« first day (3207 days earlier)      last day (1727 days later) » 

4:31 AM
@FaheemMitha I find it easier to stick with Debian Testing. Rolling updates means almost never having to reinstall (although I hosed my machine not long ago and had to…) But once in 5 years isn't bad. Plus TeXLive is kept up to date in Testing, (though it's a big download every month because of how Debian packages things). I even run Debian on Mac hardware (though this is a pain). With the current generation of Macs it's basically impossible, so I'll be going back to PCs for my next laptop.
 
 
1 hour later…
5:49 AM
@DavidPurton So you're a Debian user? For how long?
 
@FaheemMitha Oooo. more than 20 years. It was the first Linux distro I used. And haven't used anything else as my main OS since Windows 98.
 
@DavidPurton I've been using it since 2001. I haven't really used anything else since.
I'm administered other distributions, but I don't really use them.
Before that SUSE, and before that RedHat. When I installed RedHat on my home machine, it came preinstalled with Windows 95.
That was in 1998.
BTW, have you tried persuading your fellow Tikzling to be sensible and come back?
 
@FaheemMitha I fear I have no sway in these matters. I just stole his avatar ideas…
2
 
6:35 AM
Is luaotfload a general purpose thing? As in, you are using it if you are using LuaTeX? I don't really know what Opentype fonts are.
And is more manual style memory management in Lua not an option? Quoting from
> and memory usage in lua can be large as allocations doubles when more is needed
Eep.
 
@FaheemMitha luaotfload is always loaded with lualatex (look in the log) but not in plaintex.
 
6:51 AM
@UlrikeFischer So it is used by default for font loading in LuaLaTeX?
 
@FaheemMitha yes.
 
@UlrikeFischer I see. Thank you for letting me know.
 
@FaheemMitha you also need it or something equivalent in plain luatex if you want to use Unicode fonts, but plain luatex defaults to 7bit tfm fonts like classic tex.
 
@DavidCarlisle So plain LuaTeX uses something else?
 
7:13 AM
Good morning everybody.....
There is a italian link (excuse me)....for 60000 or more ducks :-)
 
8:02 AM
@FaheemMitha plain users can do whatever they want (that's the plain way) although luaotfload works in plain as well, so probably they mostly use that,
 
@DavidCarlisle I see.
So, is Lua memory management problematic? Or is the amount of memory required just unreasonable?
 
@FaheemMitha huh?
 
1 hour ago, by Faheem Mitha
> and memory usage in lua can be large as allocations doubles when more is needed
That fonts thing.
 
@FaheemMitha why do you think that it is a problem? In luatex memory is dynamic, unlike pdftex where you have to change the texmf.cnf if you need more.
 
@UlrikeFischer I don't know if it is a problem.
But are the memory issues just because things (specifically the font files) are too large?
 
8:16 AM
@FaheemMitha yes, some fonts are so large that the memory available for a 32bit luatex on my system wasn't enough to create the lua, but a 64bit luatex did manage it. But this memory is only needed once for the initial use of the font.
 
@UlrikeFischer Is that required to generate font files?
 
8:35 AM
@FaheemMitha before I was online:-)
 
8:45 AM
@texdr.aft there is no real program pointer to goto, it's just a macro language so jumping forwards is just \@gobble and jumping backwards isn't really possible.
 
8:57 AM
@JosephWright in beamer, are the beamer options also passed on to beamer themes? or are they somehow separate? I'm playing with a theme, and wanted to give a warning if the user is not using a wide aspectratio (I could of course just test on the \beamer@paperwidth/-height
 
@DavidCarlisle You could read the file back and step through
@daleif Themes are loaded using \usepackage internally; they don't automatically get the generic beamer option list. You'll likely have to check for the underlying variable values
 
@texdr.aft What's your use case?
 
@JosephWright feature request: have beamer store the value of aspectratio (it only sets the paper size) via a case list
 
@daleif All the options are keyval, so I think it's available as \KV@beamer@aspectratio (need to trace through)
 
Who is the current beamer maintainer?
Wasn't it written by the creator of PGF/TikZ?
 
9:11 AM
@JosephWright you could do, but would you?
 
yo'
@FaheemMitha AFAIK, @JosephWright
 
@yo' Oh.
 
@DavidCarlisle No idea ;)
@FaheemMitha I've ended up 'holding the baby', largely by accident, though I get assistance from a chap called Louis Stuart (at least that's who he is on GitHub!)
@FaheemMitha Till Tantau wrote it, then handed it over to Vedran Miletic and (lesserly) me, before Vedran also moved on leaving me at least trying to fix serious bugs
 
@JosephWright That sounds like a rough gig. That's a huge package with a ton of users.
So Till Tantau is not involved at all now?
 
@FaheemMitha No, not for some years
 
9:18 AM
@JosephWright that is just the macro definition. It does not seem to actually store the current value (which make sense). I'll just test on the paper instead
 
@FaheemMitha I only got involved to fix one urgent issue, really; I avoid making changes beyond clear bugs as there is a lot of code with basically no comments. It's hard to know why Till did many things.
 
@JosephWright Have you had any contact with him? Tried to get him to take an interest? Maintenance is often a problem for free software projects when the creators leave.
 
@FaheemMitha Er, no, Till seems to be happy to leave things to me/Louis
 
@JosephWright Ok.
 
@FaheemMitha We did get given some money by DANTE a few years ago, Till took the lead there :)
 
9:25 AM
@JosephWright What year was that? Is Beamer on Github now too?
 
@FaheemMitha Er, I think 2015
@FaheemMitha Yes, beamer is on GitHub; Vedran had it on BitBucket, but that repo is dead (I can't remove it, so it's sort-of stuck)
 
@JosephWright Ok. Thank you.
Looks like samcarter has been doing quite a lot too.
 
9:51 AM
@FaheemMitha Yes
 
10:26 AM
@DavidCarlisle ooh
 
 
2 hours later…
12:15 PM
Excuse me everybody. But the geometry package works correctly to you?
I had the correct margins before and I find myself with margins that cannot be changed.
 
@Sebastiano Yes
 
Joseph hi
look here
 
@Sebastiano Without some example of what you are doing it's impossible to help
 
@JosephWright It is my recent question.
 
yo'
@Sebastiano that's for recto/verso pages, probably. It's the standard behaviour, see the geometry documentation.
 
12:20 PM
yo' has always work well. Look here
 
@Sebastiano Which one?
 
1
Q: One or more equations (two-by-two on different rows) with an enumerate to the left and the another on the right

SebastianoStarting from these MWE into two links How to place and number 3 short equations in one line? (Stefan Kottwitz's answer) with this MWE: \documentclass[11pt,a4paper]{scrbook} \usepackage{amsmath} \usepackage{tabularx} \begin{document} \chapter{Equations} See equations \eqref{eqn:1}, \eqref{e...

See my folder. I still haven't processed the old one. See the .pdf file of the 9-8-19. I have done recent changed but I have not touch the margins.
Yesterday works correctly.
\ProvidesPackage{structure}[2019-01-01]

%----------------------------------------------------------------------------------
%	VARIOUS REQUIRED PACKAGES AND CONFIGURATIONS
%----------------------------------------------------------------------------------

\usepackage[top=2.5cm,bottom=2.2cm,
            left=3.2cm,right=1.5cm,headsep=10pt,
            a4paper]{geometry}  % Page margins
I'm working cutting the pieces of my long work.
What is happen? :-(
The succesive image is taken from my old version of 9-8:
The margin are correct before :-(
 
1:00 PM
@user121799 you're contributing again?
 
1:22 PM
@Sebastiano Please make a full example and post a question
 
@JosephWright @yo' Hi, again, after 1 hour I have solved my problem. There is a missed "oneside". grrr :-( Thank you very much to everyone for your collaboration. I have almost 15 different structure
 
^^ @JosephWright @UlrikeFischer Arthur's talk?
 
@DavidCarlisle oh my god. double harfbuzz ;-)
 
@UlrikeFischer I assume just one that is, rip the middle out of xetex, start from luahbtex then add back the missing xetex bits....
 
@DavidCarlisle Oooh
@DavidCarlisle Yeah ....
@DavidCarlisle, @UlrikeFischer Obvious question is why to all of this
 
1:36 PM
@DavidCarlisle perhaps we should write to more people "there can be only one".
 
@JosephWright more context is here:
@JosephWright yes I (think) the luahbtex + some suitable lua fillins can cover all of xetex except character classes, it would be easy to put them back as an engine extension but they are pretty weird really and only used in xetex as that's the best hook there is there
 
@JosephWright and who they expect to support this zoo.
 
@UlrikeFischer We need a video with a voiceover by Sean Connery ;)
@DavidCarlisle Quite: really, there is nothing significant that LuaTeX can't cover, and even interchar stuff is fixable
@UlrikeFischer I think we know that ...
 
@DavidCarlisle ooh
@DavidCarlisle Not Arthur, there's no hat. :)
@UlrikeFischer ooh
@JosephWright ooh
@JosephWright ooh I know who will be the maintainer!
 
1:52 PM
@JosephWright or this one youtube.com/watch?v=sqcLjcSloXs
 
@JosephWright I think meeting the real use cases is OK, making a drop-in replacement looked a bit hard last time I looked (input buffer callbacks a bit early and node list callbacks too late)
 
2:15 PM
@barbarabeeton how was the conference?
 
@barbarabeeton welcome back! <3
 
@DavidCarlisle -- I enjoyed it. My favorite talks were by David Fuchs (fully interactive TeX) and Tom Rokicki (changes to dvips to allow it to support searching of output files after conversion of ps to pdf), both based on Knuthian TeX. (Showing my biases here.) Don was present for that day too.
 
@barbarabeeton yay
 
@PauloCereda -- Thank you. Have to get off soon; need to finish packing. But I'll be back later; we're just shifting our location to San Francisco for a few days before going home.
 
@barbarabeeton you claim allegiance to Knuthian TeX but we know you are a Unicode devotee really
2
 
2:25 PM
@DavidCarlisle ooh
 
@DavidCarlisle -- Dealing with Unicode is a sure way to make one even crazier than how one started out.
2
 
@barbarabeeton is something known about TUG2020?
 
@barbarabeeton ooh
@UlrikeFischer ooh Germany Germany Germany
 
@UlrikeFischer I hear that it will be next year
 
I'm just trying to research the answer to the question at tex.stackexchange.com/q/503763 online. Trying to find the relevant manual in I-don't-know-which-box at home will delay it too long.
 
2:28 PM
@DavidCarlisle unless it is like latex3 then probably 3020
 
@UlrikeFischer could be.
 
@DavidCarlisle ooh integers
@UlrikeFischer ooh more integers
 
@UlrikeFischer -- We got a credible proposal from Rochester Institute of Technology, and it's been accepted. (That's Rochester, New York.) Host will be the Cary Library, which is really superb, with participation by Chuck Bigelow and Kris Holmes. Dates not known yet, since RIT hasn't yet finalized their summer schedule, but location is fixed. Proposals for 2021 et seq. are welcome!
 
@barbarabeeton this pdf makes interesting reading:-) stacks.stanford.edu/file/jy605yq4819/jy605yq4819.pdf
 
@DavidCarlisle -- Yes. That's what I'm looking at. (But I really have to finish packing. Back later.)
 
2:35 PM
@barbarabeeton OOOOOOOH
@barbarabeeton I can visit Sacred Heart Cathedral where Bishop Sheen served! YAY!
 
yo'
3:20 PM
@barbarabeeton I'll see what my schedule in 2021 is, but I would be willing to offer Prague. The only thing to sort out would be that I'll have no official contact with any of the academic bodies.
 
@yo' WOOOO
user image
5
@AlanMunn ^^
 
@PauloCereda yess!!
 
@HaraldHanche-Olsen :)
 
3:56 PM
user image
3
 
4:18 PM
@cis Not sure if you found this already, but dim(\List). I think it's undocumented though, marmot mentioned it recently in an answer.
 
@yo' -- That's an attractive prospect! Do keep me posted. (Petr Sojka and his son were at the conference, and gave a very interesting presentation. Even though Petr is in Brno, and that's not exactly next door, he might have some ideas. And I believe that Jiří Zlatuška, who was Thanh's thesis advisor at Masaryk University, was in Parliament for a while, and is now back at Masaryk as Dean of the Faculty of Informatics.)
 
Hello in $\vert \Vert x\Vert -\Vert y\Vert\vert $ how to write the absolute value. larger ?
 
@barbarabeeton I suppose you're going? I admit to being a bit envious. It's all on the wrong side of the puddle.
 
4:35 PM
@PolineSandra you should use lVert and rVert not Vert but to get a bigger version \bigl\lvert \lVert x ....
 
5:07 PM
@HaraldHanche-Olsen -- Now that I'm retired, I can spend more time wandering around, and intend to.
 
 
2 hours later…
Kind of an urgent question. I noted changes to a manuscript with {\color{red}next text}. Now I need a version without the red. Can I just replace \color{red} with nothing and leave the {} in the text or will that mess things up? It appears to work fine, but it is creating problems I am just not noticing? Entire paragraphs are in red....
 
@Canageek should be fine, if you really want to make sure that you turn off colour without changing any spacing, just leave everything and use the monochrome option to the color package.
 
@DavidCarlisle Sadly I use the \color command elsewhere in the paper. In future I should make an alias I can just disable, but I didn't think of that
 
@Canageek the group can't really hurt (well It is possible to make edge cases where it will do something but in those cases it would already have affected things with the red colour)
 
Thanks
 
7:07 PM
@Canageek you can also use latexdiff and it will find the changes for you.
 
7:27 PM
@PauloCereda I am "well prepared" (but only got 6 out of 12 :-)
 
@DavidCarlisle ooh :)
 
@DavidCarlisle I got them too ;-)
 
7:42 PM
@DavidCarlisle I'm well prepared too.
 
I just upgraded to TeX Live 2019. It's now saying LuaTeX 1.10. That's Lua 5.3, right? But it doesn't say anything about Lua 5.3 if I invoke it.
 
@FaheemMitha \directlua{print(_VERSION)}\bye
 
Yes, it says Lua 5.3. Thanks.
 
@egreg we could have a party with @UlrikeFischer but not invite @PauloCereda as he's unfit
 
7:56 PM
@DavidCarlisle oh
 
However, my Lua file is still using 5.2 paths, but the compile is working. Shouldn't it error out?
And in any case, should I update to 5.3 paths?
 
@FaheemMitha you are pulling in files from a stock Lua installation? That's brave.
@FaheemMitha Most Lua 5.2 code still works in 5.3
 
@DavidCarlisle Brave, possibly. Unavoidable, definitely. At least if I want to do anything with Lua.
Are you saying this isn't supported?
@DavidCarlisle Sure. I think in nearly all cases, it's the same code.
Oh, Lua 5.3 is ok getting code from Lua 5.2 paths? Why not, I guess.
But I guess I should switch to 5.3 paths?
 
@FaheemMitha integers being the most obvious difference.
 
@DavidCarlisle Right. But most code doesn't use it yet, I think.
Isn't 5.3 relatively new?
 
8:00 PM
@DavidCarlisle my husband's comment:
 
@FaheemMitha What are you up to? Lua's a lightweight scripting language, most of the time you can quickly re-code any functions you need
 
@FaheemMitha no but previously a literal such as 2 was a float, and now it's an integer
 
@JosephWright Just trying to use Lua libraries. Nothing fancy. E.g. LuaSQL.
@DavidCarlisle That's an improvement, I guess.
 
@FaheemMitha presumably, I have never done such a thing myself.
 
@DavidCarlisle Ok.
 
8:01 PM
@FaheemMitha yes but it can easily break existing code.
 
@DavidCarlisle True.
 
@UlrikeFischer by the way, lesson learned this weekend, never do what your wife suggests, bad things happen.
 
@DavidCarlisle ?
 
@DavidCarlisle I always thought that people do what their wives say, otherwise bad things happen.
Not a Shakespeare fan?
 
8:07 PM
@FaheemMitha Play was good, but Lady Macbeth isn't the world's nicest person:-)
 
I once politely suggested that Shakespeare was overrated on an internet forum. And got told off by an upset Shakespearean fan. An American, as it happened.
I'm not sure why she cared what I thought.
@DavidCarlisle If you find her unpleasant, much of modern media must also be unappealing.
 
@DavidCarlisle ooh documentation
@DavidCarlisle ooh the Scottish play
@DavidCarlisle she has an antivirus company
oh wait that's McAfee
 
@DavidCarlisle ah, well one of the more interesting questions of the Scottish Play is whether Macbeth hadn‘t made up his mind already.
 
@UlrikeFischer I blame the Lady:-)
 
8:23 PM
@DavidCarlisle well you have practice here ...
 
8:41 PM
@UlrikeFischer exactly:-)
 
You are in blood
Stepp’ in so far that
Should you wade no more
Returning were as tedious as go o‘er.
(III, 4)
 
@UlrikeFischer longtable support:)
 
@DavidCarlisle ;-)
 
9:21 PM
What is l3build trying to tell me here? It seems like it can't find it's own files and I can't reproduce the problem locally.
 
@MarcelKrüger hm perhaps the first test failed (aaaaa-luakern is a bit manual, I should correct this...) and the code to show the log-file doesn't work on your system.
 
@UlrikeFischer Thanks, it was the first test. I forgot that I have a changed unicode-data version.
 
@MarcelKrüger I made the dev version. It seems to work fine, but curiously changing caches_dir in the configuration didn't work, I had to change caches.namespace.
 
In the TeX Live Debian packaging, everything is 2019.20190710-1, except texlive-binaries, which is 2019.20190605.51237-2. Would anyone have a clue why?
 
9:39 PM
@UlrikeFischer I just looked into it. These config options are only used by luaotfload --cache=..., otherwise the ConTeXt variables are used. We should probably change that...
 
9:50 PM
@MarcelKrüger names_dir worked, but imho caches.namespace is better anyway, creating a generic-dev is much saner.
 
@UlrikeFischer Right. Fixing the caches_dir problem would requireeither patching the fontloader or doing weird rewriting in -basics-gen, which still is kind of part of the fontloader. What do you think about just removing the caches-dir configuration option?
@FaheemMitha I don't know the official reason, but I think it makes sense: Normally in TeX Live the binaries never change between 2019 and 2020, so if Debian doesn't need to rebuild them for internal reasons there is no point in updating the binary package. The remaining parts of TeX Live do get regular updates, so they need more releases.
 
@MarcelKrüger well it doesn't work anyway ;-). I just trying to decide how to sort and name the git branches without making a mess. (dev-version should always end with -dev). Perhaps I should make a ctan upload with what is now done, to have a clean state.
 
10:11 PM
@MarcelKrüger I see. Thank you for explaining.
 
@UlrikeFischer About branches: I think it would be nice to introduce one branch which always points to the latest changes. It would be okay if this has to be rebased or reset every now and then, but one branch which can always be tracked to find all the new stuff, even before it reached the -dev version. (Maybe next-dev?)
 
10:35 PM
@MarcelKrüger I think that would ok. I like to know the version numbers, but since the tagging works and all files have correct declarations the branch names are less important. I will try to make a ctan upload and the start from there - it fits It will get version 3.
 
10:55 PM
@UlrikeFischer Great. I also wouldn't have a problem if there are additional branches with version numbers or we use lots of (temporary?) development tags such that git describe --tags gives more helpful output.
 
11:41 PM
@FaheemMitha I don't have one. I was just curious
@DavidCarlisle Of course, you could do something like this: gist.github.com/texdraft/fb93aa8edca268f657895ff8c4b529b0
 
@DavidCarlisle -- Can I come to the party? I got 7, so I'm well prepared for a visit.
 
@texdr.aft well yes but not really in the spirit of things, as it's not jumping in the main tex flow at all, so it's no different from say referencing the basic interpreter that someone encoded and then saying basic has goto. But also note \ifnum\pc=3% should be \ifnum\pc=3 % or bad things happen:-)
@barbarabeeton 7! you can be the queen of the ball:-) Oh by the way congratulations I see you just became a life member (ie, there is no escape:-)
 
@DavidCarlisle Fair enough. Also, thanks for the note about the comment. It works for now, but I'll keep that in mind.
 
@texdr.aft yes because your following tokens were not digits, try it with 1 2 3 4 instead of A B C D
 
@DavidCarlisle Oh I see. I guess I got lucky then. Thank you again.
 

« first day (3207 days earlier)      last day (1727 days later) »