IRC channel logs

<ArneBab>(luckily the development teams of most programs I need use Mercurial)

<mark_weaver>well, the thing is, someone has to create a proper commit for the VCS. if you don't do it, it means that someone else has to do it.

<ArneBab>need→need to contribute to

<ArneBab>I can create a proper commit and push it somewhere.

<ArneBab>hg-git works transparently

<mark_weaver>so there's no solution for creating something suitable for input to "git am" ?

<ArneBab>I do not know what git am requires

<ArneBab>…let me google

<mark_weaver>I guess I should be able to pull it from a git repo that you have hosted somewhere.

<mark_weaver>I haven't yet used git in that way, so I'd need to learn about it.

<ArneBab>no probs

<ArneBab>mark_weaver: here’s the commit you need: https://gitorious.org/guile/arnebabs-guile/commit/287c5514bfdb193a78c598d9382e5c0ed3ce3a49

<ArneBab>I need to go make some late-night housework now… (I promised my wife to carry some stuff up into the attic)

<mark_weaver>ah good, https://gitorious.org/guile/arnebabs-guile/commit/287c5514bfdb193a78c598d9382e5c0ed3ce3a49.patch is in the right format

<mark_weaver>okay, good night!

<ArneBab>you can also simply pull from git://gitorious.org/guile/arnebabs-guile.git

<mark_weaver>indeed

<dje42>Is there a builtin primitive to apply a procedure to each element of a list and stop when the result is true, returning that value (or #f if all invocations return #f) ?

<dje42>Easy enough with a named let, just wondering if there's something simpler.

<dje42>and-map?

<dje42>Errrr, or-map

<taylanub>dje42: I think that should be `any' from SRFI-1.

<taylanub>(Not a "builtin primitive" but I hope that wasn't really a requirement.)

<dje42>Ah. [no, I wasn't lookup for a primitive per se, just a utility]

<dje42>boot-9.scm has or-map, but any will do too

<stis>Morning guilers!

<wingo>greets!

<sneek>Welcome back wingo, you have 3 messages.

<sneek>wingo, nalaginrut says: Is this warning harmful? "fatal: Not a valid object name"

<sneek>wingo, nalaginrut says: I'm using tekuti for the blog

<sneek>wingo, ijp says: if by some chance you haven't already, check out "secrets of the glasgow haskell compiler inliner"

<wingo>happy new year &c!

<stis>happy new year& c to you too! Welcome!

<wingo>how goes?

*wingo was mostly away from the machine, but managed to fix some things in master in the end

<stis>cool! Wel familywize this has been a really nice christmas!

<wingo>excellent :)

<stis>And I'm soon to completa a reasanble iso-prolog implementation!

<stis>for guile!

<stis>And you?

<wingo>excellent!

<wingo>good, saw all the families back in the us

<wingo>back in geneva since tuesday

<wingo>slowly paging all the things back in :)

<stis>Warmer? or did you bring the cold weather with you ;.-)

<wingo>it's been warmer here than in the us -- there it was slushy and cold :)

<stis>paging in... I find it a very good exercise to do something else and the come back, you tend to improve the quality that way

<stis>e.g. seeing things that was on the black spot when one is 'in' the area

<wingo>landed a dead code elimination pass, and some generic beta/eta reductions

<wingo>effects analysis too, and some reachability things (what expressions are reachable from cont N that are not also reachable from cont M, etc)

<wingo>also $ktrunc (which is now $kreceive) appears everywhere that guile truncates multiple values, though it can be optimized out

<wingo>should help avoiding all kinds of values-related bugs that we have currently

<stis>cool!

<wingo>in the pre-cps side of things

<wingo>currently trying to finish the docs on cps

<stis>Cool, I would really like to enjoy some doc's and thoughts about 2.2

<wingo>there are some things in git already

<civodul>Hello Guilers!

<civodul>our custom-binary-input-port implementation cannot be used to implement a port wrapper that computes a hash over what's actually been read

<civodul>because it has built-in buffering

<civodul>that sucks

<civodul>soft ports don't have this problem, but they have other problems

<civodul>any suggestions?

<civodul>hey, howdy wingo!

<wingo>heya civodul :-)

***Guest3074 is now known as `micro

<jemarch>hi

<civodul>hello jemarch

<dje42>Setting aside various kinds of critical sections, is it bad to longjmp out of scm_with_guile? E.g. suppose a user-provided scheme function does the longjmp.

<dje42>[I don't know if this can happen for me today, and I'm going to avoid it where I can. At this point I'm just wondering.]

***add^_` is now known as add^_

<wingo>updated the "master" docs on the site

<wingo> http://www.gnu.org/software/guile/docs/master/guile.html/CPS-in-Guile.html#CPS-in-Guile

<stis>wingo yey!

<dje42>civodul: saw your message re: gdb 7.8

<civodul>ah yes

<dje42>Way back when I didn't expect 7.7 to be as delayed as it was.

<dje42>But no matter, 7.8 needn't be similarly slipped.

<dje42>It was problematic for various reasons to get guile into 7.7 so I never pushed for it.

<civodul>ok

<dje42>Now that 7.7 has branched, I'm hoping to get the patches in soon.

<civodul>cool!

<dje42>Last real TODO is SIGINT handling.

<civodul>right, and i didn't really have any solution :-/

<civodul>BTW, i noticed tromey will talk about gdb at FOSDEM

<dje42>Ah.

<tromey>civodul: if you're going to be at FOSDEM I'd love to meet up. I was hoping Dodji would give me an introduction to you & wingo but it turns out he isn't going to make it :(

<wingo>bummer about dodji! but I will try to go; depends on some things at home

<civodul>tromey: yes, that'd be great

<civodul>hopefully there'll be a GNU dinner on Friday

<tromey>I'll have to check my schedule :)

<tromey>but I'm sure we can find some time

<civodul>heh

<civodul>most likely i'll attend your talk, so at least we could meet there :-)

<tromey>was the dinner mentioned on gnu-prog-discuss?

<civodul>mostly ghm-discuss

<tromey>ah

<tromey>not sure I"m on that

<civodul>that's the helper list to meet gnus in real life ;-)