05 PM
@AlanMunn Instead of JavaScript, you could do a http redirect on the web site, such as <meta http-equiv="refresh" content="0; url=https://amunn.github.io/index.html" />
@StefanKottwitz Yes, I know. There seems to be some disagreement on which is the better route.
@AlanMunn the lower level the redirect, the better. Not requiring JavaScript for the basic redirect may be good
.. or that http redirect within a <noscript> tagged part as fallback to the JavaScript way if that is not active on user side
@AlanMunn for latex-project a local jekyl is needed but for the others (and my own sites) I just let github run jekyll
@UlrikeFischer ^ looks like a beginner document to me
@DavidCarlisle But then aren't you developing with the live site? Or is there a way to test changes before it's live?
@AlanMunn but as I never make typos or errors that isn't an issue. (for testing setup i usually have a spare test site but once it is up and running, at say souldern.org or texfaq.org yes we edit live)
10:12 PM
@DavidCarlisle Lucky you had time to fix the typo you didn't make. :)
How did I manage to get this second frame?
@UlrikeFischer does it go if you click edit
@AlanMunn exactly
@DavidCarlisle ah yes, now it is gone.
@UlrikeFischer pretend it never happened
10:16 PM
@DavidCarlisle yes, I got an error first because of some chars at the begin.
@AlanMunn at the village website for example it was sort of politically important to be able to say that it was a community resource and anyone with a web browser would be able to edit it. If I'd said all you need to do is install the right version of ruby and a pile of gems to support jekyll then it wouldn't have been quite the same. In practice no one other than me edits it and people just send me stuff by email but...
@DavidCarlisle does this output look right? (it is from tex.stackexchange.com/questions/552666/…).
@DavidCarlisle Yes, that makes a lot of sense for sure.
@UlrikeFischer hmm not sending it as utf8? I thought it would be utf8 by default these days, I'm not explicitly setting any headers just serving the log file and letting apache sort that out as text. I'll check
@DavidCarlisle no doesn't look like utf8, grüße came out as
....\TU/lmr/m/n/10 r
....\TU/lmr/m/n/10 ü
....\TU/lmr/m/n/10 ß
....\TU/lmr/m/n/10 e
10:25 PM
@UlrikeFischer oh no, I went to the source code to make it default to utf-8 and see what it says
# Read the documentation before enabling AddDefaultCharset.
That seems very unfair
@DavidCarlisle ;-)
@StefanKottwitz Yes that's good idea. Thanks.
@UlrikeFischer basically it warns that it's too strong and will over-ride encodings specified in files, need to check how to set it just for log files, must be easy enough
@DavidCarlisle yesterday I spent an hour to debug a miktex crash (starting from a long document ...): it failed because pdftex send a char 169 to the terminal and miktex didn't like this. github.com/MiKTeX/miktex/issues/581
@UlrikeFischer can you try now
10:32 PM
@DavidCarlisle no I see still bytes.
@UlrikeFischer OK I'll set up a simpler test and experiment possibly tomorrow though, @AlanMunn suggests I shouldn't experiment on live sites
@DavidCarlisle it's fine, I have to go to bed now anyway ...
@UlrikeFischer night
@UlrikeFischer latexcgi.xyz/tst.log doesn't work but easier to test than your original
@DavidCarlisle that is from me too isn't it?
11:01 PM
@UlrikeFischer yes I grabbed it before it got deleted:-) (I think it should work now)

« first day (3537 days earlier)      last day (36 days later) »