« first day (2624 days earlier)      last day (2309 days later) » 

Dave Jarvis
Dave Jarvis
1:44 AM
@HenriMenke Would you be able to build a Pandoc installer for Windows 7?
 
Henri Menke
Henri Menke
@DaveJarvis Maybe. It's definitely harder than anything Linux-based. I'll try.
 
Dave Jarvis
Dave Jarvis
2:07 AM
@HenriMenke Ah, I thought you might have had a cross-compiler installed. Don't worry about it -- I'll run the program on a Linux box to convert the document over.
@HenriMenke Thank you so much, by the way. This is for a proof of concept: wiki (XWiki) to Markdown to PDF. So having beautiful tables will go a long way to showing people what's possible with separating content from presentation.
 
Henri Menke
Henri Menke
2:20 AM
@DaveJarvis I realised that pandoc uses the AppVeyor hosted continuous integration service to generate Windows build (unfortunately only x86). You can download the build artifact from here eventually.
@DaveJarvis That will take some more time though. The build is quite slow.
@DaveJarvis Follow it live: ci.appveyor.com/project/hmenke/pandoc
@DaveJarvis You're welcome. I always wanted to learn Haskell and this at least gave me a starting point. I recall, that you once asked another question about pandoc-generated tables in ConTeXt here. I think it was something about programmatically inserting a rule before the last line of the table which would also be fixed by my pull request.
@DaveJarvis Also, if you have other issues with the ConTeXt writer in pandoc, I'll be happy to take a look. I'm very interested in improving it myself.
 
Henri Menke
Henri Menke
2:49 AM
@DaveJarvis The build has finished. You can now collected the artifacts: ci.appveyor.com/project/hmenke/pandoc/build/1.0.1/artifacts
 
 
2 hours later…
Dave Jarvis
Dave Jarvis
4:32 AM
Since you asked... Besides tables that can be styled, the ConTeXt writer for pandoc could use captions and cross-references. Those features would allow for high-quality, beautifully typeset technical documentation from Markdown. There's a wealth of information about these, the proposed syntax, and so forth in the issue tracker. Here are some related links:

https://github.com/jgm/pandoc/issues/2106#issuecomment-342613369
https://github.com/jgm/pandoc/issues/2609
https://github.com/jgm/pandoc/issues/813
@HenriMenke Being able to produce custom environments (e.g., \startpoem and \stoppoem) is also one of the major missing features, in my opinion.
 
 
1 hour later…
Henri Menke
Henri Menke
5:37 AM
@DaveJarvis The section divs thing looks doable and general fenced divs shouldn't be a long way from there. The internal cross-refs are probably a bit too ambitious for me. I'm really unsure about the raw_tex issue, i.e. it might be either extremely easy to fix or ultra-hard with a million cornercases.
 
Dave Jarvis
Dave Jarvis
5:56 AM
The cross-references are definitely an ambitious chunk. I'm not too concerned about the `raw_tex` issue, merely noticed the issue. Here are some additional links to cross-references, should you decide to take up the torch:

http://galahad.well.ox.ac.uk/repro/#better-figure-and-table-captions
https://hackage.haskell.org/package/pandoc-crossref
https://github.com/lierdakil/pandoc-crossref
https://github.com/tomduck (Tom is a knowledgeable resource, too)
 
 
1 hour later…
Werner
Werner
7:04 AM
...interesting... to date you have 444 posts using the phrase "Like this"... — Werner 31 mins ago
 
 
1 hour later…
Paulo Cereda
Paulo Cereda
8:07 AM
Oh no
 
 
4 hours later…
NieDzejkob
NieDzejkob
12:09 PM
@anderstood I've generated all possible date "suffixes" using a Python script, exported it as hexadecimal and used Hashcat's mixed mode
 
Paulo Cereda
Paulo Cereda
@NieDzejkob @anderstood I broke it by writing a pure Java solution where I compared the hashs of a cartesian product against @egreg's hash and printed the solution. It took me about 2 minutes.
Not too efficient, but it was a fun exercise. :)
 
David Carlisle
David Carlisle
@PauloCereda one of the two minutes waiting for the jvm to warm up I assume?
 
Paulo Cereda
Paulo Cereda
@DavidCarlisle :) Surprisingly, newer JVM's are getting better start. :)
 
 
3 hours later…
David Carlisle
David Carlisle
2:58 PM
@JosephWright I think I broke something...
 
David Carlisle
David Carlisle
3:22 PM
... or maybe it was broken already
 
David Carlisle
David Carlisle
3:47 PM
         --> failed

  tlb0063
          --> failed

  tlb0073
          --> failed

  tlb0077
          --> failed

  tlb0084-2015
          --> failed

  tlb0084
          --> failed

  tlb0105
  tlb0109
  tlb0130
          --> failed

  tlb0143
          --> failed

  tlb0150
          --> failed

  tlb0152
          --> failed

  tlb0162
 
Joseph Wright
Joseph Wright
@DavidCarlisle Ooos
@DavidCarlisle Have nephew/niece here at the moment: will sort as soon as I can
 
David Carlisle
David Carlisle
@JosephWright no real rush, to fix just was surprised how much failed after I changed \@ifundefined, but then found it failed if I changed it back, but if that is same for everyone I can just carry on and just watch out for "new" failures when I change ifundefined
 
Joseph Wright
Joseph Wright
@DavidCarlisle Checking now
 
David Carlisle
David Carlisle
@JosephWright I'm using an l3build installed from head of gh l3build repo not the tl one
 
Joseph Wright
Joseph Wright
4:03 PM
@DavidCarlisle Ah, that might well do it
@DavidCarlisle I'm checking ...
 
David Carlisle
David Carlisle
@JosephWright I can uninstall that if that's the issue (rather than checking all the changed tlgs now)
 
Joseph Wright
Joseph Wright
@DavidCarlisle Yes but there should not be any change in the .tlg files that I know of: I'll track it down
 
anderstood
anderstood
@NieDzejkob So that's the hybrid attack I guess: hashcat.net/wiki/doku.php?id=hybrid_attack. I've got trouble running hashcat so I cannot try it. Thanks for the reply.
 
Joseph Wright
Joseph Wright
@DavidCarlisle Once that is done, we do need to look at activating the TU tests in the kernel: l3build can now deal with multiple set ups ...
 
anderstood
anderstood
@PauloCereda Did it take 2 min because you were lucky, or because your code is efficient? Just generating the dictionary of [phone numbers] [all dates of one month] takes 500Mb (1 million times 30 times 15 bytes, roughly), so 2 minutes would be the time to export it for me... And running hash on the "dictionary" takes a few ms, I tried both using sha256sum in bash, and Mathematica function.
 
David Carlisle
David Carlisle
4:10 PM
@JosephWright OK meanwhile I'll let this lot run to the end, save all the diffs then run again with a version of Bruno's ifundefined and check a diff of diffs
 
Alan Munn
Alan Munn
4:48 PM
What does 4 x 0.4\linewidth give you (plus the enumerations)? Does that exceed the current \linewidth? — Werner 1 min ago
Sometimes expertise is simply stating the obvious. :)
 
Joseph Wright
Joseph Wright
@DavidCarlisle OK, there is something up: I'll track it down (l3build not LaTeX2e)
 
David Carlisle
David Carlisle
@JosephWright Ok but if you'd rather play with your nephew/niece then it'll wait:-)
 
Joseph Wright
Joseph Wright
@DavidCarlisle I think I know where the change has occurred
 
Harald Hanche-Olsen
Harald Hanche-Olsen
5:11 PM
@anderstood @PauloCereda Clearly, @egreg should have used a Pedersen commitment. Or more simply: Added some random salt before hashing. Italians know about salt, don't they?
 
anderstood
anderstood
5:29 PM
@HaraldHanche-Olsen That's what he did with the phone number, isn't it? Anyway, I'm not interested in the date, rather in solving the problem in a reasonable time :)
 
Harald Hanche-Olsen
Harald Hanche-Olsen
@anderstood The phone number is too short (and non-random) to serve as a salt, that's the problem.
 
Paulo Cereda
Paulo Cereda
6:27 PM
@anderstood I can post my code at some time, if you are interested. It's more of knowing the "audience", that is, the domain of each element of the date. :)
 
mickep
mickep
Is it considered impolite to leave a comment to a question and to suggest against what OP tries to do, if you consider it to be bad typesetting (for example make something less readable)?
 
David Carlisle
David Carlisle
@mickep it depends (on how politely you phrase the comment:-)
 
mickep
mickep
@DavidCarlisle Something like "Are you sure you want to do this? I think it will make the text less readable."
 
David Carlisle
David Carlisle
@mickep it's probably OK, if it gets a bad response you can always delete it.
 
Stefan Kottwitz
Stefan Kottwitz
@mickep Phrase it so it sounds like an idea to think about, especially open and friendly, since written text can easily be misinterpreted (polite = patronizing). Treat Internet readers like kids being near the threshold to tears to be on the safe side. :-)
 
mickep
mickep
6:35 PM
@DavidCarlisle True.
@StefanKottwitz Thanks, I'll think about that
 
anderstood
anderstood
6:47 PM
@PauloCereda Sure, I'd be interested.
 
Joseph Wright
Joseph Wright
@DavidCarlisle I've worked out which commit causes the issue, I'm just not sure why yet!
 
David Carlisle
David Carlisle
7:08 PM
@JosephWright thanks
 
Joseph Wright
Joseph Wright
7:20 PM
 
David Carlisle
David Carlisle
7:53 PM
@JosephWright I helpfully changed the title:-)
 
Torbjørn T.
Torbjørn T.
8:07 PM
@DavidCarlisle Oh dear, the OP isn't happy. See comment on Herbert's answer.
 
David Carlisle
David Carlisle
@TorbjørnT. OPS SHOULD NOT SHOUT, ESPECIALLY IF THEY ARE ASKING FOR (OR DEMANDING) CODE TO BE WRITTEN.
 
Joseph Wright
Joseph Wright
@DavidCarlisle I've tided up
 
Torbjørn T.
Torbjørn T.
@DavidCarlisle Certainly.
 
 
1 hour later…
Alan Munn
Alan Munn
9:32 PM
@DavidCarlisle Perhaps the team could refund the money he paid for the software.
 
David Carlisle
David Carlisle
@AlanMunn We sent it already
 
Alan Munn
Alan Munn
@DavidCarlisle Now that's customer service.
 
David Carlisle
David Carlisle
9:51 PM
@JosephWright running test suite again, only test failed so far is a test that \@ifundefined doesn't define its argument to \relax .....
 
Joseph Wright
Joseph Wright
@DavidCarlisle Ah, good
 
David Carlisle
David Carlisle
@JosephWright I'll let this complete then put the changed defn back and run again. Currently I'm a bit worried about Bruno's luatex-specific definition I have put it in (as it's a lot quicker and a lot less convoluted for luatex) but I'm not sure if it affects any of the base tests but it will force luatex-specific tlg for any test that runs \tracingmacros past anything using \@ifundefined which may break some peoples tests....
 
Joseph Wright
Joseph Wright
@DavidCarlisle I was wondering about the policy on that: we have the same question for expl3, I guess, though there we could make the change
 
David Carlisle
David Carlisle
@JosephWright I should get it checked in tonight then will flag it for team once they have the code and/or test results to see
 
Joseph Wright
Joseph Wright
@DavidCarlisle Great
 
David Carlisle
David Carlisle
9:59 PM
just finished: 
  Check failed with difference files
  - ../build/test/tlb-ifundefined.etex.diff
  - ../build/test/tlb-ifundefined.luatex.diff
  - ../build/test/tlb-ifundefined.xetex.diff
so I claim if I save this buffer in emacs and re-run, they should all pass:-) (unless luatex bites us)
ooh difs (not tracing macros, box output...) @JosephWright
 
Joseph Wright
Joseph Wright
@DavidCarlisle I'll look forward to details
 
David Carlisle
David Carlisle
@JosephWright just blame Bruno:-) (missing \long :-)
 
Joseph Wright
Joseph Wright
@DavidCarlisle Oh that: I just 'read in' the \long ;)
 
David Carlisle
David Carlisle
@JosephWright it took me long enough to find all the hidden @F
@JosephWright although that explained a couple of the fails with errors not sure about the different box output but perhaps there were errors in the log not tlg, anyway re-running
 

« first day (2624 days earlier)      last day (2309 days later) »