Da
Show enters and exits. Hide enters and exits.
| 00:01:20 | evan | Defiler: I HATE $! and rescue nesting. |
| 00:03:23 | tenderlove | evan: why do you hate the dollar so much? I mean, I know America isn't the best country, but it's still *our* country |
| 00:03:36 | evan | thats $$ |
| 00:03:40 | evan | i'm a fan. |
| 00:03:42 | evan | this is $! |
| 00:03:51 | evan | perhaps i should say |
| 00:03:52 | evan | this is $!. |
| 00:03:57 | tenderlove | OH |
| 00:04:03 | tenderlove | ;-) |
| 00:04:06 | evan | :) |
| 00:04:24 | tenderlove | sorry, it's almost christmas holiday. I can't pay attention to work |
| 00:04:26 | tenderlove | :-( |
| 00:04:41 | evan | it happens. |
| 00:29:42 | mistergibson | ah tenderlove, but if the money isn't really yours ... is it really your country? |
| 00:30:16 | mistergibson | looks to me like it all belongs to the Fed Reserve Cartel |
| 00:51:40 | boyscout | Enable passing multiple files to #rm_r helper. - c305508 - Brian Ford |
| 00:51:41 | boyscout | Added #cp helper to mspec. - f99285a - Brian Ford |
| 00:51:41 | boyscout | Removed remaining uses of FileUtils in specs. - 6ad8b07 - Brian Ford |
| 00:51:41 | boyscout | Clean up use of #rm_r helper. - 865388c - Brian Ford |
| 00:54:38 | boyscout | CI: Build 865388c failed. http://ci.rubini.us/rubinius/builds/865388c7b7afe6c0e886504b451998eb1eee2b44 |
| 00:55:21 | brixen | what? |
| 00:58:51 | brixen | oh hrm, did boyscout not report earlier? |
| 00:59:08 | evan | i thought he did |
| 00:59:26 | brixen | dbussink's commit did not pass |
| 00:59:34 | brixen | the one removing Enumerator tags |
| 00:59:49 | evan | bad dbussink! |
| 00:59:51 | evan | can you fix it? |
| 00:59:55 | brixen | which is odd |
| 00:59:58 | brixen | I'm building on elle now |
| 00:59:59 | evan | he went to bed i think. |
| 01:00:01 | evan | k. |
| 01:00:10 | brixen | there's a second failure for what I just pushed |
| 01:01:39 | brixen | why would this enumerator spec fail on linux but not osx... |
| 01:02:20 | brixen | wtf, and this other spec fails now? |
| 01:02:59 | evan | :/ |
| 01:04:31 | brixen | stupid stupid #require and #load specs |
| 01:04:35 | brixen | le sigh |
| 01:04:43 | brixen | oh well, I can focus on them now |
| 01:05:18 | evan | did you figure out whats wrong? |
| 01:05:49 | brixen | still waiting for build to finish |
| 01:05:58 | brixen | to poke at the enumerator failure |
| 01:06:02 | evan | k |
| 01:06:02 | brixen | doesn't fail on osx |
| 01:08:02 | evan | so, i think the exception stack idea has the same problem as the current approach |
| 01:08:12 | evan | so i'm going to have to rethink this. |
| 01:09:31 | brixen | hmm |
| 01:09:48 | evan | because the stack requires logic to keep it consistent |
| 01:09:57 | justin-george | yeah the exception stack stuff is *retarded*. /me suppresses his hate |
| 01:10:28 | evan | and that logic has to run in rescue handlers |
| 01:10:37 | evan | justin-george: MRI has no stack |
| 01:10:48 | evan | well, it does |
| 01:10:52 | evan | sort of. |
| 01:10:55 | justin-george | heh... exactly. |
| 01:11:24 | justin-george | rescue Exception => e; raise e; end <-- maddening |
| 01:11:26 | brixen | holy shit, 74 sec for spec run from a clean build |
| 01:11:32 | evan | justin-george: whats wrong with that? |
| 01:11:46 | evan | brixen: niiice. |
| 01:11:49 | justin-george | evan: you get a new backtrace from where the exception was reraised |
| 01:12:07 | justin-george | whereas rescue Exception => e; raise; end does *not* give you a new backtrace |
| 01:12:07 | evan | oh the backtrace logic |
| 01:12:15 | evan | right, thats a bug in MRI imho. |
| 01:12:29 | evan | in rubinius, you can do |
| 01:12:32 | justin-george | and of course that's just the edge of the dark area that is exception handling |
| 01:12:34 | evan | rescue E => e |
| 01:12:46 | evan | raise Whatever, "more generic", e |
| 01:12:46 | evan | too |
| 01:12:50 | evan | chained exceptions! |
| 01:13:04 | evan | the new exception references the original one |
| 01:13:07 | evan | and both are shown |
| 01:13:07 | justin-george | yup, eagerly await it :) |
| 01:13:22 | justin-george | er, I mean, eagerly await being able to use that as the standard. |
| 01:13:37 | evan | yeah, need to figure out how to let people use it in standard code |
| 01:13:39 | boyscout | Removed nonsense cleanup code from #load specs. - bd6d466 - Brian Ford |
| 01:14:42 | brixen | I don't know why that enumerator spec failed for boyscout |
| 01:14:49 | brixen | I couldn't repro it on elle |
| 01:15:13 | evan | hrm. ok. |
| 01:15:20 | evan | we'll see shortly |
| 01:15:29 | evan | i can clean and redo if it still fails. |
| 01:15:38 | brixen | I love that the tmp files are all in the cwd now |
| 01:15:46 | brixen | I can run the specs and not clash with boyscout CI |
| 01:15:51 | evan | heh |
| 01:15:52 | evan | yeah |
| 01:15:58 | brixen | evan: I did a rake clean |
| 01:16:01 | brixen | for boyscout |
| 01:16:10 | evan | k |
| 01:20:47 | brixen | hah, just got that enumerator failure |
| 01:20:54 | brixen | 3 times no failure, then failure |
| 01:20:56 | brixen | wtf bitch |
| 01:21:07 | evan | WIERD. |
| 01:21:11 | evan | is it a JIT bug? |
| 01:21:14 | justin-george | heisenbug. |
| 01:21:32 | brixen | evan: ohh could be |
| 01:21:44 | brixen | I'll run with it off |
| 01:22:22 | brixen | I can't get it on osx |
| 01:25:26 | brixen | well, I don't get it with -Xint |
| 01:25:33 | evan | hm. |
| 01:25:34 | evan | ok |
| 01:28:56 | brixen | boyscout: where are you son? |
| 01:29:17 | brixen | http://ci.rubini.us/rubinius shows the run completed, but no report here |
| 01:29:22 | evan | hm. |
| 01:29:29 | evan | let me restart him |
| 01:29:43 | justin-george | nightall |
| 01:29:49 | evan | justin-george: nite |
| 01:29:51 | brixen | cya justin-george |
| 01:35:06 | mistergibson | odd thing. I have iconv.h installed at /usr/include - it is there. but I get errors attempting to install rbgccxml that it does not exist. output: http://pastie.org/755271 |
| 01:35:46 | mistergibson | any ideas? |
| 01:36:26 | evan | ummm |
| 01:36:33 | evan | hm. |
| 01:36:39 | mistergibson | odd huh |
| 01:37:15 | evan | did you install rbx? |
| 01:37:22 | mistergibson | yes |
| 01:37:24 | mistergibson | I'm running it |
| 01:37:40 | mistergibson | rbx gem install ... |
| 01:37:46 | evan | did you set your prefix to /usr ? |
| 01:37:56 | mistergibson | for rbx, yes |
| 01:37:59 | evan | ug. |
| 01:38:07 | evan | it might have hosed something |
| 01:38:08 | mistergibson | ug? it will be doing networking |
| 01:38:11 | mistergibson | ok |
| 01:38:17 | mistergibson | it needs its own root? |
| 01:38:19 | evan | because that means it created /usr/runtime |
| 01:38:28 | mistergibson | oh |
| 01:38:32 | mistergibson | hem |
| 01:38:36 | mistergibson | good to know |
| 01:38:47 | evan | it treated /usr as the root of the rubinius installation |
| 01:38:51 | evan | which is likely not what you want. |
| 01:38:53 | mistergibson | what is the recommended prefix? |
| 01:38:55 | evan | :/ |
| 01:38:58 | mistergibson | k |
| 01:39:08 | mistergibson | wha whaaa |
| 01:39:09 | evan | i do /usr/local/rubinius |
| 01:39:18 | mistergibson | but its going to be doing networking |
| 01:39:19 | evan | then link /usr/local/rubinius/bin/rbx to /usr/local/bin |
| 01:39:24 | evan | networking? |
| 01:39:25 | mistergibson | I'll put it in /opt |
| 01:39:27 | mistergibson | yes |
| 01:39:32 | mistergibson | all networked apps go into /usr |
| 01:39:37 | mistergibson | fhs ? |
| 01:39:37 | evan | ha. |
| 01:39:40 | brixen | well, I get the same thing installing into a dev dir, ie, no rbx install |
| 01:39:44 | evan | never heard that before. |
| 01:40:02 | mistergibson | might want to review the std. FHS for linux |
| 01:40:10 | mistergibson | but, its good to know |
| 01:40:11 | brixen | I have iconv.h in /usr/include |
| 01:40:30 | evan | rbx's install is not based on linux FHS. |
| 01:40:35 | brixen | why does it say "Gem files will remain installed in /Users/brian/devel/rubinius/gems/1.8/gems/nokogiri-1.4.1" |
| 01:40:36 | mistergibson | ok, so I need to go yard out all the rbx files I installed and reinstall in a sep. place |
| 01:40:39 | mistergibson | I see that |
| 01:40:42 | mistergibson | should be |
| 01:40:43 | brixen | something is messed up here |
| 01:40:45 | mistergibson | I think |
| 01:40:50 | mistergibson | oh well |
| 01:40:59 | evan | you should have done |
| 01:41:07 | evan | ./configure --prefix=/usr/local/rubinius |
| 01:41:10 | evan | rake install |
| 01:41:12 | mistergibson | I see |
| 01:41:14 | evan | is that what you did? |
| 01:41:24 | mistergibson | well, if its the same to you, I'll go with opt |
| 01:41:41 | evan | sure, but don't do /opt directly |
| 01:41:51 | mistergibson | at least that is fhs compliant |
| 01:42:03 | mistergibson | / opt / rbx ? |
| 01:42:14 | evan | sure |
| 01:42:18 | mistergibson | so, it basically needs its own root, got it |
| 01:42:51 | mistergibson | well, you might want to review that as most distros will follow the fhs |
| 01:43:16 | brixen | mistergibson: you can configure it to do a lot of special things |
| 01:43:20 | brixen | see configure --help |
| 01:43:20 | mistergibson | frankly, I'd like to evenually just use rbx and put links in so its callable like ruby |
| 01:43:24 | mistergibson | k |
| 01:43:29 | brixen | what's wrong with it being in it's own dir? |
| 01:43:37 | brixen | how does that contravene fhs? |
| 01:43:52 | mistergibson | not directly, I suppose |
| 01:43:52 | brixen | uninstalling is very easy then |
| 01:44:02 | brixen | rm -rf /opt/rbx |
| 01:44:05 | brixen | or whatever |
| 01:44:10 | mistergibson | yes, I see |
| 01:44:21 | mistergibson | well at least I know what I did |
| 01:44:36 | brixen | anyway, none of this has anything to do with rbgccxml not installing |
| 01:44:39 | mistergibson | I'm not sure what to do if it hosed a prior runtime file |
| 01:45:08 | brixen | I get the same error |
| 01:45:12 | mistergibson | well, I'm getting by implication that it may be the reason it cannot find iconv.h when it is in its search path |
| 01:45:14 | evan | mistergibson: did you build without --prefix initially |
| 01:45:15 | mistergibson | oh |
| 01:45:17 | evan | then reconfigure? |
| 01:45:21 | evan | and build |
| 01:45:28 | mistergibson | no, but i can - that's the easy part |
| 01:45:30 | evan | no no |
| 01:45:32 | evan | don't do that. |
| 01:45:37 | mistergibson | cleaning out what I have will take some care |
| 01:46:12 | mistergibson | I configured and built with --prefix=/usr initially |
| 01:46:48 | evan | mistergibson: do you do that with most software? |
| 01:46:59 | evan | i've been a sysadmin a long time |
| 01:47:05 | evan | and i never install custom stuff into /usr |
| 01:47:06 | mistergibson | yes, if it is a network capable thing |
| 01:47:09 | evan | it's impossible to track |
| 01:47:23 | evan | i don't get what networking has to do with it? |
| 01:47:26 | mistergibson | I guess I'm used to thinking of it as 'ruby' |
| 01:47:35 | mistergibson | which hardly seems custom, my mistake |
| 01:47:57 | mistergibson | again, all networking capable apps go into /usr |
| 01:48:02 | mistergibson | unless /opt |
| 01:48:26 | evan | why? |
| 01:48:38 | evan | that seems very arbritrary |
| 01:48:40 | mistergibson | fhs, or my understanding of it - that's what I was taught |
| 01:48:48 | mistergibson | no, quite standard |
| 01:49:00 | mistergibson | I've been working under that concept for years |
| 01:49:09 | mistergibson | why am I defending it? |
| 01:49:25 | evan | i work hard to not mix package installed and non-package installed on linux |
| 01:49:26 | mistergibson | its what I understand distros to adhere to as a standard file layout |
| 01:49:40 | mistergibson | I see |
| 01:49:53 | evan | because having "random" stuff /usr/bin is just a mess. |
| 01:49:58 | mistergibson | I guess its because I don't quite understand the classification of rubinius |
| 01:50:02 | evan | imho, /usr should be managed by the system entirely |
| 01:50:12 | mistergibson | I see |
| 01:50:21 | evan | it's classification doesn't really matter |
| 01:50:27 | evan | if you're buliding it yourself, it's not from a package |
| 01:50:31 | evan | and thusly would not go into /usr |
| 01:50:38 | mistergibson | well, as far as I know: /usr/local is for things that run local and do not use networking |
| 01:50:41 | evan | at least, thats how I manage things. |
| 01:50:44 | evan | oh |
| 01:50:45 | evan | no no |
| 01:50:45 | mistergibson | k |
| 01:50:49 | evan | thats not what local means |
| 01:51:02 | evan | the local in /usr/local means "locally installed" |
| 01:51:06 | evan | ie, custom |
| 01:51:11 | evan | not from packages |
| 01:51:12 | mistergibson | I see |
| 01:51:30 | evan | you using redhat? ubuntu? debian? |
| 01:51:38 | mistergibson | I thought /opt was for custom apps |
| 01:52:03 | evan | people use /opt and /usr/local, yes |
| 01:52:09 | evan | i tend to use /usr/local |
| 01:52:18 | evan | because /usr is on a different partition than / |
| 01:52:22 | evan | and / is usually pretty small. |
| 01:52:42 | evan | thought i've noticed most people don't use different mount points anymore |
| 01:52:45 | mistergibson | well, by way of repair ... what about this /usr/runtime ? is that a special file that may have been over written? |
| 01:52:53 | mistergibson | I do |
| 01:53:08 | mistergibson | old schoo I suppose |
| 01:53:09 | evan | should be a few directories with .rbc files |
| 01:53:14 | evan | you should be able to remove the whole directory |
| 01:53:20 | evan | check in your build's runtime/ dir |
| 01:53:30 | evan | you can see everything it installed. |
| 01:53:30 | mistergibson | k, thanks |
| 01:53:38 | mistergibson | I'll go clean it up and do it right |
| 01:53:41 | mistergibson | cool |
| 01:56:04 | brixen | mistergibson: you'll have a mess of files in /usr/lib now too |
| 01:56:16 | evan | ooh ewww |
| 01:56:16 | evan | yeah. |
| 01:56:18 | evan | :( |
| 01:56:25 | mistergibson | :{ |
| 01:56:28 | mistergibson | :P |
| 01:57:11 | mistergibson | safe to rm *.rb *.rbc ? |
| 01:57:15 | evan | yeah |
| 01:57:18 | evan | but there are directories too |
| 01:57:25 | evan | you can see them in build_dir/lib |
| 01:57:37 | evan | brixen: maybe if prefix is passed a directory |
| 01:57:44 | evan | it should check to see if there is a bin/lib/ etc |
| 01:58:05 | evan | and use a distributed directory setup |
| 01:58:15 | evan | something to think about |
| 01:59:02 | brixen | evan: yeah, could |
| 01:59:16 | brixen | evan: assuming every convention has the same /lib /bin etc |
| 01:59:34 | brixen | might be better to just add rubinius if only --prefix is given |
| 01:59:44 | brixen | unless it ends in rbx|rubinius |
| 01:59:52 | evan | brixen: sure |
| 01:59:56 | brixen | there are individual options for the FHS ppl |
| 01:59:59 | evan | or if --prefix is a directory that doesn't exist |
| 02:00:04 | brixen | sure |
| 02:00:11 | brixen | that's probably the best |
| 02:00:13 | evan | if it does exist, add rubinius on the end |
| 02:00:19 | brixen | yeah |
| 02:00:19 | evan | configure needs to tell the user that too |
| 02:00:24 | brixen | k |
| 02:00:26 | mistergibson | well, perhaps my mishap will do some good afterall :) |
| 02:00:39 | evan | "hey, i'm not going ot litter things around, so i've tried my best" |
| 02:00:44 | brixen | mistergibson: yeah, we try hard to learn from other's mistakes :) |
| 02:00:57 | mistergibson | better than pulling arrows out of yer butt |
| 02:01:02 | brixen | heh |
| 02:01:08 | evan | mistergibson: hehe |
| 02:08:06 | mistergibson | odd rubinius isn't in the ubuntu pkg manager |
| 02:08:15 | mistergibson | ought to be |
| 02:08:28 | mistergibson | trust me, there are *far* less supported packages in there |
| 02:08:55 | evan | mistergibson: just a case of no one putting it in |
| 02:09:04 | mistergibson | odd |
| 02:09:07 | mistergibson | k |
| 02:25:15 | boyscout | Made ./configure --prefix smarter. - 7a074d2 - Brian Ford |
| 02:28:12 | boyscout | CI: Build 7a074d2 failed. http://ci.rubini.us/rubinius/builds/7a074d2c79ab7fee54a6aec915f0d05fa9417a2e |
| 02:28:34 | brixen | let me guess |
| 02:28:52 | evan | great. |
| 02:28:55 | evan | :/ |
| 02:29:07 | brixen | something to do with arguments passed to a method |
| 02:29:31 | brixen | dbussink removed a (*args) from the Enumerator#each_with_index |
| 02:29:48 | evan | maybe we should revert that |
| 02:29:49 | mistergibson | hrm |
| 02:29:58 | mistergibson | still getting that err gem installing ... same thing |
| 02:30:00 | evan | because thats when it started to go wrong. |
| 02:30:10 | mistergibson | I even installed iconv into the rbx prefix |
| 02:30:17 | brixen | mistergibson: yeah, I assured you it had nothing to do with installing rbx ;) |
| 02:30:19 | evan | mistergibson: NO |
| 02:30:23 | evan | don't do that?! |
| 02:30:24 | mistergibson | ok |
| 02:30:36 | evan | don't put random files into the rubinius directory |
| 02:30:38 | mistergibson | just dorking it up all over the place huh? |
| 02:30:45 | brixen | heh |
| 02:30:46 | mistergibson | brixen, what do you think it is? |
| 02:30:57 | evan | not sure whats up |
| 02:31:01 | evan | try install nokogiri directly |
| 02:31:01 | brixen | mistergibson: could be something wrong with mkmf running in rbx |
| 02:31:05 | mistergibson | k |
| 02:31:11 | brixen | wait! |
| 02:31:12 | mistergibson | ok, will do |
| 02:31:14 | mistergibson | oh? |
| 02:31:16 | brixen | this is not nokogiri |
| 02:31:23 | brixen | this is rbgccxml yes? |
| 02:31:30 | mistergibson | nokogiri is a dep of ... right |
| 02:31:35 | evan | i think it's acutally nokogiri |
| 02:31:35 | brixen | ok |
| 02:31:41 | evan | which should be working |
| 02:31:44 | mistergibson | y |
| 02:31:47 | brixen | so, why was nokogiri working before |
| 02:31:48 | brixen | yeah |
| 02:32:00 | mistergibson | I'll try it directly |
| 02:32:05 | brixen | I'm trying it directly too |
| 02:32:14 | brixen | yeah, it is, ok |
| 02:32:26 | brixen | what's up noko, beotch? |
| 02:32:29 | brixen | hmm |
| 02:32:31 | mistergibson | hehe |
| 02:32:48 | mistergibson | same err |
| 02:32:58 | brixen | yeah |
| 02:33:01 | mistergibson | cannot find a .h file right in its search path |
| 02:33:10 | evan | it worked yesterday.... |
| 02:33:15 | mistergibson | mkmf? |
| 02:33:22 | evan | installing nokogiri |
| 02:33:23 | evan | so yes. |
| 02:33:26 | brixen | where said iconv.h does in fact exist in the *listed* dirs |
| 02:33:54 | brixen | evan: dbussink's change seems harmless, and the failure is not consistent |
| 02:33:59 | brixen | evan: still want me to revert? |
| 02:34:04 | evan | well |
| 02:34:16 | evan | i'm not going to have a chance to look for a few days |
| 02:34:21 | evan | headed to Bozeman tomorrow morning |
| 02:34:26 | mistergibson | let me know what you decide - I'll just bork my 'local' branch pull and rebranch |
| 02:34:26 | brixen | ah yes |
| 02:34:34 | evan | so i'd rather have it stable |
| 02:34:35 | brixen | mistergibson: I'll look at it in a bit |
| 02:34:38 | mistergibson | ok |
| 02:34:43 | mistergibson | thanks for the effort |
| 02:34:45 | mistergibson | :) |
| 02:34:47 | brixen | n/p |
| 02:34:51 | mistergibson | now, THAT's support |
| 02:34:58 | brixen | mistergibson: oh, what's your platform btw? |
| 02:35:06 | mistergibson | Ubuntu Karmic |
| 02:35:08 | brixen | k |
| 02:35:09 | mistergibson | 32-bit |
| 02:35:15 | brixen | I've got that in a vbox |
| 02:35:18 | brixen | I think |
| 02:35:24 | mistergibson | I like vbox |
| 02:35:24 | brixen | oh, maybe 64bit vbox |
| 02:35:28 | mistergibson | fancy |
| 02:35:37 | mistergibson | some day I'll have a 64bit unit |
| 02:35:49 | brixen | I'm on a mbp |
| 02:35:57 | brixen | vbox is very handy |
| 02:36:01 | mistergibson | the used ones are starting to show up on the shelves |
| 02:36:04 | mistergibson | y |
| 02:36:25 | mistergibson | never quite got mac osx on it though - no pae mode |
| 02:36:57 | mistergibson | looking to xen now or some such |
| 02:37:02 | brixen | evan: I say probably not a big deal, there won't be much activity and I can see it's that one spec failing |
| 02:37:15 | evan | ok |
| 02:37:20 | brixen | evan: I could try to look at it too, just for shits n giggles n such |
| 02:37:25 | evan | sure |
| 02:37:34 | mistergibson | odd thing, I did run the mspec:all before install .. and did get failures |
| 02:37:40 | evan | it's possibel i can look at tomorrow evening |
| 02:37:41 | mistergibson | didn't mention it though |
| 02:37:46 | evan | i get into Bozeman at like noon |
| 02:37:53 | evan | because the flight is at 6am :/ |
| 02:37:56 | brixen | mistergibson: mspec:all? what is that? |
| 02:38:04 | mistergibson | not literally |
| 02:38:07 | mistergibson | rake mpec |
| 02:38:08 | brixen | evan: yikes! are you packed? |
| 02:38:12 | brixen | mistergibson: ah ok |
| 02:38:14 | evan | brixen: not yet |
| 02:38:18 | brixen | mistergibson: you got failures from that? |
| 02:38:22 | mistergibson | y |
| 02:38:23 | brixen | evan: go pack! hah |
| 02:38:26 | evan | hah. |
| 02:38:28 | mistergibson | I can redo it |
| 02:38:31 | evan | going to an xmas dinner first |
| 02:38:33 | brixen | evan: you should be asleep son :D |
| 02:38:34 | evan | then pack |
| 02:38:36 | brixen | ok |
| 02:38:39 | evan | i'm a late packer |
| 02:38:45 | evan | it only takes me about 10 minutes anyway |
| 02:38:55 | mistergibson | I tend to travel with one bag ... easy squeezy |
| 02:39:01 | mistergibson | nice |
| 02:39:07 | mistergibson | light has its advantages |
| 02:39:14 | brixen | evan: don't forget your mittens and hat |
| 02:41:21 | evan | I shan't! |
| 02:41:24 | evan | i'm off to dinner. |
| 02:41:28 | brixen | k |
| 02:41:32 | brixen | headed out myself |
| 02:42:46 | brixen | I definitely have libiconv installed via macports, so wtf |
| 02:42:52 | brixen | anyway, I'll poke at it later |
| 02:45:14 | mistergibson | brixen: should I just use the tar ball for now? |
| 02:45:26 | mistergibson | rc1 |
| 02:46:20 | brixen | you can try |
| 02:46:28 | brixen | but there's a number of fixes in master |
| 02:46:32 | mistergibson | oh |
| 02:46:46 | brixen | we're close to releasing rc2 |
| 02:46:49 | mistergibson | hrm, this would be one of those wait for the soup moments then |
| 02:46:55 | brixen | yeah |
| 02:47:00 | mistergibson | no worries |
| 02:47:06 | brixen | or hack on something else |
| 02:47:10 | brixen | rbx related |
| 02:47:30 | mistergibson | yeah, like to ride that christmas morning thing all at once I suppose |
| 02:48:37 | mistergibson | np, I can wait |
| 03:26:51 | Defiler | evan: exceptions being accessible/determined/known/aaaaaaah when they are not lexically in scope is so evil |
| 03:27:01 | Defiler | The darkest of pacts |
| 05:28:25 | brixen | ok, we have definitely broken something |
| 05:28:45 | brixen | I can install nokogiri at 7a074d2c |
| 05:28:55 | brixen | looks like something with String perhaps |
| 05:29:10 | brixen | the mkmf discovery code is borked |
| 05:29:31 | brixen | and it looks like the string for the program to check for iconv.h is garbage |
| 05:31:29 | brixen | hrm lots of commits since then |
| 07:32:39 | brixen | bingo |
| 07:33:00 | brixen | 7128425b |
| 07:34:37 | brixen | now to figure out what is broken |
| 08:58:10 | dbussink | brixen: my change triggered some bug? |
| 09:08:49 | boyscout | Change Struct#inspect to use Class#inspect instead of Class#name - 4d48208 - Dirkjan Bussink |
| 09:08:49 | boyscout | Remove tags for now passing Struct#inspect and Struct#to_s specs - afac246 - Dirkjan Bussink |
| 09:11:48 | boyscout | CI: afac246 success. 3019 files, 11628 examples, 35765 expectations, 0 failures, 0 errors |
| 09:25:01 | boyscout | Remove tags for now passing Math specs - b458290 - Dirkjan Bussink |
| 09:27:59 | boyscout | CI: Build b458290 failed. http://ci.rubini.us/rubinius/builds/b4582904b6c52285781bbe0e7dceed96f7dd57e2 |
| 09:38:11 | dbussink | brixen: is that the failure you were discussing about with evan? |
| 09:38:26 | dbussink | brixen: that only sometimes occurs? |
| 13:10:15 | rue | Who knows |
| 17:17:06 | brixen | dbussink: yes |
| 17:17:13 | brixen | it first appeared here http://ci.rubini.us/rubinius/builds/13e05f0bbb5908e7a353ca3cd179748c10122cdc |
| 17:40:34 | dbussink | evening |
| 17:55:29 | brixen | dbussink: fyi, if you are removing tags, make sure you really know what you are doing if they are tagged unstable |
| 17:55:43 | brixen | there's a reason I go to the trouble to mark them unstable |
| 17:55:54 | brixen | by definition, that means they sometimes pass |
| 17:56:01 | brixen | ergo, they sometimes fail |
| 17:56:09 | dbussink | brixen: well, that's why i've ran them on two different platforms and didn't see any issues even after multiple passes |
| 17:56:18 | brixen | hence, if they pass for you, how do you know they are fixed |
| 17:56:23 | brixen | unless you fixed it |
| 17:56:44 | brixen | unstable doesn't necessarily have anything to do with platform |
| 17:56:50 | brixen | it may, but doesn't have to |
| 17:56:55 | brixen | just an FYI |
| 17:57:21 | dbussink | brixen: do you want me to put the tag back? |
| 17:57:28 | brixen | I'm doing that right now |
| 17:57:36 | brixen | waiting for you changes to build |
| 18:00:41 | boyscout | Tag unstable Enumerator#each spec. - 85ac7fb - Brian Ford |
| 18:17:23 | brixen | wtf boyscout |
| 18:32:32 | mistergibson | hey brixen |
| 18:35:16 | brixen | hey mistergibson |
| 18:35:24 | brixen | I'm fixing the nokogiri issue |
| 18:37:35 | mistergibson | thanks for doing that |
| 18:37:49 | mistergibson | so, what was the deal with mkmf? |
| 18:38:05 | brixen | it's not mkmf specifically |
| 18:38:09 | mistergibson | oh |
| 18:38:12 | brixen | it's our string_build instruction |
| 18:38:16 | mistergibson | I see |
| 18:38:22 | brixen | mkmf is just manifesting the issue |
| 18:38:33 | brixen | I'm still tracking down the specific problem |
| 18:38:37 | mistergibson | symptom area, cause ... right |
| 18:38:42 | brixen | but it's definitely related to string interpolation |
| 18:38:49 | mistergibson | ok, I'll shoosh |
| 18:38:56 | brixen | heh no worries |
| 19:14:51 | mistergibson | just give me the high sign and I'll do a pull on master |
| 19:29:35 | mistergibson | do you want me to run and pastie the mspec test results after pull/bulid? |
| 19:31:46 | brixen | sure |
| 19:31:51 | mistergibson | ok |
| 19:32:03 | mistergibson | who did the garbage collector stuff? |
| 19:33:02 | brixen | evan |
| 19:33:05 | mistergibson | ok |
| 19:44:46 | brixen | man, gem command massively uses string interp |
| 19:49:16 | brixen | this is amazing |
| 19:49:43 | brixen | installing nokogiri causes the string \Znokogiri\A to be created about a million times |
| 19:49:53 | brixen | give or take a few 100k |
| 19:52:16 | mistergibson | hehe |
| 20:46:35 | rue | \A -> \Z |
| 22:32:45 | brixen | mkmf.rb is the most WJW code I have *ever* seen |
| 22:32:46 | brixen | http://gist.github.com/263396 |
| 22:37:00 | rue | Yes, this is partly why I suggested replacing it |
| 22:37:21 | brixen | I'll suggest in one hand and shit in the other... |
| 22:37:22 | brixen | :) |
| 22:38:06 | rue | Never pursued it since the idea was shot down immediately. |
| 22:38:28 | brixen | haha |
| 22:41:17 | rue | It stops benig funny after a while |
| 22:41:24 | rue | Being* |
| 22:42:31 | brixen | I'm glad we didn't stop working on rbx when all those ppl shot down the idea of fast ruby written with ruby |
| 22:44:27 | yakischloba | clearly a trash idea |
| 22:44:34 | brixen | most def |
| 22:44:35 | yakischloba | impossible, raelly |
| 22:45:31 | Zoxc | define: fast |
| 22:47:15 | Zoxc | how is argument passings done in rbx? |
| 22:47:26 | brixen | carefully |
| 22:47:58 | Zoxc | do you turn the arguments into a low level array or can it be stored on the stack (with the JIT for example)? |
| 22:48:40 | brixen | look in vm/instructions.def |
| 22:48:45 | brixen | eg send_stack insn |
| 22:57:27 | Zoxc | hm.. so how do you recover the stack space if there is a jumping instruction in the middle of the arguments? |
| 22:57:47 | brixen | show me what that looks like in ruby code |
| 23:07:45 | Zoxc | something like this: http://pastie.org/756025 |
| 23:08:40 | Zoxc | it's unlikely to appear in real code, but it's kinda bad if it's in a loop |
| 23:09:05 | brixen | and what ruby are you using in which that code even parses? |
| 23:09:38 | brixen | syntax error == unlikely to appear in real code for sure |
| 23:10:38 | brixen | I guess it does in 1.9 |
| 23:10:51 | brixen | but you get a warning, statement not reached |
| 23:12:02 | brixen | anyway, bin/rbx compile -B file.rb to see what the bytecode would actually be |
| 23:30:31 | mistergibson | is there a ready made replacement for mkmf ? |
| 23:33:14 | brixen | nope |
| 23:33:35 | mistergibson | bummer |
| 23:33:36 | mistergibson | ok |
| 23:34:11 | mistergibson | is any discussing a windows port of rubinius (non-cygwin)? |
| 23:35:54 | brixen | just need a windows dev to work on it |
| 23:36:10 | brixen | we've removed almost all deps that would be problematic |
| 23:36:39 | brixen | it's not a priority pre 1.0 final |
| 23:36:48 | brixen | but we'll get to it eventually |
| 23:37:09 | brixen | I should say, not a priority because we only have ppl asking about it |
| 23:37:22 | brixen | no contributors actually willing to work on it |
| 23:37:59 | yakischloba | in the ruby community? shocking ;) |
| 23:38:11 | brixen | yakischloba: hehe |
| 23:38:29 | brixen | I'm sure windows 7 will be so much more fun |
| 23:38:33 | brixen | I can't wait |
| 23:38:36 | mistergibson | hehe |
| 23:38:43 | yakischloba | heh. |
| 23:38:46 | brixen | evan is supposed to be getting me a license actually |
| 23:39:06 | mistergibson | its not that I use windows much at all, but if I needed to deploy on it - good to know its on the table for later |
| 23:39:11 | brixen | massively overpriced POS that it is |
| 23:39:12 | mistergibson | no biggie for me |
| 23:39:45 | brixen | I bought SL for 29USD |
| 23:39:53 | brixen | that is amazing to me |
| 23:39:56 | mistergibson | brixen: how dare you complain you dead peasant, just bend over - yes you may have another - and the price is going up |
| 23:40:05 | mistergibson | and smile dammit |
| 23:40:09 | brixen | heh |
| 23:40:11 | brixen | ear-to-ear |
| 23:40:20 | brixen | every time I call windoze a POS :D |
| 23:41:00 | brixen | I've stopped talking about m$ directly, no words do it justice |
| 23:41:46 | mistergibson | wine is as close as I get to them anymore - just to play games. Someday >sigh< even that will not be needed. |
| 23:42:13 | brixen | yeah, the memories |
| 23:42:35 | brixen | I think I first ran notepad.exe in wine in like '95 or something? |
| 23:42:44 | brixen | I was amazed |
| 23:42:55 | mistergibson | nice |
| 23:43:29 | yakischloba | hmm. i built rbx for the first time in a long time this morning and encountered some crazy segfault explosion and didnt bother to copy it down because i naively assumed it would be reproduceable. no such luck now :p |
| 23:43:36 | mistergibson | I miss GeoWorks Ensemble (yes, I'm one of those) - really liked the idea of object assembly |
| 23:43:48 | brixen | yakischloba: eeks |
| 23:44:43 | yakischloba | yea…im running the specs over and over trying to get it again but i have done about 20 runs now |
| 23:45:44 | brixen | heh |
| 23:45:52 | brixen | maybe you imagined it, just sayin' :D |
| 23:46:00 | yakischloba | :) |
| 23:46:33 | brixen | maybe you accidentally invoked notepad.exe :) |
| 23:46:43 | yakischloba | heh |
| 23:46:53 | mistergibson | when setting up a project, anyone have a spiffy n00b primer on properly mspec'ing one's code? |
| 23:47:09 | brixen | mistergibson: rspec has pretty good docs |
| 23:47:14 | mistergibson | k |
| 23:47:24 | brixen | mspec is just a custom framework that follows most of rspec |
| 23:47:28 | brixen | it's not special |
| 23:47:34 | mistergibson | ok, gotcha |
| 23:47:57 | brixen | and david chelimsky has a book on pragmatic programmers you should be able to get in pdf |
| 23:48:06 | brixen | it's not in print yet afaik |
| 23:48:41 | brixen | mistergibson: http://rspec.info/ |
| 23:49:07 | mistergibson | thanks |
| 23:49:19 | brixen | we should be running rspec too, so rbx gem install rspec |
| 23:49:25 | brixen | and let us know if you have issues |
| 23:50:21 | mistergibson | still awaiting mkmf reform push ... can it be made to behave? |
| 23:50:30 | brixen | yeah, it can |
| 23:50:52 | brixen | it actually has nothing to do with mkmf, other than that crazy code exposed a bug in the compiler :) |
| 23:51:02 | mistergibson | interesting |
| 23:51:12 | mistergibson | cool actually |
| 23:51:42 | brixen | yeah, nice that some programmer sacrificed his soul to SATAN so we can be better |
| 23:51:45 | brixen | :) |
| 23:51:53 | mistergibson | hehe |
| 23:52:08 | mistergibson | ew |
| 23:55:00 | brixen | this damn node has so much logic now ugh |
| 23:55:25 | mistergibson | getting pudgy? |
| 23:56:24 | brixen | yes |