« first day (2869 days earlier)      last day (2063 days later) » 

3:58 AM
@marmot Our discussion about clipping and the marmot transparency made thinking: tex.stackexchange.com/a/449947/36296 :)
 
@samcarter Very nice! +1. ;-)
user image
2
^^^^ that's what I am just hoping to finish: closed 3D coil. trickier than I thought because of all the little rotations....
 
@marmot Thanks :) The coil looks very nice! I can already imagine the marmot jumping through it. But now I'm back to bed.
 
@samcarter \pgfsetblendmode{multiply} is really smart
@samcarter @AlanMunn is in the news ;-)
 
 
2 hours later…
5:45 AM
I think I may have discovered the whereabouts of @PauloCereda …
 
 
1 hour later…
7:14 AM
@HaraldHanche-Olsen Oooh, that's why!
 
 
3 hours later…
9:47 AM
@mico it's just that new users might take that the wrong way. I know it was a bad question, but most of the users don't read up on any of the guidelines on how to ask a question before posting their first. I read your comment and had to laugh, but it's a humor I developed roaming this portal.
also, people in here still assume when i try to make a joke on something ducky, that I just didn't understand something correctly and either treat me as if I signed up yesterday or ... well, that's it, actually. :D
 
 
1 hour later…
10:57 AM
@JosephWright care to post an answer:-) tex.stackexchange.com/questions/449869/…
 
 
1 hour later…
@HaraldHanche-Olsen you can't trust politicians to get anything right
 
12:22 PM
@DavidCarlisle Done: it's a string test so relatively easy
 
@JosephWright The robustify solution you posted here does not work
 
@Krishna Can you post a screenshot
 
@JosephWright Sure, I could. This will just lengthen the question without adding meat to it. I This is a direct link to a public overleaf document which shows the code and the rendering. Is this okay? overleaf.com/19349431rysdtbdjwcxm#/72086837
 
@HaraldHanche-Olsen This island urgently needs a better internet connection, I'm missing the ooh counter!
 
12:30 PM
@Krishna ^^^
 
That's what I got.
 
@marmot With multiply it was surprisingly easy. My first idea was to use complementary colours and then create the image two times, once with additive and once with subtractive colour mixing in the hope than upon combination the right bits would cancel each other out but luckily this was not necessary.
 
And, what I want is not the same as your screenshot either. The desired thing is soomething else. The leading tablenotes should stay close to first digit, while the columns are still lined up at the decimal point
 
@Krishna table-align-text-pre = false does that
 
@JosephWright Nope, still not working
Bernard thinks this is a bug with table-align-text-pre = false
In his words, "what's not working is table-align-text-pre=false (I tried it, but as it didn't work, I didn't use it in the code I posted). Maybe there's an interaction with the S column type."
 
12:37 PM
@Krishna Something may be up here, but you definitely have an old version causing issues too
 
@JosephWright Well, I religiously update my TL distribution twice a day :)
Wait a minute, this is overleaf, which uses TL2017...... I am going to try the table-align false flag in my local computer
 
@marmot Out of curiosity: in your coil ring, how do you make sure that the start and end will line up and nothing like this happens:
 
Hooray! That actually worked. Thank you.
So, sometime between TL2017 release and today, this was fixed.
Overleaf currently uses TL2017
From their history for the last 3 releases, it looks like Overleaf seem to update their docker images to the latest TL only in November
 
Good Morning friends!
Quick question for expl3 experts: Does expl3 have some number formatting function?
Something like a \int_print_leading_zeros:nn { 1 }{ 3 } which prints 001...
 
12:53 PM
@PhelypeOleinik no, as far as I know
 
@boycott.se-yo' Ooh, I was really hoping otherwise U_U
 
@PhelypeOleinik but what you ask for is not that difficult to achieve
 
I guess that that something like two@digits \int_digits_two:nn will have to do it for now :P
 
@PhelypeOleinik why two parameters? And remember that you are not allowed to use the \int_ namespace. Get your own namespace
 
@boycott.se-yo' I did something comparing the value of the number, but I thought that there was something built-in.
@boycott.se-yo' I didn't name it that way. Just kidding ;)
 
1:01 PM
@PhelypeOleinik ok
 
@boycott.se-yo' Thanks :)
 
well, I would use something like \prg_repliacate:nn{\int_eval:n{5-\str_count:o{\int_to_arabic:n{#1}}}}{0}#1
where you have to \cs_generate_variant:Nn \str_count:n {o} and where I many errors on the way
@PhelypeOleinik ^^
 
1:21 PM
\documentclass{article}

\usepackage{expl3,xparse}

\ExplSyntaxOn
\cs_generate_variant:Nn \str_count:n { f }
\NewDocumentCommand\FormatInteger
  {
    m m
  }
  {
    \int_set:Nn \l_tmpa_int
    {
      #1 - \str_count:f
        {
          \int_to_arabic:n { #2 }
        }
    }
    \prg_replicate:nn
      { \l_tmpa_int }
      { 0 }
      #2
  }
\ExplSyntaxOff

\begin{document}

\begin{tabular}{rr}
\FormatInteger{5}{1}     & \FormatInteger{1}{1} \\
\FormatInteger{5}{11}    & \FormatInteger{2}{1} \\
@boycott.se-yo' Thanks a lot :D
 
@PhelypeOleinik this is not expandable, but you probably don't need that.
 
@boycott.se-yo' Not right now. I couldn't get it working by nesting the \int_eval:n in \prg_replicate:nn when I changed the signature of \str_count to :f
@boycott.se-yo' And I did change that because apparently the signature :o counts \int_to_arabic:n as a bunch of characters, then \int_eval:n would return something like -45
 
1:41 PM
@JosephWright rebuilding oberdeik bundle (again) (takes about 45 minutes or so :(
 
\ExplSyntaxOn
\cs_generate_variant:Nn \str_count:n { f }
\NewExpandableDocumentCommand\FormatInteger
  { m m }
  {
    \prg_replicate:nn
      {
        #1 - \str_count:f
          {
            \int_to_arabic:n { #2 }
          }
      }
      { 0 }
      #2
  }
\ExplSyntaxOff
@boycott.se-yo' Expandable now :)
@boycott.se-yo' I really don't know why it didn't work earlier :/
 
@samcarter spirograph when the pen slips out of the wheel :-)
2
@JosephWright voted:-)
 
@samcarter There is an easy part: make sure that there is no fractional coil. This has been solved by Jake in tex.stackexchange.com/a/25689/121799, that's why it's easy. The more tricky part is to make sure that all, i.e. also the first and last coil, are 3D-like, i.e. have a white boundary.
 
2:17 PM
@marmot Oh, that is a pretty neat trick!
@DavidCarlisle Indeed, that is a very good description.
user image
2
@marmot The marmot is waiting to jump through the coil ring :)
2
 
@samcarter Yes, I am working on it. ;-) But the problem is that the shifts are in tangent space. So I really do not know how to make it work for all possible amplitudes and segment lengths...
 
2:32 PM
@marmot Maybe the trick is to just restrict the possible amplitudes and segments lengths to the ones that are working :)
 
@samcarter Well, the thing is that I am almost there.
I think it would be a nice thing to have if it worked. (I played with these things in tex.stackexchange.com/a/447546/121799 and there it seems to work. Really the only problem is to make sure that the last tiny bit is also 3D like, but this seems to be tougher than expected...)
 
3:02 PM
@Krishna Ooh, tlmgr is a religion, now?
 
@HaraldHanche-Olsen reserve such descriptions for emacs
 
3:30 PM
@PhelypeOleinik oh yeah, I think two expansions are needed. Safest would be simply :x
 
What is the best way to get a box (stored with \setbox ) into a l3 coffin? I'd use \hcoffin_set:Nn \l_tmpa_coffin { \box_use:N \l_tmpa_box }, is this ok or should I do it differently (performance should be a consideration)?
 
@boycott.se-yo' It breaks if I change :f to :x though. I tried to trace the issue but couldn't find out why :/
 
3:58 PM
@PhelypeOleinik oh, well, don't ask me :)
 
4:48 PM
@Skillmon Looks fine
@PhelypeOleinik Use f-type expansion, or the new e-type perhaps
 
@JosephWright I used the f-type. Why doesn't the x-type work here?
@JosephWright And what does the e-type do? I didn't see it listed in interface3...
 
@PhelypeOleinik Depends on the details
 
@JosephWright Is there a faster way than regex_replace_all to turn the ^^M of multi-line verbatim input into actual new lines?
 
@PhelypeOleinik e-type is \expanded, currently only in LuaTeX but emulated by Bruno in other engines and to-be added to pdfTeX/upTeX in TL'19 (XeTeX outstanding ...). It does the same as x-type except it is expandable itself and doesn't double # tokens. Very handy!
 
@Skillmon what do you mean by an actual newline? (usually after verbatim ^^M is an active token defined to be \par
 
4:55 PM
@JosephWright Looks interesting :)
 
@samcarter Done: i.stack.imgur.com/SOFYK.png , see tex.stackexchange.com/a/449706/121799 (things can be so easy... if one does not try the wrong thing. ;-)
 
@DavidCarlisle \NewDocumentCommand\foo{+v}{#1} and does return all the new lines as ^^M (the character ^^M) displayed by an apostrophe in the document.
 
@PhelypeOleinik Yup: I just need to prod Jonathan Kew again at some point, or ask Karl to add all of our stuff directly to XeTeX in TL ...
@DavidCarlisle On the HO stuff, I have a feeling we may want to roll some stuff into the kernel or expl3, and leave the original names as stubs ... I'm thinking in particular of some hooks
 
@Skillmon oh that sort of verbatim
 
@DavidCarlisle sorry, should have been more precise.
@DavidCarlisle currently I'm doing this with \regex_replace_all:nnN { \r } { \c{newline} } \l_tmpa_tl
 
5:00 PM
@Skillmon no I should have guessed, but 2e for me is modern, so didn't think of l3 (despite the immediately preceding comments:-)
 
@Skillmon You could code it up using the \tl_replace_all:Nnn function, but to be honest unless you need really fast performance, I'm not sure it's work it
 
@Skillmon that's scarily inefficient (@JosephWright do we have a supported way to do that?)
 
@DavidCarlisle Not at the moment: we've not exposed end-of-line stuff
 
@DavidCarlisle thanks I know, just load ducksay and you see how inefficient it is :)
 
@JosephWright would be nice to have an option for v to return ends of lines as something definable, a non-active ^^M isn't very friendly.
@Skillmon add \tracingall and see how long your log file is...
 
5:03 PM
@DavidCarlisle boy this needs long to complete :)
still at it
 
@DavidCarlisle Perhaps a processor might be best: that's the 'right' place in xparse
 
@Skillmon you might need to research bigger disk drives on amazon
3
 
Aaaaaand it's done
 
@JosephWright did you see my catcode table fix for the lua shell-escape test, did you have similar code in expl3?
 
Output of wc for \documentclass[]{article}\tracingall\usepackage{ducksay}\begin{document}<++>\en‌​d{document}:
 
5:05 PM
@Skillmon @DavidCarlisle I'll raise this at the next LaTeX team meeting: we might need a few ideas, but I suspect an xparse interface to sort this is doable
 
wc -l: 14854175
wc -c: 483512839
 
@DavidCarlisle I've missed that: I'll take a look (may need to add ...)
@DavidCarlisle Will look in a bit: just about to do some shopping and head to my flat (have been at 'real' work)
 
The file is so large it took my VIM ~3 seconds to open it O.o
 
@JosephWright probably not many people (other than Heiko) are doing a shell-escape query at the point that 0 has catcode 9, but returning always a catcode 12 0 seemed safest.
@Skillmon if you will use a second class editor
 
@DavidCarlisle Ah, probably not but we can be defensive, certainly
 
5:10 PM
@DavidCarlisle took emacs a few seconds as well (and it asked whether I really want to open it -- stupid OS shouldn't ask me if I call it with emacs ReallyBigLogFile.log)
 
@JosephWright actually the error message was worse as I wasn't returning 0 from the directlua but \numexpr0\relax which doesn't work too well if 0 has catcode "ignore"
 
@DavidCarlisle also an emacs installation takes 127.33 MiB, while VIM only takes 3.68 MiB. I guess that's small for an OS, though.
 
@Skillmon yes a palace is usually bigger than a garden shed.
 
@DavidCarlisle and efficient software is usually smaller.
@DavidCarlisle just compare installation size of Windows (highly unefficient) with any Linux Distro (even the bloated ones are much smaller).
 
5:15 PM
@DavidCarlisle I don't know for sure, but I guess that notepad.exe is bigger than VIM while having only a fraction of the features.
 
@Skillmon if emacs is too complicated for you, you should try this, I think you will get on fine, and it's a lot more usable than vim.
 
@DavidCarlisle it's not, it can't even display line breaks right.
@DavidCarlisle but it might be a good enough editor for your bloated OS called emacs.
 
@Skillmon actually they fixed that in this year's update (along with adding curl and tar to windows:-)
 
@DavidCarlisle Microsoft fixed anything? I'm stunned. Probably they introduced 8 new bugs by fixing this one.
@DavidCarlisle 3 of which not even related to notepad.exe
 
@Skillmon I expect so, but it's progress. Or it would be progress except i never use notepad and never use the windows command line, and I already had curl and tar in the linux subsystem and in cygwin :-)
 
5:20 PM
@JosephWright It doesn't seem so (using ^^M directly in the source code breaks it -- no idea how I'd have to do it)
@DavidCarlisle I'm still shocked by the thought that you have to use Windows every day, just so you can start the Linux subsystem.
 
@Skillmon actually I don't use the linux subsystem much but I use cygwin all the time. But I keep an eye on the subsytem development as it may be a viable replacement for cygwin at some point. Basically as a normal day to day environment I run cygwin and X and windows is mostly just acting as my X window manager, apart from a few native windows apps such as this browser.
 
@DavidCarlisle That's better, but still not good, imho. What a waste of resources
 
@Skillmon actually the resources are not wasted, the machine isn't mine, and all our development is Make based but many of our customers are on windows, and I find it easier to have a fake linux environment in windows than run linux and have a fake wine or some such environment. Last time I had a machine of my own, it ran debian linux.
 
@DavidCarlisle It's likely a good working environment. I fear the day I leave university and have to use the IT infrastructure of the company that's willing to employ me. I for myself can't work anymore under Windows, it just feels highly inefficient. But maybe I'm just a super special sort of snowflake.
 
@Skillmon compared to most companies, we have a remarkably free choice, I could run linux, but I choose this cygwin setup, and I can install stuff on my own machine, in many companies people wouldn't be able to install tex (or compile luatex from sources or whatever on a work machine)
 
5:32 PM
@Skillmon I was thinking at the 'use forbidden primitives' level: we will need to sort an interface for 'you'
@Skillmon Depends on what you are used to: like @DavidCarlisle, I have Windows plus 'stuff' (work software is all Windows-only, and I have to have Word on Windows)
 
oberdiek bundle updated, phew I hate that thing:-)
 
@JosephWright I don't get why big companies spend so much money on MS software.
 
@Skillmon Market leader and all that, plus there are a lot of Windows-only tools (most of our instrument and analysis software, Outlook, ...)
@DavidCarlisle We need to think about this, longer-term
 
@JosephWright yes:-)
 
@JosephWright Outlook might be a good piece of software (especially compared to other stuff created by MS), but certainly there are alternatives.
 
5:37 PM
@Skillmon Oh, sure: I don't use it :)
 
@JosephWright and the market leader stuff is just like saying: We do stupid stuff, because other people do stupid stuff. It shouldn't be that hard to write software which runs on multiple OSes (see almost all the big free software applications, all have a Windows version)
 
@Skillmon I do draw the line at using outlook (although it has been suggested a few tines:-)
 
@Skillmon Yes, true, but the reality is that many things are not, and the push if anything is for tablet versions
 
@PhelypeOleinik I saw you created a template for thesis and the like. Do you have plans on translating the documentation to English? Though I wouldn't use it (at least not any time soon), it is interesting to see what features others implemented who wrote a very specific template.
 
@Skillmon I will, as soon as I finish the one in portuguese.
 
5:47 PM
@PhelypeOleinik sounds good for me, thanks in advance!
 
@marmot Fantastic answer, I wish I could upvote a second time!
 
@samcarter How do you make the marmot run and jump?
 
@marmot You should sell this quote, would make great T-shirts, stickers, mouse pads ...
 
@Skillmon The template is for the postgraduate program I'm in, so the priority is for Brazilian users. As soon as I get the time to I'll upload a newer version with the documentation.
 
@samcarter There are already some nice alternatives: ebay.com/itm/MARMOT-Republic-Climbing-STICKER-Decal-New-/…
 
5:50 PM
@marmot One seconds, I'll move the code to github
 
@Skillmon One of the things I have to do before is to create an equivalent in MS Word /screams in despair
 
@PhelypeOleinik really? Why would that be your job?!
 
@marmot Please don't look at it too closely, instead of nice coding I opted for stupid copy&paste: github.com/samcarter8/tikzlings/blob/master/ViewingPlatform/… -- you also need the new latest marmot version: github.com/samcarter8/tikzlings/blob/master/…
 
@PhelypeOleinik I hope your Word skills are better than mine (mine are limited to the use case of a poor excuse for notepad.exe)
 
@Skillmon The whole "new template" thing started because I suggested it saying "the current template looks like the draft of the thesis" (See for yourself: drive.google.com/file/d/0B4YUv4YY3h1cNE9xZTIzai1HVkE/view )
@Skillmon Then I made the LaTeX one and when I handed it over I automatically got the task to make the MS Word one. I accepted because it's better than using the old template :P
 
5:55 PM
@PhelypeOleinik any fellow student/post doc who is fond of using Word who can do it for you?
 
@Skillmon Probably not much better. Last time I tried it, I spent hours trying to insert both a header with the name of the chapter and the page number.
 
@PhelypeOleinik last time I tried it, it took me more than 30 minutes to change a single picture.
 
@Skillmon There's a colleague of mine whom I helped with a paper. I'll ask her when I start this task
@Skillmon Then our skills are similar :)
 
@PhelypeOleinik does seem so (I had help in the end or else it would have taken me more than a day I guess)
 
@Skillmon After I started using LaTeX I noticed there are many trivial things that are basically impossible (at my level of knowledge) doing in Word, for example something as simple as \noindent\llap{Hello }World :/
 
6:03 PM
Is there a possibility to align 3 coffins in such a way, that handle1 of coffin1 is at the same height as handle2 of coffin2 and handle3 of coffin2 is horizontally aligned underneath handle4 of coffin3 (and they don't have to touch each other)?
 
6:16 PM
user image
3
@samcarter ^^^^^ ;-)
\begin{document}
\foreach \X in {0,...,35}
{\begin{tikzpicture}
\path[use as bounding box] (-4.5,0) rectangle (4.5,4);
\begin{scope}[shift={(2.5cm,0.5cm)}]
\duck
\end{scope}
\begin{scope}[xscale=-1,shift={(2.5cm,0.5cm)}]
\duck
\end{scope}
\ifnum\X>18
\draw[decoration={3d coil color=blue,aspect=0.35, segment length=3.1mm,
amplitude=3mm,3d complete coil},
decorate] (-3,1.5) to[out=0,in=180] (0,{1.5+1.2*cos(10*\X)}) to[out=0,in=180] (3,1.5);
\fi
\begin{scope}[yshift={max(-cos(10*\X)*1cm,0)}]
\ifodd\X
 
@marmot Awesome! This is so great!
 
@samcarter ^^^^ apart from the lengthy preamble the document is actual not too long. (With \ifodd and \foreach you can save some c & p.)
 
@marmot Yeah, I was just to lazy :) I could watch them forever, they are kind of hypnotising...
 
@samcarter Yes, but after all this is what TeX is made for: chat.stackexchange.com/transcript/message/46621215#46621215
 
6:53 PM
LaTeX3 is capable of cool error messages as well:
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!
! LaTeX error: "kernel/unknown-coffin-pole"
!
! Pole 't' unknown for coffin '\l_tmpa_coffin'.
!
! See the LaTeX3 documentation for further information.
!
! For immediate help type H <return>.
!...............................................
Seems like giving the poles inside of macros is not a good idea.
 
7:41 PM
@PhelypeOleinik Is this YAABNT template? (Yet Another ABNT template)?
 
@AlanMunn D: Nooooo!
@AlanMunn It goes as far from ABNT as I could get
 
@PhelypeOleinik Aiming for sanity then. :)
 
@AlanMunn If not for sanity, then for decency at the very least :)
 
 
3 hours later…
10:58 PM
Hello everyone
 

« first day (2869 days earlier)      last day (2063 days later) »