Da
Show enters and exits. Hide enters and exits.
| 00:00:17 | brixen | heh |
| 00:00:31 | brixen | I'll try to fix that before next week |
| 00:04:35 | context | Unable to find FFI 'ffi_environ' in: this process |
| 00:04:48 | context | and ffi_errno |
| 00:05:04 | evan | huh. |
| 00:05:05 | evan | thats wierd. |
| 00:05:10 | context | hold on |
| 00:05:37 | context | hah! |
| 00:05:40 | context | its cause i strip'ed it |
| 00:05:44 | context | dont strip vm/vm! |
| 00:05:48 | evan | and so the "my half ass ruby implementation is a thousand times faster than anyone else" game continues http://smalltalk.felk.cvut.cz/projects/smallruby |
| 00:05:51 | evan | context: um yeah |
| 00:06:03 | evan | you're really trying to push all of it's buttons aren't you? |
| 00:06:14 | context | evan: yes! its what im good at :p |
| 00:06:20 | evan | open tickets on these things |
| 00:06:26 | evan | don't expect us to remember what goes on in IRC. |
| 00:06:26 | context | kk |
| 00:07:55 | brixen | haha selected Ruby benchmarks |
| 00:07:59 | brixen | mmkay |
| 00:08:09 | brixen | no selected rubyspec results though |
| 00:08:20 | evan | oh |
| 00:08:22 | evan | puts that on the list |
| 00:08:24 | brixen | will have to add mspec --score before RC |
| 00:08:25 | context | using github's issue tracker? |
| 00:08:28 | evan | we talked about a rubyspec score |
| 00:08:30 | brixen | context: yeah |
| 00:08:31 | evan | we should sort that out this week |
| 00:08:34 | evan | er. next week. |
| 00:08:38 | evan | hah |
| 00:08:38 | brixen | evan: yah |
| 00:08:42 | evan | :) |
| 00:09:04 | brixen | evan: more important than that is you tiered benchmarks |
| 00:09:14 | brixen | you should make that a github project actually |
| 00:09:22 | brixen | RealRuby |
| 00:09:26 | evan | yeah |
| 00:09:34 | evan | i gotta work on them this weekend and next week |
| 00:12:44 | rue | evan: That is not right, calling it a half-assed implementation... |
| 00:13:00 | rue | It is a quarter or so |
| 00:13:10 | evan | yes |
| 00:13:14 | evan | i was being generous |
| 00:13:56 | evan | ok, 10.6.2 upgrade. |
| 00:13:58 | evan | brb. |
| 00:15:06 | evan | or not |
| 00:15:10 | evan | gotta download it I guess. |
| 00:15:15 | context | ;) |
| 00:18:54 | context | and Safari 4.0.4 |
| 00:18:56 | rue | OS X gives you a 'you must restart' before you have downloaded the patch, Windows restarts before giving you a 'downloading a patch' |
| 00:46:43 | evan | and i'm back. |
| 00:47:33 | imajes | evan: i've missed you. :/ |
| 00:47:57 | evan | aah, thanks! |
| 00:48:21 | imajes | also, how come --prefix is broken |
| 00:48:29 | imajes | that's like, hard to break |
| 00:48:37 | evan | actually it's not. |
| 00:48:41 | evan | because --prefix implies install |
| 00:48:44 | evan | and it's install thats busted. |
| 00:48:48 | imajes | ah |
| 00:49:01 | imajes | that makes more sense. |
| 00:49:19 | imajes | --prefix implying install is abit strange tho |
| 00:49:29 | evan | why? |
| 00:49:36 | evan | how do you get something to that prefix unless you install it? |
| 00:50:06 | imajes | that's install. --prefix just sets the path you want stuff. and so it sets where build should set it's relative deps |
| 00:50:21 | rue | It should not attempt to install by itself... |
| 00:50:28 | evan | nor does it. |
| 00:50:32 | imajes | i'm just dealing in semantics tho |
| 00:50:39 | evan | yeah |
| 00:50:41 | evan | your point is valid |
| 00:50:46 | evan | but doesn't matter. |
| 00:50:49 | imajes | right. :D |
| 00:50:57 | rue | Theoretically, right, configure and build should not touch install |
| 00:51:16 | evan | nor do they. |
| 00:51:51 | imajes | this probably explains why it's been crufty on rvm lately |
| 00:52:16 | imajes | wayneeseguin: you should probably disable 0.13 from rvm install i guess if it can't actually, you know, install |
| 00:52:55 | evan | 0.12 couldn't either |
| 00:52:58 | evan | so it's nothing new. |
| 00:53:06 | evan | we limped it along and then broke it again |
| 00:53:09 | wayneeseguin | imajes: rvm currently only supports rbx head |
| 00:53:11 | imajes | i think it might have been rbx head |
| 00:53:14 | imajes | right |
| 00:53:24 | wayneeseguin | untill installer works |
| 00:53:44 | imajes | evan: is it some core issue? or something else? |
| 00:53:50 | imajes | (That prevents install from working) |
| 00:54:42 | evan | just haven't gotten it to the stage we want it |
| 00:54:52 | evan | there are issues related to where to find the runtime to sort out |
| 00:54:56 | evan | we've been busy with other things |
| 00:54:59 | imajes | k |
| 00:55:23 | evan | feel free to stop complaining and help fix it though! |
| 00:55:32 | wayneeseguin | lol |
| 00:55:44 | imajes | evan: hah! i'm not complaining but i will try and help fix.. easier than trying to write the engine :D |
| 00:56:01 | evan | talk to brixen he's got the scheme worked out. |
| 00:56:07 | evan | he's actually probably working on it right now. |
| 00:56:10 | imajes | ah, cool |
| 00:56:11 | imajes | will do |
| 00:57:49 | wayneeseguin | brixen: how can we help? |
| 01:00:00 | brixen | let me finish it |
| 01:01:18 | wayneeseguin | ::grin:: |
| 01:02:42 | imajes | heh |
| 01:03:37 | brixen | give people the option of 0 (zero) additional benefit and more complication |
| 01:03:44 | brixen | and they will choose it every time |
| 01:04:05 | evan | heh |
| 01:05:09 | brixen | instead, we make is simpler by not having to install |
| 01:05:16 | brixen | and you can't keep people from it |
| 01:05:21 | brixen | people are amazing |
| 01:07:35 | brixen | I should get xkcd to write this into a comic :) |
| 01:07:40 | brixen | maybe they already have |
| 01:15:36 | imajes | brixen: haha, |
| 01:15:39 | imajes | just don't have a make install |
| 01:15:45 | imajes | and don't tell people it's not broken |
| 01:15:49 | imajes | problem solved :D |
| 01:16:31 | evan | as I recall, jruby doesn't have an install. |
| 01:17:02 | rue | Just put it there and with CLASSPATH and whatnot |
| 01:17:18 | rue | Suppose the binaries have to go somewhere...dun remember now |
| 01:17:24 | rue | Last time I installed, used rvm |
| 01:17:40 | rue | I blame rvm for destroying my JRuby-install skills |
| 01:18:57 | brixen | imajes: good point. telling people "don't do X" is about the surest way to get them to try it |
| 01:19:22 | imajes | yes. |
| 01:19:25 | brixen | even god knew that! muhaha |
| 01:19:35 | imajes | .. and yet... |
| 01:20:46 | wayneeseguin | rue: wha? |
| 01:20:54 | imajes | but, brixen: seriously, if you want help doing install stuff, i'm happy to contribute. |
| 01:22:08 | wayneeseguin | brixen: does rbx care what dir its' in? Can't we just copy it out to another dir for "install" ? |
| 01:23:09 | evan | wayneeseguin: no. |
| 01:23:14 | evan | we went over this last time. |
| 01:23:31 | evan | by default, it encodes it's build dir into the binary |
| 01:23:32 | wayneeseguin | sorry, I forget things quickly ;) |
| 01:23:39 | wayneeseguin | ahh ok |
| 01:23:39 | evan | so if ou move it, it doesn't know that. |
| 01:23:49 | evan | otherwise, it doesn't know where to find lib, runtime, etc. |
| 01:24:02 | wayneeseguin | makes sense, thats why I made build dir == src dir for rvm |
| 01:24:17 | wayneeseguin | Sorry for being dense :) |
| 01:24:23 | evan | no prob. |
| 01:25:54 | brixen | wayneeseguin: just ln the vm/vm wherever you want it |
| 01:26:15 | brixen | if you really need to have rbx in a particular dir right now, clone it there |
| 01:26:30 | wayneeseguin | understood thank you |
| 01:28:26 | imajes | hey, evan: that makes the --prefix thing make more sense. :D |
| 01:29:13 | evan | thinks he and brixen are getting tired of people assuming everything is easy. |
| 01:29:14 | evan | :/ |
| 01:29:40 | Zoxc | But everything is easy ;D |
| 01:29:40 | imajes | evan:writing an install block for a makefile is easy. knowing how it inter-relates to rbx is hard |
| 01:30:02 | brixen | imajes: bingo |
| 01:30:24 | evan | so, copying files, easy! |
| 01:30:30 | evan | good to know. |
| 01:30:30 | evan | thanks. |
| 01:30:41 | evan | :) |
| 01:30:42 | imajes | evan: dude, no need for the snark |
| 01:30:43 | brixen | heh |
| 01:30:54 | evan | sorry sorry. |
| 01:30:59 | imajes | trying to help is hard when you don't know shit |
| 01:31:03 | evan | i should go chill out before I head to abby's work party. |
| 01:31:14 | brixen | chill out everyone |
| 01:31:15 | rue | Some additional documentation would help the knowledge level and understanding the problem space |
| 01:31:19 | imajes | and you guys have been day-in-day-out for the past billion commits |
| 01:31:20 | imajes | and |
| 01:31:20 | imajes | btw |
| 01:31:32 | imajes | evan, brixen: you guys are super awesome so ou have lots of stuff in your brain :D |
| 01:31:34 | rue | On the other hand, adding documentation takes knowledge and time |
| 01:31:55 | rue | So it is one of those tricky equations |
| 01:32:00 | imajes | aye, it's hard |
| 01:32:00 | evan | alright. |
| 01:32:07 | evan | hey everyone! |
| 01:32:10 | evan | you're all dismissed |
| 01:32:11 | evan | it's friday |
| 01:32:13 | evan | go enjoy yourself! |
| 01:32:20 | brixen | yay! |
| 01:32:20 | rue | Does not compute. |
| 01:32:22 | imajes | have a good party, evan :) |
| 01:32:32 | imajes | say hi to abby! |
| 01:32:36 | evan | will do |
| 01:32:43 | evan | later guys |
| 01:32:44 | wayneeseguin | yes! |
| 01:32:50 | wayneeseguin | w00t! |
| 06:14:13 | Defiler | Hrm.. how do I push to the gh-pages branch? |
| 06:15:08 | Defiler | git push origin gh-pages says: "src refspec gh-pages does not match any." |
| 06:18:33 | Defiler | or this when I start over, wtf? https://gist.github.com/412221be3a7e6038a9a3 |
| 06:29:42 | rue | That is just from the nonexplicit remote, I think |
| 06:31:11 | Defiler | How is that not an explicit remote? |
| 06:31:19 | Defiler | Oh, I see what you mean |
| 06:31:35 | Defiler | but it's a branch that I created with 'git checkout -b pages origin/gh-pages' |
| 06:31:43 | Defiler | which sounds sane to me |
| 06:31:55 | rue | Ah |
| 06:32:01 | rue | Name mismatch |
| 06:32:19 | rue | So you need to explicitly designate that pages is to update gh-pages |
| 06:32:21 | Defiler | it has the correct name in .git/config though, and I get that error when I specify it |
| 06:32:58 | Defiler | I'll rename it and see what happens |
| 06:34:02 | boyscout | Fix some typos and stuff in the release notes - 2bbf28b - Wilson Bilkovich (gh-pages) |
| 06:34:02 | Defiler | aah wtf |
| 06:34:32 | Defiler | that fixes it? why can I push to remote branches easily with alternate names? |
| 06:34:35 | Defiler | and not that? crazy. |
| 06:35:59 | rue | You can, but you need to be explicit |
| 06:36:42 | Defiler | git push origin gh-pages is what I tried first |
| 06:36:47 | Defiler | is there a more explicit syntax? |
| 06:37:15 | rue | In the correct branch |
| 06:37:29 | rue | $ git push origin HEAD:gh-pages I think |
| 06:37:34 | Defiler | oh wow |
| 06:37:53 | rue | The refspec gets a bit...wrong at points |
| 06:38:34 | Defiler | oh, ok.. source ref and dest ref |
| 06:38:54 | Defiler | OK, name local branches after remote ones ha ha |
| 06:39:21 | Defiler | OK, you can also apparently do git push origin pages:gh-pages |
| 06:39:36 | rue | Ah, of course |
| 06:40:03 | Defiler | I thought both args to push were parts of the remote repo spec |
| 06:40:10 | rue | Yeh, the origin is just the repository |
| 06:40:16 | Defiler | but it's actually one remote and one src/dest pair. wacky times |
| 06:40:17 | rue | The second+ are the refspec |
| 06:52:22 | Defiler | man, this Charles Bailey has some git chops http://stackoverflow.com/questions/1270514/undoing-a-git-push |
| 06:52:38 | Defiler | git update-ref hardcore |
| 07:10:13 | Defiler | brianmario: Did I hear that you know how to build x86_64 LLVM on Snow Leopard? I've tried several different approaches, and it is tricky |
| 07:10:57 | Defiler | I tried archflags/cxxflags/cflags -arch x86_64 and EXTRA_OPTIONS=-m64 fun |
| 07:16:30 | brianmario | I actually just let the rbx build scripts do it all |
| 07:16:32 | brianmario | worked great |
| 07:16:53 | brianmario | are you trying to install it into your system separately from rbx? |
| 07:21:04 | Defiler | Yeah, I am |
| 07:21:08 | Defiler | so that I can build macruby |
| 08:49:41 | dkubb | I'm trying to install rubinius via rvm, and I got an error when installing: https://gist.github.com/48ec274983d806614a8a |
| 08:50:17 | dkubb | this is on a fresh 10.6.2 Snow Leopard installed today |
| 08:51:18 | dkubb | I installed libev 3.6 via macports, but it made no difference in the error |
| 09:55:28 | naeu | What's the best way to get Rake for Rubinius 0.13? |
| 09:55:38 | naeu | I feel I'm clearly doing something wrong here: http://gist.github.com/234464 |
| 09:59:37 | cyndis | rake should still have installed even though the rdoc failed |
| 09:59:57 | naeu | cyndis: where will it reside? |
| 10:00:11 | cyndis | you can run it using "rbx -S rake" |
| 10:00:33 | cyndis | or hmm |
| 10:00:43 | cyndis | at least it's in rubinius/bin/rake |
| 10:01:10 | cyndis | but yeah, -S seems to run the correct one too |
| 10:01:22 | naeu | cyndis: rake isn't in my rubinius/bin dir |
| 10:01:48 | cyndis | oh.. i wonder how i got mine (clean clone from today, anyway) |
| 10:02:16 | naeu | and -S doesn't work as my gemdir isn't in my PATH, it's specified in gemhome in ~/.gemrc |
| 10:02:54 | naeu | hmmm, I downloaded the tgz of 0.13 today |
| 10:03:11 | naeu | then ran ./configure --enable-llvm |
| 10:03:14 | cyndis | does rubinius/gems/1.8/gems/rake-0.8.7/bin/rake exist? |
| 10:03:16 | naeu | and then rake |
| 10:04:08 | naeu | no, I don't have a gems dir in my rubinius dir |
| 10:06:07 | cyndis | i'll try compiling 0.13 from scratch |
| 10:06:31 | naeu | ooh ok |
| 10:07:09 | cyndis | i'm not really familiar with rbx or anything though, so probably i'm not going to be able to help |
| 10:07:54 | naeu | no, me neither :-) |
| 10:08:08 | naeu | i just wondered if I could get my rspec test suite running with rubinius |
| 10:08:24 | naeu | i keep trying whenever a new release is made |
| 10:08:30 | cyndis | :) |
| 10:08:36 | naeu | but there are always obstacles in my way :-) |
| 10:11:09 | dbussink | naeu: there are know issues with rdoc / ri generation |
| 10:11:26 | dbussink | naeu: so for now i'd advise installing gems with --no-rdoc and --no-ri |
| 10:11:34 | naeu | dbussink: I just want rake for the moment |
| 10:11:47 | cyndis | built 0.13, bin/rbx gem install rake, got bin/rake |
| 10:12:09 | dbussink | naeu: ./bin/rbx -S gem install rake rspec --no-rdoc --no-ri if you want rake and rspec |
| 10:12:26 | dbussink | naeu: in your project dir you can do something like ../rubinius/bin/rbx -S rake spec |
| 10:15:30 | naeu | dbussink: where does rubinius install its gems? |
| 10:15:43 | naeu | the installation of rake was reported as successful |
| 10:15:57 | dbussink | naeu: in rubinius/gems |
| 10:16:34 | dbussink | naeu: installing rubinius is broken atm, so you should just use everything relative to the clone / download directory |
| 10:16:58 | naeu | hmmm, I don't see a rubinius/gems dir |
| 10:17:18 | dbussink | naeu: rubinius here is the toplevel dir here |
| 10:17:24 | dbussink | i have rubinius cloned in a rubinius dir |
| 10:17:31 | dbussink | so there is a gems dir in that toplevel directory |
| 10:17:33 | naeu | yeah, sure |
| 10:17:51 | naeu | no gems dir in there |
| 10:18:09 | dbussink | strange, since i have installed some gems to test stuff and i see it there |
| 10:18:37 | naeu | nevermind, I can wait till the next release :-) |
| 10:19:10 | naeu | it's great to see development moving at a good pace though |
| 10:19:24 | dbussink | naeu: did you download the release or clone the repo? |
| 10:19:54 | dbussink | naeu: i've been running some specs too to expose bugs etc. |
| 10:20:01 | dbussink | it's really helpful in fixing bugs etc. |
| 10:20:56 | dbussink | but i'm heading out for some shopping, back later |
| 17:31:55 | wayneeseguin | evan / brixen around? I have a slight issue when trying to run latest git head rbx: http://pastie.org/private/z5hvrsfeifz9cm5ob4wgmw |
| 17:37:08 | evan | wayneeseguin: thats the issue I was talking about |
| 17:37:14 | evan | you must be moving the build dir |
| 17:37:31 | evan | how are you building it? |
| 17:41:39 | wayneeseguin | evan: ok so I'm checking out the repo itself to ~/.rvm/rbx-head |
| 17:41:42 | wayneeseguin | then cd to that dir |
| 17:41:51 | wayneeseguin | RBX_PREFIX="$rvm_ruby_home" ; export RBX_PREFIX |
| 17:42:19 | wayneeseguin | ./configure --prefix=$rvm_ruby_home |
| 17:42:42 | wayneeseguin | if you pass --jit it adds: --enable-llvm |
| 17:43:55 | wayneeseguin | then |
| 17:43:56 | wayneeseguin | rvm 1.8.7%rbx ; rake --trace |
| 17:44:11 | wayneeseguin | and that's pretty much it |
| 17:44:23 | wayneeseguin | the rvm 1.8.7%rbx switches to 1.8.7 with rake & ParseTree installed |
| 17:45:37 | wayneeseguin | evan: could it be that RBX_PREFIX ? |
| 17:46:46 | wayneeseguin | or configuring with the rvm_ruby_home prefix? |
| 17:52:18 | Defiler | wayneeseguin: --prefix is kinda broken right this moment |
| 17:52:34 | Defiler | see gotchas at the top of: http://rubini.us/release-0.13.html |
| 17:52:55 | Defiler | brixen is working on fixing the prefix stuff though, to enable the three-part config where you put bin, lib, and include in different places |
| 17:53:18 | wayneeseguin | Defiler: sweet I'll remove that from rvm then and hopefully people can use rbx again with rvm |
| 17:53:32 | Defiler | Today, though, it might just make sense for rvm to copy a built version of rbx somewhere, rather than actually doing rake install |
| 17:55:09 | Defiler | Just trying rvm for the first time today... how do I specify a version (that I installed with rvm) to run with 'rvm ruby'? |
| 17:55:13 | Defiler | It's not clear from the rvm usage output |
| 17:55:40 | wayneeseguin | Defiler: You do 'rvm rbx' then 'ruby -v' should output rubinius |
| 17:56:22 | Defiler | OK. Is there a way to do it for just one command as well? |
| 17:56:36 | Defiler | or is that just what 'rvm rbx some_file.rb' would do |
| 17:56:55 | wayneeseguin | Defiler: I'm about to push a fix to rvm also |
| 17:57:05 | wayneeseguin | Defiler: exactly correct 'rvm rbx some_file.rb' will run some_file.rb against rbx |
| 17:57:12 | Defiler | Sweet |
| 17:58:54 | Defiler | Also, 'rvm system' is expecting /bin/ruby and /bin/gem to exist; is there a way to tell it that mine is in /usr/local/bin? |
| 18:00:08 | wayneeseguin | Defiler: if you do 'rvm update --head' you should be able to 'rvm install rbx' now and it should build. |
| 18:00:21 | wayneeseguin | Defiler: yeah have it in your path :) |
| 18:00:35 | Defiler | It is though |
| 18:00:36 | wayneeseguin | Defiler: 'rvm system' actually simply disables rvm |
| 18:00:44 | wayneeseguin | Defiler: ahh ok do this: |
| 18:00:51 | Defiler | Oh, I was trying to do 'rvm system some_file.rb' |
| 18:01:00 | wayneeseguin | a) make sure rvm is sourced last in your profiles (.bash_profile / .zshrc / .bashrc) |
| 18:01:02 | Defiler | When I was switched to something else, but wanted a quick compare against 1.9 |
| 18:01:04 | Defiler | It is |
| 18:01:05 | wayneeseguin | ahhh ok yeah that won't work |
| 18:01:37 | wayneeseguin | I'll add a PT story to look into maybe adding that as a feature |
| 18:01:48 | Defiler | Yeah, thinking about how that would work.. |
| 18:02:01 | wayneeseguin | Defiler: I was using 'rvm system' as the way to disable rvm |
| 18:02:02 | Defiler | Just having a config file option for it would be perfect for me |
| 18:02:25 | Defiler | but I can just make an alias for flip-flopping. Not a big deal |
| 18:03:47 | wayneeseguin | Ok I added a PT story to at least add a system option to the do actions |
| 18:04:15 | Defiler | Oh, can I just set my default to 'system', actually? |
| 18:05:49 | Defiler | also, I get a big flood of these after doing 'rvm update --head', followed by 'rvm ruby' which told me to reload, and then an 'rvm reload' |
| 18:05:59 | Defiler | .rvm/scripts/ruby-installer:274: parse error near `]]' |
| 18:06:52 | Defiler | hopefully not because I use zsh though |
| 18:07:54 | Defiler | Oh, the installer makes it source that script, rather than just running it.. so it loads a bash script under zsh I guess |
| 18:08:06 | wayneeseguin | oh haha Defiler one sec I'll fix that |
| 18:08:46 | Defiler | changed it to if [[ -x /Users/wilson/.rvm/scripts/rvm ]] ; then /Users/wilson/.rvm/scripts/rvm ; fi |
| 18:08:50 | Defiler | and it works nicely |
| 18:09:00 | wayneeseguin | actually Defiler... I don't see a syntax error ... |
| 18:09:23 | Defiler | maybe zsh doesn't do that [[ "string" ]] test without any -options? |
| 18:09:41 | wayneeseguin | hrm which line are you seeing that ? |
| 18:10:06 | Defiler | if [[ "$rvm_trace_flag" ]] ; then |
| 18:10:09 | Defiler | (line 274) |
| 18:10:41 | Defiler | this script just needs to run under bash always |
| 18:10:56 | Defiler | So we can't source it into .zshrc |
| 18:11:33 | wayneeseguin | ahhhhhh I see that now |
| 18:12:16 | wayneeseguin | ok Defiler just pushed a fix for that, thank you. |
| 18:12:34 | wayneeseguin | it was supposed to be " if [[ "$rvm_trace_flag" -eq 1 ]] ; then |
| 18:12:48 | Defiler | aha |
| 18:12:54 | Defiler | cool |
| 18:13:14 | wayneeseguin | looks like rvm is workign in zsh here |
| 18:14:42 | Defiler | Yeah, it must just work. |
| 18:15:32 | wayneeseguin | tries rbx now |
| 18:15:53 | Defiler | ouch. rvm reload just crashed zsh and brought up the problem report window. heh |
| 18:16:40 | wayneeseguin | o.O |
| 18:16:40 | wayneeseguin | wow |
| 18:17:14 | wayneeseguin | ok now rvm is only doing "./configure ; rake" and it's still crashing |
| 18:19:42 | wayneeseguin | trying again with --trace |
| 18:19:46 | wayneeseguin | seems to be running longer this time |
| 18:23:01 | wayneeseguin | http://pastie.org/private/zfvo1x9grafetsbcdroaq |
| 18:23:11 | wayneeseguin | evan: ^ stack trace from failed `rake` on clean rbx |
| 18:23:26 | wayneeseguin | err , fresh git clone |
| 18:24:59 | Defiler | huh. weird. |
| 19:04:15 | Defiler | Hrm. I'm trying to get 'rbx gem install json' to build the extension |
| 19:04:28 | Defiler | and gcc barfs on this line: if (st_lookup((assert(???? && "RHASH() is not supported")->tbl), tmp, 0)) { |
| 19:04:44 | Defiler | I have no idea what those four question marks are about |
| 19:06:22 | brianmario | if you need a working JSON lib in rbx, I fixed up yajl-ruby to be compatible |
| 19:06:44 | Defiler | Yeah, I'm already using it on rbx. Good stuff |
| 19:06:51 | brianmario | ah cool |
| 19:07:10 | brianmario | that being said, I do think we should try and get as many gems working on rbx as possible |
| 19:07:26 | brianmario | to help flesh out missing parts of the capi and other bugs |
| 19:07:46 | Defiler | Yeah. Typhoeus builds nicely |
| 19:08:02 | Defiler | but json complains about rb_hash_foreach and rb_obj_id |
| 19:09:01 | Defiler | Oh, right, I see.. json uses RHASH |
| 19:09:38 | brianmario | hm |
| 19:12:11 | Defiler | I'm not sure what the best thing to replace st_lookup(RHASH(blah)->tbl, sym, 0) with though for best compatibility |
| 19:12:45 | Defiler | I'm apparently not sure sentence grammars eithers. |
| 19:20:09 | Zoxc | should redo reset parameters? |
| 19:21:38 | Defiler | What do you mean by reset? |
| 19:21:53 | Defiler | like, restore to their previous state? |
| 19:22:28 | Zoxc | restore them to the original parameters? |
| 19:23:56 | Defiler | Well, if you have a redo in, say, an 'each' |
| 19:24:04 | Defiler | it will repeat it with the same block params |
| 19:24:19 | evan | Defiler: yeah, we don't implement RHASH |
| 19:24:22 | Defiler | but if the block param is some object that you destructively modified in the previous run at it, it stays changed |
| 19:24:26 | evan | Defiler: not sure if we will actually. |
| 19:24:39 | Defiler | I don't see how we could, right? It exposes the hash data structure |
| 19:24:47 | Zoxc | it doesn't seem to be that way on 1.9.1 :/ |
| 19:25:23 | evan | Defiler: yeah, we could fake it a bit |
| 19:25:33 | evan | Defiler: ie, if only the st_* functions are used on it |
| 19:25:47 | evan | we could implement special st_* functions that use rb_funcall to get things. |
| 19:25:56 | Defiler | Zoxc: What's your test code look like? I just wrote a little script and ran it under 1.9 and it works that way |
| 19:25:59 | evan | the issue there is conflicts with legic usages of st |
| 19:26:48 | evan | wayneeseguin: i have no clue how that can be |
| 19:26:53 | evan | wayneeseguin: i've never seen that |
| 19:27:02 | evan | and don't know why libev's configure would freak out |
| 19:27:05 | evan | you're going to have to investigate. |
| 19:27:17 | Zoxc | http://pastebin.com/m7165e23b |
| 19:27:20 | evan | i could try and repro it. |
| 19:29:13 | Defiler | Zoxc: http://gist.github.com/234710 |
| 19:29:31 | wayneeseguin | evan: The weird thing is that I am simply doing the three steps from the website on a clean repo ... so odd |
| 19:29:39 | evan | hm, |
| 19:29:44 | Defiler | Your test doesn't pass because fixnums are immediates |
| 19:29:44 | evan | wayneeseguin: i'll try installing via rvm |
| 19:29:47 | evan | one sec |
| 19:29:59 | Defiler | You're not modifying the same object, you're creating a fresh fixnum each time |
| 19:30:22 | Zoxc | And? |
| 19:30:38 | Defiler | what I mean is x += 1 doesn't persist after a redo |
| 19:30:48 | Defiler | whereas modifying some object's data will |
| 19:30:51 | wayneeseguin | evan: if you do 'rvm install rbx --trace' it will spit out *everything* and it logs the individual commands run to ~/.rvm/log/rbx-head/*.log |
| 19:30:53 | Zoxc | isn't redo block local? |
| 19:31:02 | Defiler | Yes |
| 19:31:13 | Zoxc | then pass should not be touched :/ |
| 19:31:13 | evan | wayneeseguin: k. |
| 19:31:20 | Defiler | but on the stack, a string is a pointer to somewhere else, and a fixnum is just right there encoded in the stack |
| 19:31:32 | evan | wayneeseguin: it builds it's own ruby 1.8.7? |
| 19:31:35 | evan | (this is my first time with rvm) |
| 19:31:43 | wayneeseguin | :D |
| 19:31:51 | Defiler | Aah, I see what you are saying Zoxc |
| 19:32:05 | wayneeseguin | evan: yes, tis the only way to guarantee 1.8 + parsetree. Arch Linux default is 1.9 for example. |
| 19:32:23 | evan | wayneeseguin: you don't need parsetree anymore btw. |
| 19:32:44 | evan | what if the existing ruby is 1.8.7? |
| 19:32:49 | evan | seems like you should be able to use that. |
| 19:33:07 | Defiler | Zoxc: The behavior of block locals in 1.9 is definitely different |
| 19:33:13 | wayneeseguin | evan: really? w00t. Ok so I can remove that line and just ensure 1.8.7 then |
| 19:33:38 | Zoxc | so redo doesn't restore the parameters in 1.9 then |
| 19:33:55 | evan | Zoxc: redo and retry in blocks is... crazy. |
| 19:34:00 | evan | we don't support retry in a block |
| 19:34:01 | evan | never will. |
| 19:34:12 | Zoxc | I was talking about redo now ;D |
| 19:34:24 | evan | wayneeseguin: rvm install rbx --trace worked fine for me |
| 19:34:28 | brixen | seriously wishes folks would chill on the rbx install for one freakin day |
| 19:34:37 | brixen | evan: here's a legit issue you could look at http://gist.github.com/234715 |
| 19:34:56 | brixen | evan: that's running a request through sinatra |
| 19:34:58 | evan | brixen: mmmm |
| 19:34:59 | evan | k |
| 19:35:11 | evan | people are crazy. |
| 19:35:17 | evan | you define a method with define_method |
| 19:35:23 | evan | then detach it with method() |
| 19:35:27 | evan | to call it. |
| 19:36:08 | wayneeseguin | evan: interesting, I must have some environmental thing lurking |
| 19:36:14 | evan | wayneeseguin: you gotta fix the error log. |
| 19:36:16 | wayneeseguin | evan: thanks |
| 19:36:20 | evan | it's put the ENTIRE trace output as one line. |
| 19:36:23 | wayneeseguin | evan: what's up with it? |
| 19:36:24 | wayneeseguin | ahh |
| 19:36:26 | wayneeseguin | ok |
| 19:36:37 | Defiler | Huh, that is absolutely not how I thought redo worked |
| 19:36:52 | evan | wayneeseguin: i can't use rvm currently |
| 19:36:55 | evan | it's still using --prefix |
| 19:36:59 | evan | which doesn't work. |
| 19:37:00 | wayneeseguin | evan: rvm update --head |
| 19:37:04 | wayneeseguin | that has been removed |
| 19:37:28 | Defiler | Does redo do something different/special for blocks vs. procs? |
| 19:39:22 | Defiler | OK, no. I see how it works |
| 19:41:38 | Defiler | Zoxc: http://gist.github.com/234719 |
| 19:42:14 | Defiler | That code won't run under 1.9, but it shows how it behaves; it starts again at the top of the block, but it doesn't restore any names or states |
| 19:42:38 | Defiler | so if x = 3 has happened, the label 'x' isn't restored next time; you just get the new x |
| 19:42:54 | Defiler | I still have no idea when I would want redo |
| 19:43:26 | evan | wayneeseguin: how can I see the output? |
| 19:43:31 | evan | rake.error.log doesn't work |
| 19:43:41 | evan | the output doesn't make any sense |
| 19:44:08 | evan | I wanna see the actual output of running 'rake' in the rbx build dir |
| 19:44:20 | wayneeseguin | evan: well you can always do the command manually: |
| 19:44:34 | wayneeseguin | cd ~/.rvm/rbx-head ; rvm 1.8.7 ; rake --trace |
| 19:44:56 | wayneeseguin | That is exactly what rvm is doing |
| 19:45:10 | wayneeseguin | (latest version anyway) |
| 19:45:30 | evan | m. |
| 19:45:44 | evan | you're setting up some wierd ass flags |
| 19:45:46 | evan | or something |
| 19:45:51 | evan | because libev has |
| 19:45:56 | evan | -Wl,4 |
| 19:46:04 | wayneeseguin | aha! I know what might cause that |
| 19:46:07 | evan | in the cli to gcc |
| 19:46:44 | evan | wayneeseguin: why are you setting LDFLAGS? |
| 19:46:56 | evan | and CFLAGS |
| 19:46:59 | evan | and ARCHFLAGS |
| 19:48:08 | wayneeseguin | evan: OSX has been pure evil to get ruby to run on, that was the only way I was able to consistently get all MRI rubies to compile for everyone :/ |
| 19:48:13 | wayneeseguin | evan: I'm removing that call for rbx |
| 19:48:25 | Defiler | Yeah, I set pretty much all of those |
| 19:48:39 | Defiler | alias native_configure='./configure CFLAGS="-arch x86_64" CXXFLAGS="-arch x86_64" LDFLAGS="-arch x86_64" OBJCFLAGS="-arch x86_64" --build=x86_64-apple-darwin10.2.0' |
| 19:48:48 | brixen | wayneeseguin: you should be using the flags we use to compile rbx |
| 19:50:09 | wayneeseguin | brixen: Awesome, what are they? |
| 19:50:24 | wayneeseguin | brixen: trying to figure out proper OSX compilation for all users has caused me nothing but pain |
| 19:50:43 | brixen | cd rbx_root; rake -q clean; rake -t build |
| 19:50:46 | wayneeseguin | The dual architectures are pure evil |
| 19:50:49 | brixen | you can see the flags in the commands |
| 19:51:20 | Defiler | I don't understand why 32 bit is the default architecture on MacOS |
| 19:51:34 | Defiler | there should at least be a single gcc config you could set on your system that would make it work that way |
| 19:51:52 | evan | wayneeseguin: so |
| 19:52:04 | Zoxc | didn't they move to llvm already? |
| 19:52:09 | evan | wayneeseguin: you must be trying to pass in "-compatibility_version 4" |
| 19:52:28 | evan | but thats getting turned into -Wl,-compatibility_version -Wl,4 |
| 19:52:31 | evan | which is wrong. |
| 19:52:39 | wayneeseguin | ok |
| 19:53:19 | evan | this might be a bug in libev |
| 19:53:37 | evan | because I see a similar pattern in all the libtool's |
| 19:53:38 | wayneeseguin | This is encouraging... it's actually working now that I am not calling make_flags |
| 19:53:55 | wayneeseguin | well compiling at any rate |
| 19:55:10 | evan | i'm not sure whats up with libtool |
| 19:55:22 | evan | but i don't have the patience (as in, in my whole life) to debug libtool. |
| 19:55:37 | brixen | wayneeseguin: you're trying to use a single set of flags to compile every ruby impl? |
| 19:55:45 | wayneeseguin | evan: yeah tha'ts not worth it |
| 19:56:11 | wayneeseguin | brixen: no only the MRI based ones on OSX. I made the mistake of putting it in front of all ruby interpreter's compilation |
| 19:56:45 | brixen | wayneeseguin: oh ok |
| 19:57:12 | wayneeseguin | This is awesome, I think that we'll be able to have people using rbx under rvm again :) |
| 19:57:23 | evan | ok guys |
| 19:57:26 | wayneeseguin | w00t |
| 19:57:27 | wayneeseguin | ∴ ruby -v |
| 19:57:28 | wayneeseguin | rubinius 0.13.0 (1.8.7 2bbf28bf 2009-11-06) [x86_64-apple-darwin10.0.0] |
| 19:57:28 | evan | i'm going to go watch some USC football. |
| 19:57:33 | wayneeseguin | evan: enjoy !! |
| 19:57:34 | wayneeseguin | Thank you |
| 19:57:36 | evan | later! |
| 19:57:37 | evan | np. |
| 20:15:21 | Zoxc | Finally my test suite passes :D |