Index

Show enters and exits. Hide enters and exits.

04:41:10boyscoutUpdated MSpec source to 8537c769. - b79cd96 - Brian Ford
04:41:10boyscoutUpdated CI specs to RubySpec 4fd07cfb. - 5ab9bc8 - Brian Ford
04:41:10boyscoutEnable features for with_feature guard. - fcda3dd - Brian Ford
04:41:10boyscoutUpdated CI tags for merged rubyspecs. - d8f4b4c - Brian Ford
04:58:26boyscoutCI: Commit d8f4b4c failed. http://github.com/evanphx/rubinius/commit/d8f4b4ceb1832abad24beaa6b6eb306865d25693
04:59:17brixeninteresting
05:00:40brixentempted to say rbx is not compliant on that nonsense
05:05:24brixenotoh, we have some wacky code here...
05:46:18brixenruby -e 'p Math.atanh(1.0)' => ERANGE on debian/lenny
05:46:23brixenbut EDOM on ubuntu
05:46:32brixenfucking please already
07:59:33dbussinkbrixen: just a fyi, i've succesfully installed freebsd with virtualbox
07:59:49dbussinkdidn't hit any glitches to be honost
07:59:51dbussinkhonoust
08:13:43brixendbussink: what version of vbox and what version of freebsd?
08:13:59brixenhave to say, I'm seriously underwhelmed by vmware fusion so far
08:14:09brixenI'd rather use vbox
08:14:18dbussinkbrixen: virtualbox 3.1.4, freebsd 8.0 64 bit
08:14:30brixenit would not boot for me
08:14:33brixensame versions
08:14:37dbussinki think i installed it before the 3.1.4 release with 3.1.3
08:14:46dbussinkbut it still works now with 3.1.4
08:14:55brixen8.0-RELEASE-amd64-disc1.iso
08:15:01dbussinkbrixen: where does it hang?
08:15:17brixenle'me try it again
08:15:24dbussinki've used the net installer
08:15:29dbussinknot the full iso
08:16:06dbussinkbrixen: i've chosen the freebsd 64 bit os type
08:16:24brixenCPU doesn't support long mode
08:16:34brixenthen gives me the OK prompt instead of booting
08:16:47brixenI installed fine from this iso under vmware
08:17:01dbussinkthat's weird
08:17:05brixenyes it is
08:17:09brixenanyway...
08:17:17brixenso these #@$@#$ Math.atanh specs
08:17:23brixenthey make me so blue
08:17:33brixenI'm considering a new rubyspec guard
08:17:43dbussinkbrixen: btw, i think the reason that ubuntu and lenny are different is that ubuntu already switched to uclibc instead of glibc
08:18:12dbussinkand that uclibc behaves slightly different from glibc
08:18:54brixensomething like this http://gist.github.com/306427
08:19:20brixenif MRI refuses to make specify sane behavior, then rubyspec can essentially exclude MRI
08:19:34brixenand let implementations that will specify sane behavior do so
08:19:50brixendbussink: I don't really care
08:20:10brixendepending on the libc like this is absolutely lame
08:20:26dbussinkwell, mri has the habbit of just exposing behavior like this
08:20:34dbussinkand not go through any means to unify it
08:20:41dbussinkor only few means
08:20:45brixenjust because they can't find their ass with 2 hands and a flashlight when it's something so clear as a math function doesn't mean all ruby users should suffer
08:21:14brixenlibc behavior in this case is inexcusable
08:21:21brixenbut MRI should not just pass the buck
08:21:27brixenthat is doubly inexcusable IMO
08:22:48dbussinkyou don't have to convince me about this :)
08:23:07brixenI'm just pissed at wasting so much time on this
08:23:15dbussinkit's going to be fun when porting rbx to windows
08:23:21brixenand actually having to argue with ruby-core that we should have sane behavior in MRI
08:23:24brixenugh!
08:23:42brixendbussink: Factor does a very good job being platform independent
08:23:48brixenit's entirely possible
08:23:56brixenyou just have to value sanity
08:24:03dbussinkthe problem is not whether it's possible, it's what the behavior will be
08:24:16dbussinki know the jruby guys have chosen to mimic mri windows behavior on windows
08:24:27dbussinkwhich creates all kinds of exceptions just for windows
08:24:35brixenthat's not being platform independent
08:24:37dbussinkwhile behaving consistently would have been for free
08:24:42dbussinkacross platforms that is
08:25:01brixenwell, they should push back more
08:25:08brixenthose are MRI bugs IMO
08:25:39dbussinkthe number of differences should be as minimal as possible and if they can be fixed they should
08:25:46dbussinkand behave the same everywhere
08:28:58dbussinkbrixen: but from the discussion in rubyspec i got the idea that there is this wall of misunderstanding
08:29:47brixenI don't get what is so hard to understand
08:33:36dbussinkbrixen: well, i see an ivory tower thing, "we don't want to specify this"
08:33:48dbussinkthey don't seem to grasp that what people do with code is also the spec
08:33:56dbussinksince that's code that works and people depend on
08:36:56brixenyes, exactly
08:37:42slavahi brixen
08:38:56brixenhi slava
09:37:13boyscoutUpdated CI specs to RubySpec 48fe98c6. - 498fb88 - Brian Ford
09:37:34brixenwhat a freaking mess Math is
09:37:39brixenand Complex
09:37:44brixenle sigh
09:37:47brixeng'nite
09:42:31boyscoutCI: rubinius: 498fb88 successful: 3037 files, 11905 examples, 36184 expectations, 0 failures, 0 errors
09:44:16dbussinkbrixen: nite!
16:27:38ruedbussink: Definitely true there is a perception some things can be unspecified where they clearly cannot, like Hash ordering
16:27:55rueBut some things legitimately can be
16:54:37evanmorning.
16:58:37kronos_vanoevening.
17:03:09brixenmorning
17:12:29kronos_vanoIs it ok, that running code with -Xint option faster than without it?
17:12:37evanno
17:12:43evanit's not.
17:12:50evanwhat code?
17:13:03kronos_vano1 sec
17:15:01kronos_vanoevan, http://gist.github.com/306819
17:15:36evanweird.
17:16:04kronos_vanoruby18, ruby19 and jruby 1.4 is 2x faster on this code.
17:18:00evanhm, i get about the same with and without -Xint
17:18:01evanbut still
17:18:40evanI should note how much faster s << is that += here
17:18:43evanabout 100x faster.
17:19:26kronos_vanoIt is benchmark from ruby v8
17:20:50evanruby v8?
17:24:04kronos_vanohttp://github.com/macournoyer/rbv8
17:26:16evankronos_vano: i'll bet this benchmark spends 90% of the time copying data
17:26:28evanthats why -Xint seems slightly faster
17:27:26kronos_vanoProfiling point to GC.collect_mature and String#+
17:28:30evanthe fact that I can run with -P and it doesn't seem to effect the time at all
17:28:56evanpoints to the idea that it's stuck elsewhere
17:31:31kronos_vanobbiab
17:59:27brixenf@#$@#!! http://redmine.ruby-lang.org/issues/show/1708
17:59:33brixenyou have got to be kidding me
17:59:58brixenMath/Complex is not just "unspecified" in 1.9
18:00:02brixenit's undefinable
18:00:18brixensomeone has really jumped the shark
18:02:46evanbrixen: don't bother with 1.9 there i think for now
18:02:52evanwe need to push back hard to get them to fix that.
18:04:03slavahi evan
18:04:07evanhi!
18:12:25luislavenabrixen: ping?
18:12:29luislavenasorry, hello guys :)
18:13:24dbussinkluislavena: howdy luis :)
18:13:29dbussinkluislavena: got rbx to compile on windows? ;)
18:13:49brixenluislavena: sup?
18:13:58luislavenadbussink: what are you talking about? I have plenty of windows at home, but they don't compile, they just make light get in.
18:14:16luislavenadbussink: ;-)
18:14:33dbussinkluislavena: ah, light? what is that? i've seen this yellow shiny thing yesterday, but i can't remember what it was
18:14:39luislavenabrixen: do you have a place for me to drop you 450MB of 7-zip compressed gold?
18:14:41dbussinkit disappeared again today though
18:15:02dbussinkluislavena: evan has a machine that might have that
18:15:43evan7lbs of gold!
18:15:45evanrad!
18:15:47evani'll take it.
18:16:12evanluislavena: use drop.io
18:16:17brixenluislavena: can I just download it from you, or do you need to drop it?
18:16:36dbussinkevan: what's an lbs? ;)
18:16:44evan:)
18:17:23luislavenabrixen: I'm on a sucky french internet connection (not associated with the movie)
18:17:29brixendbussink: it's a kg for imperialists ;)
18:17:37brixenluislavena: heh
18:17:39dbussinki might bring stroopwafels this year to railsconf if people are interested ;)
18:17:59brixendbussink: woot!
18:18:17dbussinkdepending on whether my talk is accepted or not mostly
18:18:48brixenluislavena: well, dunno then, will drop.io work as evan suggested?
18:18:58evanshould
18:19:07evandrop.io is just a play to upload and download files.
18:19:10evanplain and simple
18:19:13luislavenabrixen: is my private installation of open source software.
18:19:13evanused it a number of times
18:19:21luislavenabrixen: so guess no big deal then ;-)
18:19:55brixenluislavena: I won't share it with anyone :P
18:20:23luislavenabrixen: lol, is outlined here: http://github.com/luislavena/binfiles
18:21:11dbussinkluislavena: ah, good be used to work on windows compilation?
18:21:12brixenluislavena: what? omg, you're sharing codez!
18:21:15dbussinkcould be used
18:22:20luislavenabrixen: ;-)
18:22:43dbussinknow we can all be as cool as luis and do his magic too :)
18:24:28luislavenadbussink: you can know the tricks, but you will never present them like I do ;-)
18:24:50luislavenabrixen: will send you some drop.io then tomorrow your time.
18:25:07luislavenabrixen: playing catchup with OSS stuff and pinned this conversation last time :(
18:25:39brixenluislavena: no worries, I appreciate it
18:25:45brixenluislavena: enjoy France! :)
18:27:20dbussinkluislavena: in paris again?
18:27:39luislavenadbussink: yes, arrived Feb 2, was in manchester the week before.
18:27:48luislavenadbussink: will be here until April 4 I think.
18:28:57jvoorhishello
18:29:29dbussinkluislavena: ah, cool :)
18:29:53luislavenadbussink: coming to visit? ;-)
18:30:19dbussinkluislavena: hehe, well, at least the trains are going again :P
18:30:50jvoorhisbrixen: that Complex issue is kind of interesting – i really don't like the situation :\
18:31:01jvoorhisbrixen: this looked interesting too – http://redmine.ruby-lang.org/issues/show/2189
18:34:45brixenjvoorhis: yes, 2189 is related
18:34:50brixenand unresolved
18:34:53brixenand ridiculous
18:35:38jvoorhisyeah, it's probably a good time to speak up
18:36:08brixenwell, and raise what point that was not ignored in http://redmine.ruby-lang.org/issues/show/1708 ?
18:36:20brixenI'm filing tickets now, but seriously
18:36:22brixenbullshit
18:36:37jvoorhisjust re-raise the issues :) be precise and consistent
18:36:42jvoorhisget support
18:36:46jvoorhisand patches
18:37:05evanjvoorhis: are you offering to help? :)
18:37:15jvoorhisevan: i'm beginning to feel inclined to
18:37:20evanyay!
18:38:05jvoorhisi really think Math is broken :\
18:39:06jvoorhisi'd actually prefer if we had FMath and CMath but no Math (but i won't hold my breath)
18:45:10jvoorhisand even though "Ruby is magic", it would be nice to have semantics for some subset of it :)
23:49:59evanof String#split
23:50:04evans/of/oh/
23:50:06evanyou're so stupid.
23:50:15evan"hi!".split("") => ["h", "i", "!"]
23:50:22evan"hi!".split("", -1) => ["h", "i", "!", ""]
23:50:25evan"hi!".split("", -2) => ["h", "i", "!", ""]
23:50:28evan"hi!".split("", -3) => ["h", "i", "!", ""]
23:50:30evanle sigh.
23:50:35evanwhat a stupid ass edge case.
23:54:05brixenugh
23:54:59evani don't see why you'd do that.
23:55:51brixenme neither
23:56:21brixenif there is an implicit match between ! and $ then why is there not between ^ and h
23:56:38evan*shrug*
23:56:44brixenyeah
23:57:06evanI should note that the split code is a mindfield of insanity
23:57:09evanit's so hard to work with.
23:57:18brixen:(
23:57:29evani'm resorting to adding special cases to try and get this one new spec to pass
23:57:40brixenmindfield of insanity conjures many interesting images
23:59:09brixengiven this example: "1,2,,3,4,,".split(',', -4) # => ["1", "2", "", "3", "4", "", ""]
23:59:15brixenit kinda make sense
23:59:22brixenbut not for ""
23:59:24evanwhy -4?
23:59:29evandoes -1 get you the samet hing?
23:59:31brixenyeah
23:59:34brixennot the -4
23:59:44brixenbut the final two "" in the array
23:59:46evanoh, the extra ""
23:59:52evansure, because there is nothing between the commas
23:59:53evansure
23:59:57brixenie, the trailing , is really an empty field