IRC channel logs

2014-11-14.log

back to list of logs

***fangism is now known as fangism-hongry
<nalaginrut>morning guilers~
*nalaginrut just received T-shirt from Lisp in summer project http://instagram.com/p/vXTzHHqYCQ/
<davexunit>neat :)
<nalaginrut>;-)
<nalaginrut>wow, Turing Award will be million dollars
<civodul>Hello Guilers!
<ArneBab>moin
<ArneBab>did someone already port GNU greg to guile 2.0? (couldn’t find info on that)
<ArneBab>I found this: https://github.com/Hoops/greg/commits/master ← does that mean support would be there but wasn’t merged?
<civodul>no idea
<ArneBab>that’s one of the two blockers for guile 2.0.11 in Gentoo (the other is texmacs)
<civodul>ok
<civodul>i think greg is (or used to be) part of GNUstep, no?
<ArneBab>from the info-docs that was the intention. It uses Guile to drive test-scripts.
***adhoc_drupalcamp is now known as adhoc
<agumonkey_>hey there
<agumonkey_>I wanted to try 2.2 branch
<agumonkey_>but there's no branch yet, will master do ?
<taylanub>agumonkey_: master is the upcoming 2.2, yes. the version number will say 2.1, which is the development version of 2.2.
<agumonkey_>oh ok
<agumonkey_>thanks
<dsmith-work>Happy Friday, Guilers!!
<taylanub>anyone looked into Snow? http://trac.sacrideo.us/wg/wiki/Snow https://github.com/sethalves/snow2-client I might start working on a Guile client.
<civodul>this is Marc Feeley's package manager, right?
<civodul>i looked at it around 2007, IIRC :-)
<civodul>around the same time, Serrano came up with a competitor
<civodul>and none of theme caught up
<civodul>*them
<taylanub>hm, this might be a reboot of it. this one is for R7RS. maybe that's why Seth Alves is calling it Snow2.
<civodul>yeah, probably
<mark_weaver>taylanub: Alex Shinn was working on a R7RS package manager as well. He emailed some selected Scheme implementors who were close to R7RS support about it.
<taylanub>Chibi has a Snow client too apparently, yes.
<taylanub>mark_weaver: by the way what's the status of the R7RS branch? if there's something you think I could do, please tell :)
<mark_weaver>taylanub: the main blocker at this point is that my implementation of cyclic-data handling in 'write' is too slow.
<mark_weaver>part of the problem is not being able to change the print state data structure, because it's public and thus part of the ABI.
<mark_weaver>so the current branch has a rather convoluted attempt to get the job done despite that.
<mark_weaver>also, I'm rather discouraged with the R7RS-large effort, which is not exactly helping to motivate me.
<mark_weaver>taylanub: I agree with wingo's comments in http://article.gmane.org/gmane.lisp.scheme.reports/4005 and http://article.gmane.org/gmane.lisp.scheme.reports/4019
<mark_weaver>anyway, despite all this, I guess we should still move ahead on R7RS, if for no other reason than to facilitate code sharing with the chicken community.
<mark_weaver>so, the other blocker is that I do not yet have sufficient confidence that the code has been adequately tested to merge with stable-2.0. maybe you could help with that.
<mark_weaver>I think I know how to solve the efficiency problem in 'write': by implementing a simple specialized eq?-hash-table that uses open addressing instead of chaining to handle collisions, with a constraint that only the most recently added element can be removed (stack-like).
<mark_weaver>that should allow me to avoid allocation (which is very expensive in bdwgc) during hash table operations.
<wingo> https://wingolog.org/archives/2014/11/14/on-yakshave-on-color-on-cosines-on-glitchen
<stis>evening guilers!
<dsmith-work>wingo: nope
<dsmith-work>wingo: not with https anyway
<dsmith-work>wingo: probably some daconian corporate filtering on my end.
<dsmith-work>dRaconian
<stis>thanks wingo: a really entertaining read!
<dsmith-work>Yes, very!
<ArneBab>sneek: later tell wingo: Huge kudos on seeing your Firefox Javascript work polished and merged! (I read a news entry about it - not yet finished with your blog entries, though the first half of the yakshave sounds great!; maybe this weekend)
<sneek>Okay.
<ArneBab>sneek: later tell wingo: the first half *was a great read*
<sneek>Okay.
<ArneBab>sneek: botnsack
<ArneBab>sneek: botsnack
<sneek>:)
<mark_weaver>wingo: congrats on your fast generator work getting merged! and thanks for guile-jpeg, and the informative and entertaining post about it! :)
<wingo>mark_weaver: heh, thanks ;)
<sneek>Welcome back wingo, you have 2 messages.
<sneek>wingo, ArneBab says: Huge kudos on seeing your Firefox Javascript work polished and merged! (I read a news entry about it - not yet finished with your blog entries, though the first half of the yakshave sounds great!; maybe this weekend)
<sneek>wingo, ArneBab says: the first half *was a great read*
<wingo>haha the first half of that article was written with much wine, as you might be able to tell ;)
*stis is sipping wine, maybe that's why it was so nice :-)
<wingo>hehe :)
<dsmith-work>Seems like I missed a few of the yaks in the middle...
*stis shouled go to conferences and talk about guile-log, just for fun!!, any suggestions?
<sethalves>snow2 / chibi-snow is a reboot on snow. Marc Feeley has blessed the use of the name
<sethalves>the repository spec is written by Alex Shinn
<sethalves>I'm not discourgaed by lack of progress in r7rs-large. I sounds like they were mainly planning on writing and blessing srfis
<sethalves>*It
<sethalves>I tried to figure out how to run the guile r7rs branch recently, but I failed
<taylanub>mark_weaver: I'll do my best to test the branch then. other than Guile's own test suite, is there any "strategic" way for this? maybe just running the test suites of as many libraries and programs as possible? and wildly using those which have no good test suite?
***fangism-hongry is now known as fangism