IRC channel logs

2020-08-14.log

back to list of logs

<chrislck>sneek: later tell janneke \o/ for some windows patches
<sneek>Okay.
***apteryx is now known as Guest78459
***apteryx_ is now known as apteryx
<wingo>moin :)
<RhodiumToad>mornin'
<RhodiumToad>I tested the armv7 abi fixes btw, seems good
***dddddd_ is now known as dddddd
***deesix_ is now known as deesix
<janneke>o/
<sneek>Welcome back janneke, you have 1 message!
<sneek>janneke, chrislck says: \o/ for some windows patches
<janneke>chrislck: in a minute :)
*janneke sends mail
***rubdos_ is now known as rubdos
<wingo>RhodiumToad: great!
*wingo landed switch optimization, yay
<wingo>case dispatch should be much faster now
<janneke>yay!
<chrislck>It's Friidayyy \o/
*chrislck knows nothing about cross-compilation... are things like compiling win32 & arm in ci possible?
<str1ngs>hello, is there anything like elastic tabs for guile? http://nickgravgaard.com/elastic-tabstops/
<str1ngs>
<RhodiumToad>wingo: btw there's still one bug that kicks in on 32-bit freebsd
<RhodiumToad>gen-scmconfig isn't coping with the case where off_t is in fact an int64, but HAVE_STAT64 etc. is not defined - it assumes that off_t is a long, which is wrong
<wingo>interesting
<RhodiumToad>freebsd (as with all 4.4BSD derivatives) doesn't have any of the "largefile" nonsense, because off_t is 64 bits even on 32-bit systems
<RhodiumToad>so there is no need for 64-bit variants of system calls
<dsmith-work>Happy Friday, Guilers!!
<chrislck>\o/ it's not Friday until dsmith says so!!
*sneek does the Friday dance
<janneke>chrislck: not sure what you mean by 'ci', but this
<janneke>guix build --target=i686-w64-mingw32 guile@2.0 --with-source=guile=https://ftp.gnu.org/pub/gnu/guile/guile-2.0.11.tar.gz --verbosity=1
<janneke>builds an i686-mingw version of guile-2.0 using guix
<chrislck>:-o
<chrislck>ci = continuous integration... ie every guile commit tries a win32 build
<chrislck>would be _nice_
<janneke>yeah, well start as soon as 2.0.11 is released, it bitrotted after that :-/
<chrislck>easily done by hooking up to travis-ci.org
<janneke>you could setup a cuirass service to do that
<janneke>ah, could be / no idea
*chrislck not too hot on gnu vocabulary
<chrislck>the idea is to keep guile hackers on their toes
<janneke>yeah, right
<janneke>us hackers are slacking all day ;)
<janneke>(and night)
<chrislck>see https://travis-ci.org/github/gnucash/gnucash - builds both cutting edge distros (arch) and oldest supported (ubuntu-18.04)
<chrislck>if a commit breaks a build, travis yells at us
<chrislck>that's all it is
<dsmith-work>ci is great
<janneke>yeah that's nice
<dsmith-work>We had a team that was on the other side of the planet, doing mostly windows work, while locally was more linux. The linux builds would break regularly until we installed a ci system (jenkins).
<chrislck>so... guile doesn't have CI???
<dsmith-work>So before, with all the time lags, it could take days of communication to fix things.
<chrislck>travis is nice... it tests gnucash main repo *and* each committer's branches
<dsmith-work>And after, we would see the breakage and the fixes before we came in in the morning.
<chrislck> https://travis-ci.org/github/christopherlam/gnucash/branches has all of my experimental branches
<chrislck>janneke: if you know the nix command to build guile win32, then maybe I can try create a travis profile to CI it
<dsmith-work>The lightEning jit has ci, as it's hosted on gitlab.
<janneke>chrislck: ah, you could ask on #nixos, i don't know how to build for w32/mingw
<janneke>using nix
<a_v_p>Hello Guilers! I keep experimenting with Metabash, this time I'm trying to implement a 'diff' analogue that will allow me to compare two data streams.
<a_v_p>Currently I can do something like this: https://gist.github.com/artyom-poptsov/f76be508899d2f8645f032cf4682204c
<a_v_p>But I don't like the implementation : https://github.com/artyom-poptsov/metabash/blob/master/modules/metabash/diff.scm
<a_v_p>Suggestions are welcome!