Show enters and exits. Hide enters and exits.
| 00:00:07 | dgtized | oh you updated the .rb code, compiled, then changed the stables, compiled, and then swap the shotgun code? |
| 00:01:00 | ctennis leaves the room. | |
| 00:01:15 | brixen | I changed RET(Qfalse) to RAISE_FROM.., and changed #close to be: io_close; nil |
| 00:01:26 | brixen | specs pass, so I'm rebuilding so I can build stables |
| 00:01:39 | brixen | then I'll try changing #close to just io_close |
| 00:01:40 | agile leaves the room. | |
| 00:01:52 | brixen | and change the primitive to RET(Qnil) |
| 00:04:23 | brixen | what I'm not clear about is why we need io_close and #close |
| 00:04:32 | brixen | @descriptor is a field, so accessible from C |
| 00:04:39 | brixen | why not just IO#close as a primitive |
| 00:05:45 | dgtized | hmm |
| 00:06:24 | brixen | perhaps I should just try that |
| 00:09:12 | brixen | dgtized: what did you determine about why IO is in bootstrap/primitives.rb ? |
| 00:09:41 | dgtized | I determined that I was supposedly supposed to ask evan |
| 00:09:48 | brixen | heh |
| 00:09:57 | brixen | I think I shall try a little experiment :) |
| 00:10:04 | brixen | evan can worry about C++ for now :P |
| 00:10:26 | dgtized | I moved them out of bootstrap/primitive and into bootstrap/io and it all passed just fine |
| 00:10:42 | dgtized | but I was informed that this was somehow "not what I was supposed to do" |
| 00:10:47 | brixen | yeah, I would expect as much, there are no clear dependencies |
| 00:10:49 | brixen | rationale? |
| 00:11:08 | dgtized | rationale was that I should ask evan because he might be grouchy about it |
| 00:11:46 | brixen | well, he gets grouchy when things are broken |
| 00:11:57 | dgtized | right, and I didn't see why that would break anything |
| 00:12:12 | dgtized | I don't see any reason why it should matter if io_close gets defined before or after io.rb |
| 00:12:13 | brixen | historically, bootstrap was really bootstrap code, then we put all the primitives in it because headius was complaining they make the Ruby libs less usable |
| 00:12:28 | brixen | and we had issues with ordering dependencies |
| 00:12:41 | dgtized | brixen: right, I remember that, but when I swapped it over it was just happy |
| 00:12:42 | brixen | but evan explicitly told me I could use depends on: in bootstrap |
| 00:13:07 | brixen | yeah, primitives.rb is just a crime against sanity |
| 00:13:09 | dgtized | but I mean presumably primitive.rb gets loaded last anyway |
| 00:13:29 | brixen | nope, it gets loaded wherever it is in load_order.txt |
| 00:13:59 | brixen | which looks like 22nd on my system |
| 00:14:33 | dgtized | where is load order again? |
| 00:14:44 | brixen | good lord, there is actually bootstrap/io.rb |
| 00:14:45 | brixen | ugh |
| 00:14:57 | brixen | dgtized: runtime/bootstrap/.load_order.txt |
| 00:17:18 | dgtized | it looks like there are a few things that care in bootstrap, but it doesn't look like primitives to me |
| 00:18:27 | brixen | dgtized: yeah, le'see if we can make this a bit cleaner |
| 00:18:45 | brixen | for now, I'm just defining IO#close as Ruby.primitive :io_close_ng |
| 00:18:52 | brixen | we'll leave the existing primitive |
| 00:19:17 | dgtized | ah, that makes sense |
| 00:20:06 | dgtized | ok, well if you are pushing those changes, let me just revert my IO#close patch from my local tree |
| 00:21:08 | brixen | well, haven't found out if it works yet, so don't go revertin' stuff! ;) |
| 00:22:49 | dgtized | well it still leaves it my local repo history anyway |
| 00:24:40 | qwert666 leaves the room. | |
| 00:25:18 | rue | Huh, this screen sharing thing is pretty good |
| 00:25:20 | brixen | hmm, seems to be working |
| 00:25:31 | rue | Except the screen is about half the size |
| 00:25:41 | dgtized | to just point IO#close at the primitive? |
| 00:25:54 | brixen | rue: you know, knowledge is the only thing you can share and still have as much as before |
| 00:26:26 | rue | Except the evil idea thiefs |
| 00:26:28 | brixen | dgtized: yeah, basically |
| 00:28:03 | dgtized | and that gets around the need to update stables? |
| 00:28:43 | brixen | dgtized: well, for a bit. not changing the existing primitive's behavior means the stables don't need to change |
| 00:28:50 | brixen | eventually, we'll build new stables |
| 00:29:27 | rue | Otherwise you need to do it in parts; implement the new ones, roll new stables off those, then take down the old ones |
| 00:29:36 | brixen | what's so silly is that every class in bootstrap/primitives.rb already has a file in bootstrap |
| 00:31:46 | brixen | dgtized: I can give you the diff and tell you the next 2 steps if you want to commit all this |
| 00:32:00 | dgtized | sure |
| 00:32:02 | brixen | secretly tries to shift evan's grouchiness to dgtized :D |
| 00:32:21 | brixen | heh, j/k |
| 00:32:45 | dgtized | whatever, I think it's just no one else looked at how rediculous bootstrap/primitives.rb was |
| 00:32:49 | nolan_ enters the room. | |
| 00:33:06 | brixen | dgtized: one sec, doing a full clean build and I'll have a diff for you |
| 00:33:09 | dgtized | k |
| 00:33:32 | jayWHY_ leaves the room. | |
| 00:34:14 | nolan | Hey, not sure if this is documented anywhere. How do I change the install prefix for rubinius? Tried editing shotgun/config.mk, but rake install still wants to dump stuff in /usr/local. |
| 00:34:45 | brixen | nolan_: no one's tried that yet, afaik, but it's been moved to a rakefile |
| 00:35:13 | brixen | nolan_: are you using a clone or the daily tarball? |
| 00:35:57 | nolan | Clone. |
| 00:36:21 | brixen | k, look in rakelib/configure.rake |
| 00:36:35 | brixen | line 22, did you try changing that? |
| 00:40:00 | brixen | dgtized: I wasn't on head, so one more sec while I update |
| 00:40:04 | dgtized | k |
| 00:40:20 | dgtized | nice, you can add automatic coloring to .git/config |
| 00:40:29 | dgtized | so that git diff and git status are colored |
| 00:41:16 | nolan | No, but it looks like it just takes PREFIX from the environment--or, rather, it should. 'rake configure PREFIX=/Users/nolan/Packages/rbx' doesn't appear to fix the issue. Hmmm. |
| 00:42:33 | brixen | dgtized: nice, what did you add? |
| 00:43:02 | brixen | nolan_: odd, you could file a ticket if you want, and/or send a patch :) |
| 00:43:23 | wyhaines leaves the room. | |
| 00:43:42 | dgtized | brixen: http://pastie.org/173792 |
| 00:44:41 | brixen | sweet thanks |
| 00:44:51 | brixen | are there still 3 errors on head? |
| 00:45:32 | dgtized | I think so |
| 00:45:50 | dgtized | but you remove all 3 if you just tag out the one introduced by that spec |
| 00:46:01 | dgtized | the other two fall out of the before/after blocks |
| 00:46:17 | brixen | ok |
| 00:46:20 | brixen | dgtized: http://pastie.org/173794 |
| 00:51:11 | dgtized | presumably we should probably update stables once we have that checked in for everyone so that the old io_close is removed |
| 00:51:12 | brixen | dgtized: next step, I did a clean build; then rake build:stable so all files are updated |
| 00:51:39 | brixen | yeah, running through those steps locally now so you can commit it all together |
| 00:51:52 | brixen | so, commit that diff |
| 00:52:04 | brixen | rake clean build; rake build:stable; git commit |
| 00:52:05 | dgtized | let me add the spec tags to it |
| 00:52:07 | dgtized | but yea |
| 00:52:11 | brixen | k |
| 00:53:00 | dgtized | oh if we update stables twice we should be able to remove that method right? |
| 00:53:31 | brixen | yes |
| 00:54:54 | nolan | Aha, figured it out, looks like PREFIX needs to be specified on 'rake install', not 'rake configure'. Thanks for the pointer. |
| 00:55:03 | brixen | nolan_: sure |
| 00:56:30 | brixen | dgtized: it would take rebuilding stable 3x to rename io_close_ng to io_close, might as well leave it for now |
| 00:56:58 | brixen | dgtized: so, after rebuilding stables, you should be able to remove IO#io_close and io_close primitive |
| 00:57:44 | brixen | bbiab.. |
| 00:59:15 | nolan | What does "Unable to find a bootstrap to load!" mean when "rbx compile" fails? |
| 00:59:24 | nolan | This is during rake install. |
| 00:59:33 | crossblaim leaves the room. | |
| 01:01:24 | drbrain | nolan_: something is very wrong |
| 01:01:37 | drbrain | does plain 'rake' work? |
| 01:02:50 | nolan | Seems to be rebuilding. |
| 01:03:21 | d2dchat_ enters the room. | |
| 01:04:42 | loincloth leaves the room. | |
| 01:05:27 | drbrain | ok, then after that does install work? |
| 01:08:41 | nolan | No. I'll try cleaning and rebuilding everything again. |
| 01:10:30 | d2dchat_ leaves the room. | |
| 01:10:36 | drbrain | have you installed rubinius before? |
| 01:12:48 | d2dchat leaves the room. | |
| 01:17:13 | maxpenguin enters the room. | |
| 01:18:31 | rue | Mm, that could be issue; make sure you can first build normally from within the source tree |
| 01:19:05 | binary42 enters the room. | |
| 01:19:11 | justinweiss leaves the room. | |
| 01:19:40 | crayz__ leaves the room. | |
| 01:19:57 | nolan | OK, cleaning/rebuilding seems to have done the trick, thought I'd already done that but...*shrug* |
| 01:20:06 | crayz__ enters the room. | |
| 01:23:56 | rue | You know what the world really needs? |
| 01:23:56 | rue | A third OSS UNIX-like family |
| 01:24:28 | tarcieri | OpenSolaris? |
| 01:24:31 | tarcieri | heh |
| 01:24:32 | cored enters the room. | |
| 01:24:46 | brixen | rue: what the world needs is to utterly and completely remove every trace of Windows from the universe |
| 01:24:52 | tarcieri | indeed |
| 01:25:55 | Defiler | What we need is a cleansing fire that will sweep the galaxy |
| 01:26:01 | Defiler | A holy crusade against the machines |
| 01:26:47 | djwhitt | a butlerian jihad if you will... |
| 01:27:59 | agile enters the room. | |
| 01:28:04 | tarcieri | at least MS shot themselves in the foot with Vista |
| 01:28:08 | obvio enters the room. | |
| 01:29:54 | brixen | and yet the suffering of the masses continues |
| 01:30:01 | tarcieri | heh |
| 01:30:14 | brixen | I wish they had better aim, perhaps pointing that thing at the head :) |
| 01:30:32 | jayWHY enters the room. | |
| 01:30:43 | jayWHY leaves the room. | |
| 01:31:22 | dgtized | ugh |
| 01:31:24 | brixen | dgtized: how's it going |
| 01:31:31 | dgtized | still having trouble with bootstrap updates |
| 01:31:36 | dgtized | sorry stable updates |
| 01:31:49 | dgtized | I have it working as is |
| 01:31:58 | dgtized | but as soon as I try messing with build:stable everything breaks |
| 01:32:09 | brixen | hmm, worked here |
| 01:32:13 | brixen | take me through your steps |
| 01:33:15 | dgtized | just a sec, I have a vague idea that maybe my git tree isn't exactly the way I thought it was |
| 01:33:47 | brixen | k |
| 01:34:11 | brixen | you should be able to 1. apply patch, 2. build stables, 3. remove IO#io_close and io_close primitive |
| 01:34:16 | brixen | I'd leave it at that for now |
| 01:35:06 | nolan | OK, here's another, when I run 'rbx gem', I get 'cannot find file to load: rbconfig/datadir'. Any idea what's up with that? I don't see a rbconfig/datadir in the repository. |
| 01:35:20 | dgtized | brixen: that's the order I am trying, but a rebuild on the stables keeps breaking things |
| 01:35:47 | brixen | dgtized: after 1. do you rake clean build; rake build:stable ? |
| 01:36:14 | dgtized | brixen: right, but then if I do rake clean build again bin/mspec is broken |
| 01:36:27 | brixen | hmm I'll try |
| 01:36:38 | nolan | I get the same with rbx rar, only it's bytecode/compiler that isn't found. |
| 01:36:57 | rue | What about shotgun/rubinius gem ? |
| 01:37:40 | nolan | Same. |
| 01:38:04 | nolan | And "find . -name 'bytecode/compiler'" from the root of my clone returns nothing. |
| 01:40:18 | thewoolleyman leaves the room. | |
| 01:40:46 | rue | That is compiler 1 stuff |
| 01:41:15 | dgtized | brixen: yea, so if I have a stable setup, but as soon as I do a build:stable and then rebuild everything breaks -- I presume I need to do a build:stable / rebuild just to check to see if the new stables are happy for everyone else |
| 01:41:17 | rue | lib/compiler/bytecode.rb exists but if you see it the other way around, you have an installed version somewhere |
| 01:41:39 | brixen | dgtized: trying to reproduce now |
| 01:43:53 | dgtized | in a minute I will just push what I have thus far, since that runs, it doesn't require stable updates, and it does have the update, it's just messy in between |
| 01:46:45 | brixen | dgtized: ok |
| 01:46:59 | brixen | dgtized: I've reproduced it here, but have to leave for ruby group in a few |
| 01:47:05 | brixen | I'll work on it after you push |
| 01:47:52 | dgtized | I think I figured it out actually, but I haven't got that far |
| 01:48:04 | dgtized | at teh bottom of core/io.rb is private :io_close |
| 01:48:12 | dgtized | which gets compiled at a different phase |
| 01:48:28 | brixen | hmm, yeah, that could be a problem :) |
| 01:49:44 | lopex leaves the room. | |
| 01:51:40 | dgtized | alright let's see if that fixes it |
| 01:52:18 | dgtized | nope |
| 01:53:05 | dgtized | well maybe I updated the stables too early |
| 01:57:22 | boyscout | 4 commits by Charles Comstock |
| 01:57:23 | boyscout | * removed io_close from core/io.rb; b46817f |
| 01:57:24 | boyscout | * IO#reopen should return self; 9ee5251 |
| 01:57:25 | boyscout | * switched bootstrap IO#close to use io_close_ng and merged bootstrap/primitives.rb ...; a13983a |
| 01:57:26 | boyscout | * added primitive io_close_ng and tagged IO#close spec failures; f148128 |
| 01:57:44 | dgtized | eh someone else give a shot at updating stables from there |
| 02:06:23 | justinweiss enters the room. | |
| 02:06:25 | rubuildius_amd64 | Charles Comstock: b46817fa2; 1845 files, 6268 examples, 22361 expectations, 0 failures, 0 errors; http://rafb.net/p/sCrsqB49.html |
| 02:11:17 | rubuildius_ppc | Charles Comstock: b46817fa2; 1845 files, 6271 examples, 22390 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/173844 |
| 02:24:14 | VVSiz_ enters the room. | |
| 02:27:05 | roo enters the room. | |
| 02:31:37 | gdagley enters the room. | |
| 02:32:01 | gdagley leaves the room. | |
| 02:32:44 | ezmobius leaves the room. | |
| 02:35:51 | binary42 leaves the room. | |
| 02:41:05 | rue leaves the room. | |
| 02:42:24 | VVSiz leaves the room. | |
| 02:43:36 | bitbang leaves the room. | |
| 02:45:46 | macournoyer enters the room. | |
| 02:49:24 | agile leaves the room. | |
| 02:50:47 | KirinDav leaves the room. | |
| 03:02:21 | fbuilesv enters the room. | |
| 03:04:00 | RyanTM leaves the room. | |
| 03:07:32 | RyanTM enters the room. | |
| 03:09:16 | kofno enters the room. | |
| 03:13:48 | mkescher_ enters the room. | |
| 03:15:31 | mkescher_ leaves the room. | |
| 03:18:24 | mkescher enters the room. | |
| 03:19:11 | d2dchat enters the room. | |
| 03:20:22 | rue | Arglebargle |
| 03:22:22 | cored leaves the room. | |
| 03:27:00 | fbuilesv | Everyone talks so much about the blue Smalltalk book and I could only find the "purple" Smalltalk price :P |
| 03:27:21 | crafterm enters the room. | |
| 03:27:35 | djwhitt | http://users.ipa.net/~dwighth/smalltalk/bluebook/bluebook_imp_toc.html |
| 03:27:38 | djwhitt | there you go |
| 03:27:46 | djwhitt | that's the most important part for Rubinius stuff |
| 03:27:48 | fbuilesv | djwhitt: was a joke, I got an older edition that's actually purple :) |
| 03:28:01 | djwhitt | ah, heh |
| 03:28:12 | fbuilesv | looking for some book on my college library I found it looking straight at me :O |
| 03:28:12 | djwhitt | scarasm filter must be overloaded |
| 03:28:17 | djwhitt | April 1st and all |
| 03:28:43 | fbuilesv | I'm not reading blogs today, I'm still thinking that the Amazon interface for ActiveResource and SimpleDB is a joke :p |
| 03:40:42 | wycats | is there a program that'll show me what shlibs are installed and what their names are? |
| 03:42:15 | womble | wycats: There's always /etc/ld.so.cache |
| 03:42:17 | wycats | heh |
| 03:42:30 | wycats | I'm on OSX :P |
| 03:42:47 | wycats | I just need to figure out what the js shlib is actually called |
| 03:51:27 | rue | find |
| 03:53:06 | rue | VMWare performs surprisingly well |
| 03:53:40 | rue | wycats: I do not think darwin comes with ldd |
| 04:02:44 | GMFlash leaves the room. | |
| 04:02:50 | GMFlash enters the room. | |
| 04:05:30 | macournoyer leaves the room. | |
| 04:05:31 | antares | fbuilesv: there are 4 chapters from blue book online (absolutely legal btw) |
| 04:06:19 | antares | fbuilesv: but blue book seem to be more confusing when you come to look at Rubinius code, that's what I found |
| 04:06:54 | fbuilesv | antares: check this out |
| 04:06:56 | fbuilesv | antares: http://data.tumblr.com/PDVq61dc57au5o9nCkjLE5nA_400.jpg |
| 04:07:09 | fbuilesv | antares: that's why I said that it was purple, not "the blue" book |
| 04:07:34 | fbuilesv | antares: I'm reading it from the beginning and I'm seeing stuff that I thought was new in Ruby...almost 30 years late :) |
| 04:07:49 | antares | fbuilesv: http://stephane.ducasse.free.fr/FreeBooks/BlueBook/blueBook.jpg |
| 04:08:15 | antares | fbuilesv: read a lisp book and you'll know where symbols and ?C chars come from :) |
| 04:08:38 | fbuilesv | antares: damn, and I thought symbols did come from SM, time to trace back a bit more :S |
| 04:08:43 | antares | and open? / close! method names |
| 04:09:36 | fbuilesv | damn, it's a different book, only covers up to chapter 27 of the blue one |
| 04:09:39 | antares | fbuilesv: then read a python book and perl book and CLU paper... you'll find out about lambdas, <<HEREDOCS and blocks |
| 04:10:09 | fbuilesv | antares: SM also has blocks and anon. functions were in Lisp too, weren't they? |
| 04:10:12 | antares | Ruby basically invented nothing new but a great clean syntax for old ideas |
| 04:10:32 | antares | fbuilesv: Lisp is older than my father |
| 04:10:55 | antares | fbuilesv: John McCarthy started to work on it in 1958 I believe |
| 04:11:23 | antares | fbuilesv: in Smlltalk blocks are core of control flow |
| 04:11:35 | fbuilesv | antares: yes, around there. I started trying to learn Lisp with the little schemer but decidd to go Haskell :) |
| 04:12:20 | antares | fbuilesv: Haskell is the only relatively popular language I cannot read at all :) I even had Prolog in the university... |
| 04:12:28 | jtoy enters the room. | |
| 04:13:01 | antares | fbuilesv: there are chapters 26 to 30 online and a pdf of the whole blue book. Need those links? |
| 04:13:07 | fbuilesv | antares: I had a declarative course with Prolog too, that was awful |
| 04:13:42 | fbuilesv | antares: I'd appreciate that if the channel doesn't mind supporint piracy like that :S |
| 04:14:40 | antares | fbuilesv: these 5 chapters put online with permission from Adele Goldberg (http://users.ipa.net/~dwighth/smalltalk/bluebook/bluebook_imp_toc.html) |
| 04:14:53 | fbuilesv | antares: yes, I got that link already, I m eant the PDF |
| 04:15:32 | obvio leaves the room. | |
| 04:15:49 | antares | fbuilesv: I do not know about PDF. Smalltalk is an old horse. Here's the PDF: http://tinyurl.com/2gb57m |
| 04:15:54 | rue | antares, fbuilesv: The Blue Book is the foundation of Rubinius in the same way Picasso's models were to him |
| 04:15:56 | aotearoa enters the room. | |
| 04:16:17 | fbuilesv | antares: tyvm |
| 04:16:34 | fbuilesv | rue: what about making a music analogy? Guernica is as far as my picasso knowledge goes :P |
| 04:16:42 | antares | rue: blue book explains things like method context or block context or compiled method pretty well |
| 04:16:50 | antares | I am even more ignorant in arts |
| 04:22:10 | antares | fbuilesv: by the way, once you got familiar with smalltalk, ObjectiveC is absolutely trivial to learn |
| 04:22:50 | fbuilesv | antares: I thought about that earlier, the ObjC syntax actually seemed to make sense for a few minutes |
| 04:23:01 | rue | I wish evan had gone with ObjC :) |
| 04:23:46 | antares | rue: I second that. I do now know what it means for windows support though |
| 04:25:21 | fbuilesv | rue: I heard there's some C++ in there, do you mean it by that or just by the core C code? |
| 04:25:33 | rue | Hopefully its deathblow |
| 04:26:29 | antares | rue: well it seems that gcc provides support for objc on windows |
| 04:26:42 | antares | rue: we can even experiment with it :) |
| 04:27:33 | fbuilesv | I just found out today that there's competition: http://brobinius.org |
| 04:31:01 | rue | Haha |
| 04:31:27 | Fullmoon leaves the room. | |
| 04:31:47 | Rich_Morin_ enters the room. | |
| 04:34:12 | Rich_Morin | As I understand it, the ability to modify (eg, extend) the interpreter is one of the selling points of Rubinius. Is there a recommended way to modify the behavior of the parser (eg, from within a piece of code that's being parsed? :) |
| 04:35:08 | mkescher leaves the room. | |
| 04:35:29 | rue | Right now, no--we are stuck with a modified grammar.y for a little while still |
| 04:36:09 | rue | You can hook into the compiler as soon as the sexp is built, though. The ->AST and then ->bytecode is in Ruby |
| 04:36:47 | rue | I actually just abused the parse tree to enable conditional compilation |
| 04:37:00 | Rich_Morin | rue: Erm, are you the same "rue" I was chatting with a bit ago on ruby-lang? |
| 04:37:37 | rue | "Chatting" may be a strong term for my typical #ruby-lang-drive-bys |
| 04:38:00 | rue | But yeah :) |
| 04:40:10 | drbrain | dick |
| 04:40:28 | antares | haha I found a "week in ruby" post from march 31st |
| 04:41:18 | antares | I posted "week in rubinius" on 28th :) |
| 04:42:53 | hassox leaves the room. | |
| 04:43:48 | Rich_Morin | OK, here's the kind of thing I'd like to be able to do (eventually is fine :-); maybe we can discuss possible syntax... I'd like to be able to extend the interpreter to (say) record all of the tokens that could be either local variables or method names. How should I go about redefining the code that parses method definitions? |
| 04:44:44 | antares | Rich_Morin_: you can operate on compiled methods for that in Ruby |
| 04:44:49 | binary42 enters the room. | |
| 04:45:17 | MenTaLguY enters the room. | |
| 04:45:45 | Rich_Morin | antares: cool - can you give me a specific pointer? |
| 04:45:45 | rue | :vcall node is for lvar/method |
| 04:45:57 | agardiner | Rich_Morin_: antares is right - all of that is obtainable right now from a CompiledMethod, using #local_names and the literals tuple for SendSites |
| 04:46:11 | antares | Rich_Morin_: I recently documented stuff in kernel/core/compiled_method.rb, you may find it useful |
| 04:46:43 | agardiner | you can also use the debugger to view this... you might find the source useful |
| 04:46:52 | justinweiss leaves the room. | |
| 04:47:04 | antares | Rich_Morin_: what is the end goal for you, btw? |
| 04:47:26 | Rich_Morin | antares: Can you give me a URL; Google is failing me,... |
| 04:47:43 | antares | Rich_Morin_: it is in the source code of Rubinius |
| 04:47:50 | Rich_Morin | ah |
| 04:48:06 | rue | Well, gitweb too. |
| 04:48:37 | antares | Rich_Morin_: http://tinyurl.com/2t4vja |
| 04:48:52 | Rich_Morin | Has Rubinius switched to git? |
| 04:49:09 | agardiner | 6 months ago |
| 04:49:12 | antares | Rich_Morin_: a long time ago |
| 04:49:19 | agardiner | before it became de rigeur |
| 04:49:40 | Rich_Morin | Well a few days ago I was looking at a Rubinius web page that only talked about SVN, IIRC. |
| 04:49:46 | agardiner | rubinius made it popular! :0-D |
| 04:50:09 | antares | Rich_Morin_: website is rarely updated |
| 04:50:44 | Rich_Morin | We use git at my company (Reactrix), both for our Rails development and as a way to distribute content and collect logs. |
| 04:51:27 | Rich_Morin | Given that I might want to do really odd things with Rubinius, git could allow me to interact much more sanely... |
| 04:51:36 | antares | Rich_Morin_: cool, distributed VCSs are indeed very powerful |
| 04:52:25 | antares | Rich_Morin_: don't forget to come by this channel and tell us how it is going :) |
| 04:52:28 | Rich_Morin | I work with Nick Hengeveld and Scott Schacon, both of whom have contributed to git... |
| 04:52:33 | Rich_Morin | Sorry, Chacon |
| 04:53:17 | rue | Git is quite nice. We do not use it to its fullest potential here, actually, but there has not been a need to |
| 04:54:23 | Rich_Morin | I think it will be a Long Time before git is used to its fullest potential - it's a rather large framework. |
| 04:54:55 | antares | Rich_Morin_: there are backup systems and projects like gittorent already |
| 04:55:16 | Rich_Morin | BTW, Scott has been working on a Ruby API for git, which might be handy for you. |
| 04:55:41 | antares | Rich_Morin_: I saw such project at github: Ruby wrappers for libgit |
| 04:56:00 | Rich_Morin | antares: At a recent TWiki gathering, I wasn;t the only person bugging Peter Thoeny to switch from RCS to git. |
| 04:56:13 | Rich_Morin | might well be his |
| 04:56:16 | rue | mfp wrote gibak or whatever it is called |
| 04:56:23 | antares | and I have crazy ideas about using git for distribution system on my current project |
| 04:56:51 | antares | RCS? gosh |
| 04:57:00 | antares | I started from CVS... |
| 04:57:31 | womble | RCS is old sk00l |
| 04:57:41 | antares | rue: yes, Mauricio's project is called gibak (but didn't he used darcs for it ? ;)) |
| 04:58:44 | Rich_Morin | We have a couple of hundred "interactive advertising devices" out in the field. We need to feed content and new software to them, get log files (etc). Aside from reliability, git is quite efficient at transferring content that contains repeats of stuff that has already been sent. |
| 04:58:58 | Rich_Morin | No, SCCS is old school. |
| 04:59:48 | Rich_Morin | But TWiki has been around for a decade or so, so it's not surprising that it's based on oldish tech. |
| 05:00:01 | Fullmoon enters the room. | |
| 05:05:30 | nicksieger leaves the room. | |
| 05:06:37 | Rich_Morin | antares: Looks like CompiledMethod is a Rubinius thing only, given that MRI doesn't have byte codes... |
| 05:09:15 | obiejuan enters the room. | |
| 05:09:21 | drbrain | I just wrote a wiki using RCS |
| 05:09:43 | drbrain | http://blog.segment7.net/articles/2008/02/14/an-rdoc-wiki |
| 05:10:38 | antares | Rich_Morin_: yes, MRI does not have it. JRuby does. You asked about Rubinius from what I understood :) |
| 05:10:40 | lstoll | rcshub.com, anyous? |
| 05:10:47 | lstoll | anyone even |
| 05:11:15 | womble | lstoll: And sccshub.com |
| 05:11:20 | drbrain | lstoll: NFS over internet? |
| 05:11:21 | antares | last I heard github team is going back to SVN ;) |
| 05:11:31 | Rich_Morin | antares: yes, but it's always nice to know which tools are available where... |
| 05:12:14 | antares | cvshub.com looks good |
| 05:12:34 | wmoxam leaves the room. | |
| 05:12:48 | lstoll | drbrain: works for me. |
| 05:13:47 | Rich_Morin | antares: The base TWiki code has been has been refactored down to 17 KLOC. It may have a few features that rdoc_wiki omits. Still, cute hack! |
| 05:14:53 | antares | Rich_Morin_: what was original number? |
| 05:15:15 | Rich_Morin | dunno, but Peter could tell you. |
| 05:16:09 | Rich_Morin | My understanding is that this happened in the last few years, largely as a way to trim some plugins out of the core. |
| 05:16:30 | Rich_Morin | Akin to the Perl4 -> Per5 changes. |
| 05:17:37 | obiejuan leaves the room. | |
| 05:18:14 | antares | Rich_Morin_: do they now have twiki-core and twiki-more? |
| 05:18:23 | Rich_Morin | Getting git onto this machine now (so port install git-core). Works well, but may take a while... |
| 05:19:29 | antares | dmg installer of git recently has been slimmed to 4 mb (from 50) |
| 05:19:31 | nicksieger enters the room. | |
| 05:20:09 | Rich_Morin | I haven't heard them use those names. Basically, there's a core distro and a zillion plugins. We use quite a few here, as my spouse is a TWiki Champion. |
| 05:20:45 | Rich_Morin | antares: I'll let port cook for a while, since it's started. |
| 05:21:03 | antares | Rich_Morin_: alright |
| 05:21:32 | maxpenguin leaves the room. | |
| 05:23:05 | Rich_Morin | Interesting: It's grabbing all sorts of stuff - diffutile, curl, gsed, openssh, python24, ... |
| 05:23:42 | Fullmoon leaves the room. | |
| 05:24:37 | rue | Modularity |
| 05:26:01 | Rich_Morin | I sometimes think I should simply subscribe to continuous updates of the full sets of CPAN, gem, port, etc. After all, disk is free... |
| 05:29:49 | fbuilesv | Rich_Morin_: You obviously don't use a 1st gen Macbook as your main machine :-) |
| 05:30:38 | Rich_Morin | No, I use a PM G5 (2@2.3 GHz). |
| 05:33:25 | fbuilesv | Rich_Morin_: I was talking about the HD :P |
| 05:35:09 | Rich_Morin | Well, the reason I bought a PM was so that I could add large, economical drives, etc. So, not totally unrelated, |
| 05:38:33 | fbuilesv | :-) |
| 05:39:12 | Rich_Morin | Well, port failed me. I guess I 'll go look at the DMG, after all :-/ |
| 05:40:45 | srbaker leaves the room. | |
| 05:40:46 | rue | Yep, the Dungeon Master's Guide has invaluable advice |
| 05:41:03 | rue | had a skipped game this weekend meh |
| 05:41:03 | Rich_Morin | antares: Do you have a link for the git DMG? |
| 05:43:31 | antares | Rich_Morin_: http://metastatic.org/text/Concern/2008/03/08/git-package-1543-for-os-x/ |
| 05:43:44 | antares | Rich_Morin_: I do not know whether it is that slimmed version or not |
| 05:44:14 | Rich_Morin | It says it's a universal one. |
| 05:44:40 | Rich_Morin | so it's at least 2x a PPC one |
| 05:45:09 | antares | Rich_Morin_: I heard of that new package at #git |
| 05:47:38 | srbaker enters the room. | |
| 05:47:59 | Rich_Morin | DMGs almost always work. Then again, I've generally had good luck with port. |
| 05:50:30 | Rich_Morin | rdm@cerberus [~...Ruby/HAX] 34: which git |
| 05:51:11 | Rich_Morin | sez /usr/local/bin/git - most encouraging |
| 05:56:54 | tlockney leaves the room. | |
| 05:58:07 | headius enters the room. | |
| 05:58:50 | kofno leaves the room. | |
| 06:00:27 | MenTaLguY leaves the room. | |
| 06:01:21 | agile enters the room. | |
| 06:04:45 | agardiner leaves the room. | |
| 06:04:51 | Rich_Morin | antares: In answer to you question of a while back, I have two (related) ways I'd like to use Rubinius. One, which I've been discussing here, is to use it to assist in adding optional goo to Ruby for bullet-proofing production code. This might include things such as assertions, interfaces, prototypes, types, etc. Dunno. Right now, I'm mostly exploring the limits of my ignorance and seeing what I need to read up on. |
| 06:06:24 | Rich_Morin | antares: My other Ruby project, Arti, is an automated documentation tool for Rails. If I can use Rubinius to collect data on Ruby code (including rhtml, rake, etc), I won't have to parse it myself... |
| 06:07:24 | be9 enters the room. | |
| 06:07:24 | antares | Rich_Morin_: what data you want collect? |
| 06:07:43 | Rich_Morin | Is there a roadmap of things that are in the works, planned, etc? |
| 06:09:30 | Rich_Morin | Well, for example, it would be really useful to be able to find out inter-method calling relationships. If you've ever looked at Doxygen, you'll have an idea of how this might look. |
| 06:10:04 | Rich_Morin | So, I 'd like to be able to know that foo.rhtml calls the wombat() method |
| 06:11:04 | Rich_Morin | Now, because of all the dynamic razmatazz, I realize that I won't be able to do everything automagically, but I'm quite willing to accept that and supplement it by user-entered facts, rules, etc. |
| 06:14:16 | Rich_Morin | If I think I might be wanting to add documentation to the rubinius tree, should I ask for commit rights or is there a Better Way? |
| 06:14:38 | djwhitt | Rich_Morin_: submit a patch |
| 06:14:43 | Rich_Morin | works |
| 06:14:51 | djwhitt | Rich_Morin_: if the patch gets accepted you get commit rights |
| 06:14:57 | Rich_Morin | ah |
| 06:15:11 | djwhitt | patches usually go here: http://rubinius.lighthouseapp.com/projects/5089-rubinius/overview |
| 06:15:15 | ezmobius enters the room. | |
| 06:15:24 | RyanTM leaves the room. | |
| 06:16:47 | dbussink | Rich_Morin_: http://rubinius.lighthouseapp.com/projects/5089/howto-write-a-ticket |
| 06:17:00 | dbussink | Rich_Morin_: there is a lot more documentation there on various things |
| 06:17:30 | Rich_Morin | tnx |
| 06:19:13 | Rich_Morin | BTW, this page might be of interest - http://cfcl.com/twiki/bin/view/Projects/Spect/Ruby_Errors |
| 06:26:07 | Rich_Morin | I see that you're using spec - are folks pretty happy with it? |
| 06:28:26 | rue | We use our own spec library |
| 06:30:57 | Rich_Morin | Looks like the usage is about the same, tho. I ask because we're contemplating diving into RSpec at work... |
| 06:31:52 | dbussink | we've switched from test::unit to rspec, we like it a lot better |
| 06:32:40 | Rich_Morin | Did you go back and recode your unit tests? |
| 06:33:01 | dbussink | sort of, we did it during a refactor |
| 06:33:23 | dbussink | so in the end we basically looked at the unit test to see whether we missed anything from there when writing the specs |
| 06:33:30 | dbussink | and then threw away the unit tests |
| 06:33:48 | headius | good morning |
| 06:34:00 | dbussink | morning |
| 06:34:04 | rue | Hola |
| 06:34:10 | dbussink | headius: also an early bird today? |
| 06:34:55 | headius | I'm on EU continental time, +1 |
| 06:34:57 | Rich_Morin | A Google Maps display of IRC users might be fun... |
| 06:35:00 | headius | mini-lag |
| 06:35:28 | dbussink | a lot of it people consider 7:30 early ;) |
| 06:35:39 | headius | do you guys know if the minimal yaml specs are the only ones you're using for yaml testing? |
| 06:36:55 | dbussink | we know yaml needs a lot more specs |
| 06:38:22 | headius | you guys should just port our tests so we can toss them |
| 06:38:49 | Rich_Morin | It looks like there used to be a file named ROADMAP, but I don't see it now. Help? |
| 06:38:51 | brixen | headius: you should make a ticket for that |
| 06:38:52 | headius | jruby's set of yaml tests is probably the most complete...includes MRI's test_yaml and a few others |
| 06:39:12 | brixen | headius: people keep asking, this way we can point them to the ticket and someone will do it |
| 06:39:38 | rue | Sweet, a working Haiku image |
| 06:40:04 | brixen | rue: in vmware? |
| 06:40:50 | brixen | rue: is that the 3rd oss os you were pining for? |
| 06:43:21 | brixen | dgtized: you around? |
| 06:44:32 | lstoll leaves the room. | |
| 06:49:33 | rue | brixen: Eh, it will do |
| 06:49:44 | rue | I could not get plan9 to run so it has to stay on its own |
| 06:49:51 | brixen | great, looks like our ability to build usable stables may have been broken for quite some time |
| 06:50:09 | brixen | rue: I'll have to try it in parallels |
| 06:50:39 | brixen | dgtized: no wonder you had trouble building stables, it's broken since before your commits |
| 06:51:50 | Rich_Morin | I don't see any tests in the Rakefile - am I missing something? |
| 06:52:08 | brixen | what tests do you put in a Rakefile? |
| 06:53:01 | srbaker_ enters the room. | |
| 06:53:01 | Rich_Morin | Well, SOME build packages (eg, Perl) have all sorts of tests that they run on the binary, to ensure that it's working on the target platform. |
| 06:53:30 | srbaker leaves the room. | |
| 06:53:34 | brixen | http://rubinius.lighthouseapp.com/projects/5089/specs-overview |
| 06:53:54 | brixen | after you build, you could: rake spec |
| 06:53:59 | brixen | or bin/mspec ci |
| 06:55:19 | lstoll enters the room. | |
| 06:56:17 | brixen | hmm, where is agardiner |
| 06:59:22 | Rich_Morin | My build of Rubinius broke on PPC Tiger (http://rafb.net/p/AkIMI580.html) - anyone feeling helpful? |
| 07:00:26 | rue | Ah, that one |
| 07:00:27 | Rich_Morin | FWIW - MACOSX_DEPLOYMENT_TARGET: Undefined variable |
| 07:04:21 | rue | Looks like it is from libltdl or a system header; the rest are 10.4 or 10.5 |
| 07:04:41 | binary42 leaves the room. | |
| 07:07:38 | srbaker_ leaves the room. | |
| 07:08:00 | rue | You should be able to bypass it by giving the env explicitly |
| 07:09:34 | aotearoa leaves the room. | |
| 07:09:35 | Rich_Morin | setenv MACOSX_DEPLOYMENT_TARGET 10.4 ? |
| 07:11:04 | Rich_Morin | seems happier |
| 07:14:37 | rue | I hear an offer of free webdesign talent |
| 07:15:07 | antares | rue: did you hear it too? |
| 07:18:21 | thewoolleyman enters the room. | |
| 07:19:08 | brixen | success, I have found a commit hash from which it is possible to build stables |
| 07:22:15 | rue | brixen: How far back? |
| 07:22:16 | brixen | rue: did you know, strongtalk has been ported to linux |
| 07:22:30 | brixen | if only I wasn't chasing this, I'd build it right now |
| 07:22:37 | rue | Thought it was Wine only |
| 07:22:45 | brixen | rue: a working hash is c1211a1 |
| 07:22:59 | brixen | I'm between that and a2fa01 |
| 07:23:21 | brixen | rue: http://code.google.com/p/strongtalk/wiki/Building |
| 07:24:06 | DomainSquatter enters the room. | |
| 07:24:19 | rue | U dun care about the building part so much :) |
| 07:24:22 | brixen | go figure, discontent with one little 'nil' at the end of a method has cost me hours of work |
| 07:24:29 | brixen | rue: heh, indeed |
| 07:24:36 | brixen | I want to play :) |
| 07:27:14 | dgtized | brixen: so someone else broke the stables so they couldn't be rebuilt? |
| 07:27:21 | brixen | dgtized: indeed |
| 07:27:29 | brixen | way way wayyyy back :) |
| 07:27:32 | dgtized | I was wondering about that |
| 07:27:42 | dgtized | maybe we should rebuild stables every week or so just in case |
| 07:28:10 | brixen | rebuildius ftw :) |
| 07:28:26 | brixen | rebuildius works, rubuildius checks |
| 07:30:28 | dgtized | rebuildius? |
| 07:31:29 | brixen | a specialized bot to rebuild stables :) |
| 07:31:45 | dgtized | is there a way to generate a primitive test suit automatically for each primitive id in a given stable build? |
| 07:32:12 | dgtized | basically that's the only time it has to change right? |
| 07:32:23 | dgtized | is when we mess with the numbering in the primitives? |
| 07:32:54 | brixen | any change to a primitive that affects arity or return |
| 07:33:07 | brixen | you can add primitives, but not subtract them |
| 07:33:18 | dgtized | and primitive order right? |
| 07:33:21 | brixen | and multiplication is not commutative |
| 07:33:29 | brixen | er, wrong subject :P |
| 07:33:41 | brixen | primitive order doesn't matter |
| 07:33:46 | dgtized | how far back is it broken anyway? |
| 07:33:49 | dgtized | at least thus far? |
| 07:34:03 | brixen | dc95ff1c worked |
| 07:34:04 | dgtized | I'm trying to remember when I implemented defprim |
| 07:35:06 | DomainSquatter leaves the room. | |
| 07:35:13 | brixen | I'm between that hash and 6beb50b |
| 07:35:19 | DomainSquatter enters the room. | |
| 07:35:51 | brixen | hmm, yeah, the defprim change is smack in the middle |
| 07:35:55 | brixen | I'll try it next |
| 07:36:03 | dgtized | I got rid of primitive_names.rb with defprim about 6 commits after the last one you mentioned |
| 07:36:12 | dgtized | and I definitely didn't update the stables |
| 07:36:13 | dgtized | however |
| 07:36:21 | dgtized | the numbering should have been identical |
| 07:36:46 | brixen | I'm trying dd4304f14 now |
| 07:38:10 | dgtized | k |
| 07:40:07 | antares leaves the room. | |
| 07:40:25 | dgtized | though wilson checked in two changes to eval just before that so that also has potential |
| 07:40:50 | brixen | dd4304 is broken |
| 07:40:55 | brixen | trying one before |
| 07:41:21 | rue | I think we should just blame Defiler, he does not get enough blame for stuff |
| 07:41:39 | brixen | agreed |
| 07:41:44 | rue | He thinks his not making silly mistakes insulates him from critique.. ;) |
| 07:41:55 | brixen | although, it unfortunately does not illuminate the path to fixing this |
| 07:42:06 | rue | Did you find the right one? |
| 07:42:24 | brixen | well, defprim commit is definitely broken, trying one before that |
| 07:42:27 | dgtized | it might be my change that removed primitive_names, but brixen is checking in between |
| 07:44:44 | brixen | rue: I'm pulling down this gcc-linux branch of strongtalk on my gutsy image |
| 07:45:01 | rue | I can browse it just fine on Google's dollar :) |
| 07:45:25 | brixen | heh |
| 07:45:57 | rue | I am installing sira10s and Haiku, should be some free memory when you pin it down |
| 07:46:01 | brixen | dgtized: I think you are vindicated :) |
| 07:46:08 | lstoll leaves the room. | |
| 07:46:16 | rue | I think I will invest in another GiB on both of these machines |
| 07:46:33 | dgtized | oh it's before? |
| 07:48:16 | brixen | trying 218cc7fbd |
| 07:48:30 | brixen | 2ddabc was broken |
| 07:49:55 | dgtized | man this colored git diff is awesome |
| 07:50:23 | brixen | dgtized: oh yeah! thanks for that |
| 07:50:26 | brixen | it's very cool |
| 07:51:23 | dgtized | ok, yea it's unlikely that 2ddabc was the one that actually broke it as that one didn't change the actual primitives output |
| 07:51:35 | dgtized | so I'm guessing it's one of the two on eval |
| 07:51:48 | brixen | yeah, likely |
| 07:52:41 | dgtized | so if we determine which one broke it, how exactly do we forward port the stables from there? |
| 07:53:36 | d2dchat leaves the room. | |
| 07:53:44 | rue | Step back, what was the problem right now? |
| 07:54:26 | dgtized | I made changes that are about to require a change to stables, that's the current HEAD, it's doesn't actually require a change yet, but in order to cleanup after the changes to primitives it will |
| 07:54:26 | brixen | rake clean build; rake build:stable; rake clean:rbc; rake; bin/mspec ci => boom |
| 07:54:56 | dgtized | as result we determined that we haven't been able to update stables since march 10th ish |
| 07:55:36 | VVSiz | morning, folks! :) |
| 07:55:44 | brixen | 218cc7fb builds |
| 07:55:47 | brixen | VVSiz: morning! |
| 07:56:07 | brixen | rue: would appear that Defiler will be to blame :P |
| 07:59:13 | ezmobius leaves the room. | |
| 08:02:47 | brixen | for me: 746d89d6 => broken; 218cc7fb => works |
| 08:02:56 | brixen | if anyone wants to independently verify |
| 08:03:41 | wycats leaves the room. | |
| 08:07:56 | brixen | this makes sense, since push_self_or_class sends kind_of? and that's what cpu_locate_method was choking on |
| 08:08:06 | cypher23 enters the room. | |
| 08:08:09 | dgtized | I'm going to have to go to bed soon |
| 08:08:11 | dgtized | but |
| 08:09:04 | dgtized | wait, a compiler change that just introduces a bunch of failure tags? |
| 08:09:19 | brixen | more than that |
| 08:09:38 | mediogre enters the room. | |
| 08:10:09 | tlockney enters the room. | |
| 08:12:30 | imajes enters the room. | |
| 08:12:34 | brixen | I think it may be as simple as Object#kind_of? not being defined yet |
| 08:13:26 | qwert666 enters the room. | |
| 08:13:45 | dgtized | so if we change the load order on HEAD it might fix it? |
| 08:14:06 | mediogre leaves the room. | |
| 08:14:34 | rue | Oomg stupid sira10s |
| 08:15:31 | rue | Morning, VVSiz |
| 08:15:46 | mediogre enters the room. | |
| 08:15:47 | thehcdreamer enters the room. | |
| 08:16:00 | brixen | dgtized: there's a kind_of instruction, probably want to use that instead |
| 08:16:27 | VVSiz | rue: morning to you too. I've been asleep for 8 hours, and nothing has changed here since than! :) |
| 08:16:45 | dgtized | brixen: ah, that makes sense, anyway -- I need to go to bed, I'll see if I can't assist tommorow |
| 08:16:46 | brixen | VVSiz: I am pained! |
| 08:16:56 | brixen | VVSiz: I've spent hours on this damned bug |
| 08:17:02 | VVSiz | heheheheheh |
| 08:17:11 | brixen | dgtized: sounds good, bedtime for me too |
| 08:18:07 | brixen | VVSiz: the knowledge of this bug overfloweth the channel :) |
| 08:18:38 | VVSiz | heh, I need to check the logs then. sounds like fun! :) |
| 08:20:09 | hassox enters the room. | |
| 08:21:08 | wycats enters the room. | |
| 08:22:30 | brixen | VVSiz: I might be exaggerating a bit. there's lots of comments about git hashes though :) |
| 08:22:59 | VVSiz | :) |
| 08:23:36 | benburkert leaves the room. | |
| 08:24:02 | VVSiz | I kinda like the way Mercurial deals with revision hashes, it prepends an incrementing number to them, so: 1:23423423423, 2:34234242, etc. At least, there is some ordering :) |
| 08:24:54 | brixen | how does that work in a distributed fashion? |
| 08:28:13 | VVSiz | brixen: well, it probably doesn't work in distributed fashion, but in local context it's kinda handy |
| 08:28:41 | crafterm leaves the room. | |
| 08:29:26 | VVSiz | What fricking genius in Microsoft has put into context menu for the exetrnal hard drive, into same section: |
| 08:29:40 | VVSiz | "Safely Remove" and "Format", just side by side :) |
| 08:31:02 | crossblaim enters the room. | |
| 08:31:12 | hassox_ enters the room. | |
| 08:34:21 | VVSiz | :) |
| 08:36:15 | lstoll enters the room. | |
| 08:38:42 | hassox leaves the room. | |
| 08:39:58 | rue | brixen: Reproduction? |
| 08:40:09 | rue | Or, *to reproduce, rather? |
| 08:41:22 | rue | VVSiz: Well, you could always /imagine/ that 230a329 comes after 198fac323 |
| 08:41:56 | VVSiz | yeah! |
| 08:45:29 | be9 | git is a distributed VCS and git users need to have distributed imagination. 230a329 comes after 198fac323, no problem |
| 08:50:40 | radarek enters the room. | |
| 09:00:15 | octopod enters the room. | |
| 09:01:41 | mutle enters the room. | |
| 09:07:02 | imajes leaves the room. | |
| 09:07:31 | Arjen_ enters the room. | |
| 09:11:40 | TheVoice leaves the room. | |
| 09:28:22 | rue | So wait, what was the problem? |
| 09:28:31 | rue | I think I just built stables clean |
| 09:28:31 | thewoolleyman leaves the room. | |
| 09:30:10 | brixen | I doubt that |
| 09:30:12 | brixen | :) |
| 09:30:20 | brixen | since I have not pushed the fix yet |
| 09:31:05 | brixen | if you apply this, it will work: http://pastie.org/173994 |
| 09:32:54 | rue | 'S chugging along nicely here |
| 09:33:14 | brixen | what is, compile or mspec? |
| 09:33:26 | rue | I figured I would fix it, I thought you both turned in. Only problem was that I did not know what the problem was |
| 09:33:38 | rue | Specs are about 3/4 done |
| 09:33:39 | brixen | what is your fix? |
| 09:33:52 | brixen | mine are 90% done :) |
| 09:33:59 | rue | Nothing so far as I know unless you count my magical presence |
| 09:34:58 | brixen | 'twill be magical indeed |
| 09:35:10 | brixen | I've built it 20 times today with broken stables |
| 09:35:54 | brixen | final test now |
| 09:42:08 | boyscout | 4 commits by Brian Ford |
| 09:42:09 | boyscout | * Remove obsolete io_close primitive.; 0238804 |
| 09:42:10 | boyscout | * New stables that do not use IO#io_close primitive.; 0ccd704 |
| 09:42:11 | boyscout | * Remove IO#io_close from bootstrap.; 3715d87 |
| 09:42:12 | boyscout | * Use kind_of instruction since #kind_of? is not available at all times.; 42f0b52 |
| 09:48:49 | rue | So what was the problem symptom? |
| 09:49:20 | rue | I just rebuilt stables _with_ rebuilt stables and CI went through just fine |
| 09:49:48 | brixen | http://pastie.org/174001 |
| 09:50:40 | rue | Nope, nothing like that. |
| 09:50:48 | brixen | git reset --hard 746d89d6d; rake clean build; rake build:stable; rake clean:rbc; rake; bin/mspec ci |
| 09:50:49 | rubuildius_amd64 | Brian Ford: 023880481; 1845 files, 6268 examples, 22361 expectations, 0 failures, 0 errors; http://rafb.net/p/pGG5kj46.html |
| 09:51:00 | brixen | rue: run that command and show me the output |
| 09:58:17 | BlackEdder enters the room. | |
| 09:58:24 | brixen | same result on ey slice |
| 09:58:32 | rubuildius_ppc | Brian Ford: 023880481; 1845 files, 6271 examples, 22390 expectations, 0 failures, 0 errors; http://pastie.caboo.se/paste/174004 |
| 09:58:59 | rue | b46817 was OK, think that is earlier from today |
| 10:05:35 | imajes enters the room. | |
| 10:07:01 | brixen | rue: you must have a magical machine indeed: http://pastie.org/174005 |
| 10:10:16 | radarek leaves the room. | |
| 10:10:43 | radarek enters the room. | |
| 10:11:48 | rue | Ooh, nice.. AROS is coming along! Apparently now you can *choose* which sectors it installs itself |
| 10:13:39 | radarek leaves the room. | |
| 10:14:29 | imajes leaves the room. | |
| 10:15:41 | imajes enters the room. | |
| 10:19:05 | VVSiz | headius: WHOA!!!! I tried the test on JDK 5 and JDK 1.4 with very weird results. |
| 10:20:01 | VVSiz | both of them were running MUCH, MUCH faster with value 2, with times like 600. JDK 6 on the other hand reaches times up to 2000 |
| 10:20:30 | VVSiz | duh, wrong channel :) |
| 10:20:53 | VVSiz | nothing to see here, move along... :) |
| 10:21:20 | rue | Well now, two thousand times anything is usually exciting |
| 10:22:29 | VVSiz | heheh, headius found a weird test case where JDK 6 runs much slower on multi-core CPUs than expected. |
| 10:24:26 | chris2 enters the room. | |
| 10:42:34 | scoopr | hm, shotgun doesn't compile on arm currently :/ |
| 10:45:56 | sholden enters the room. | |
| 10:48:35 | scoopr | fun stuff too |
| 10:48:43 | scoopr | ../external_libs/libltdl/.libs/libltdl.a(ltdl.o): In function `sys_dl_open': |
| 10:48:44 | scoopr | ltdl.c:(.text+0x87c): warning: Using 'dlopen' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking |
| 10:48:53 | scoopr | subtend/nmc.c:335: warning: warning: makecontext is not implemented and will always fail |
| 10:49:07 | scoopr | and then bunch of undefined references to pthread funcs |
| 10:49:30 | rudebwoy leaves the room. | |
| 10:49:42 | rudebwoy enters the room. | |
| 10:59:53 | JimMc leaves the room. | |
| 11:16:28 | imajes leaves the room. | |
| 11:18:33 | Skip enters the room. | |
| 11:29:33 | GMFlash leaves the room. | |
| 11:33:34 | jtoy leaves the room. | |
| 11:38:50 | scoopr | who was it that had done some win32 building effort on rbx? |
| 11:46:10 | BeRUE enters the room. | |
| 11:46:26 | BeRUE | Hahaa, this was just hilariously awesome |
| 11:47:01 | rue | Apparently BeOS is coming back to life |
| 11:47:16 | BeRUE leaves the room. | |
| 11:47:54 | rue | On that note, night |
| 11:51:33 | imajes enters the room. | |
| 11:52:10 | olabini enters the room. | |
| 11:59:38 | cyndis_ enters the room. | |
| 12:01:03 | cyndis leaves the room. | |
| 12:02:01 | wdperson enters the room. | |
| 12:07:38 | cyndis enters the room. | |
| 12:08:14 | cyndis_ leaves the room. | |
| 12:17:40 | govinda enters the room. | |
| 12:20:48 | govinda | Just installed rubinius from git on my ubuntu 7.10 machine. Specs seem to run fine. I can run a simple HelloWorld (rbx hello_world.rb) program, but if I type anything in irb I get an exception. Can anybody help? |
| 12:24:30 | cremes_ leaves the room. | |
| 12:29:08 | brainopia enters the room. | |
| 12:36:13 | ttmrichter_ enters the room. | |
| 12:39:55 | chris2 leaves the room. | |
| 12:47:13 | ttmrichter leaves the room. | |
| 12:55:56 | headius leaves the room. | |
| 12:58:51 | qwert666 leaves the room. | |
| 13:06:27 | Fullmoon enters the room. | |
| 13:06:50 | Fullmoon leaves the room. | |
| 13:15:41 | mutle leaves the room. | |
| 13:15:58 | mutle enters the room. | |
| 13:16:31 | headius enters the room. | |
| 13:23:58 | crossblaim_ enters the room. | |
| 13:24:01 | crossblaim leaves the room. | |
| 13:31:01 | JimMc enters the room. | |
| 13:31:34 | abc123 | Just installed rubinius from git on my ubuntu 7.10 machine. Specs seem to run fine. I can run a simple HelloWorld (rbx hello_world.rb) program, but if I type anything in irb I get an exception. Can anybody help? |
| 13:35:34 | webmat enters the room. | |
| 13:36:31 | _helo enters the room. | |
| 13:43:37 | ttmrichter_ leaves the room. | |
| 13:45:16 | ttmrichter enters the room. | |
| 13:46:47 | ttmrichter leaves the room. | |
| 13:49:41 | ttmrichter enters the room. | |
| 13:53:59 | jtoy enters the room. | |
| 13:57:03 | _helo leaves the room. | |
| 13:57:08 | imajes leaves the room. | |
| 14:14:11 | probablycorey enters the room. | |
| 14:15:20 | AndrewO enters the room. | |
| 14:18:15 | qwert666 enters the room. | |
| 14:23:11 | macournoyer enters the room. | |
| 14:26:13 | wmoxam enters the room. | |
| 14:31:18 | binary42 enters the room. | |
| 14:33:22 | brainopia leaves the room. | |
| 14:33:25 | imajes__ enters the room. | |
| 14:34:00 | imajes__ leaves the room. | |
| 14:35:25 | imajes enters the room. | |
| 14:41:15 | imajes__ enters the room. | |
| 14:41:28 | imajes__ leaves the room. | |
| 14:41:42 | imajes leaves the room. | |
| 14:42:07 | imajes enters the room. | |
| 14:44:26 | imajes leaves the room. | |
| 14:44:50 | imajes enters the room. | |
| 14:46:49 | RyanTM enters the room. | |
| 14:49:37 | imajes__ enters the room. | |
| 14:49:41 | agile leaves the room. | |
| 14:51:18 | binary42 leaves the room. | |
| 14:51:31 | imajes__ leaves the room. | |
| 14:51:52 | imajes leaves the room. | |
| 14:52:05 | imajes enters the room. | |
| 14:52:29 | sholden leaves the room. | |
| 14:52:50 | qwert666 leaves the room. | |
| 15:00:14 | d2dchat enters the room. | |
| 15:02:43 | evan | hows things! |
| 15:02:52 | imajes | hey |
| 15:02:57 | imajes | evan: ou see this bit now? |
| 15:04:28 | evan | yep |
| 15:04:48 | imajes | cool, here's the paste again http://pastie.org/174076 |
| 15:05:09 | evan | ctrlproxy had a TON of stuff to pipe to me |
| 15:08:48 | srbaker enters the room. | |
| 15:09:34 | imajes___ enters the room. | |
| 15:10:19 | imajes leaves the room. | |
| 15:10:50 | imajes | hey |
| 15:11:00 | imajes | evan: so - since rake passes, should perhaps we get a spec to prove that it fails? |
| 15:11:02 | imajes | i guess on arity checking? |
| 15:13:48 | evan | yeah |
| 15:13:52 | evan | not sure where someone broke something |
| 15:14:31 | imajes leaves the room. | |
| 15:15:04 | imajes enters the room. | |
| 15:16:20 | imajes | evan: fair enough, will see if i can't find something to spec |
| 15:16:22 | imajes | :) |
| 15:17:15 | imajes_ leaves the room. | |
| 15:17:23 | imajes_ enters the room. | |
| 15:18:20 | djwhitt | hey evan, how goes it? |
| 15:20:27 | skaar enters the room. | |
| 15:23:12 | JimMc enters the room. | |
| 15:26:37 | fbuilesv leaves the room. | |
| 15:27:12 | obvio enters the room. | |
| 15:27:25 | yipstar enters the room. | |
| 15:30:03 | GMFlash enters the room. | |
| 15:32:19 | obiejuan enters the room. | |
| 15:34:22 | imajes leaves the room. | |
| 15:39:31 | binary42 enters the room. | |
| 15:50:49 | imajes enters the room. | |
| 15:52:52 | GMFlash leaves the room. | |
| 15:54:51 | moofbong enters the room. | |
| 15:55:09 | therealadam enters the room. | |
| 15:57:16 | rby enters the room. | |
| 15:59:17 | agile enters the room. | |
| 16:03:04 | agile leaves the room. | |
| 16:04:24 | Fullmoon enters the room. | |
| 16:06:40 | agile enters the room. | |
| 16:16:44 | binary42 leaves the room. | |
| 16:17:08 | KirinDav enters the room. | |
| 16:19:56 | abc123 | Just installed rubinius from git on my ubuntu 7.10 machine. Specs seem to run fine. I can run a simple HelloWorld (rbx hello_world.rb) program, but if I type anything in irb I get an exception. Can anybody help? |
| 16:22:46 | headius | afternoon |
| 16:24:12 | crossblaim enters the room. | |
| 16:24:12 | crossblaim_ leaves the room. | |
| 16:25:04 | JimMc leaves the room. | |
| 16:25:23 | rype leaves the room. | |
| 16:34:59 | KirinDave leaves the room. | |
| 16:35:53 | KirinDave enters the room. | |
| 16:36:04 | anonuser enters the room. | |
| 16:39:19 | thewoolleyman enters the room. | |
| 16:46:21 | imajes leaves the room. | |
| 16:52:12 | tmm1 leaves the room. | |
| 16:54:10 | KirinDave leaves the room. | |
| 16:54:35 | tmm1 enters the room. | |
| 16:56:41 | Skip leaves the room. | |
| 16:57:49 | imajes enters the room. | |
| 16:57:50 | skaar leaves the room. | |
| 16:59:48 | binary42 enters the room. | |
| 17:00:45 | joshbuddy enters the room. | |
| 17:02:14 | Skip enters the room. | |
| 17:03:21 | d2dchat_ enters the room. | |
| 17:04:16 | thewoolleyman leaves the room. | |
| 17:04:28 | Defiler | abc123: What exception do you get? |
| 17:04:49 | thewoolleyman enters the room. | |
| 17:06:23 | francois enters the room. | |
| 17:07:53 | macournoyer_ enters the room. | |
| 17:08:01 | francois | Hi all, I just cloned rubinius from git://github.com/evanphx/rubinius.git and I get a build error: http://pastie.caboo.se/174171 |
| 17:08:05 | macournoyer leaves the room. | |
| 17:08:26 | francois | gcc: grammar.c: No such file or directory |
| 17:09:28 | d2dchat_ leaves the room. | |
| 17:10:38 | d2dchat leaves the room. | |
| 17:11:52 | Defiler | francois: We aren't actively using the github version |
| 17:12:09 | Defiler | francois: So it is probably out of date, unless someone has automated it |
| 17:13:18 | Defiler | git://git.rubini.us/code is the URL you awnt |
| 17:13:20 | Defiler | (want) |
| 17:15:02 | imajes leaves the room. | |
| 17:16:24 | francois | Oh cool, thanks |
| 17:16:34 | abc123 | You can see the exception here: http://www.pastie.org/174179 |
| 17:16:56 | thehcdreamer leaves the room. | |
| 17:18:47 | Fullmoon leaves the room. | |
| 17:19:12 | Defiler | Hrm. It's like the core version of 'puts' isn't getting loaded |
| 17:19:35 | Defiler | what does: STDOUT.prim_write "Hello\n" do? |
| 17:19:43 | Defiler | (in irb, for you) |
| 17:20:43 | macournoyer | francois: I think your error is cause you don't have bison installed maybe |
| 17:22:38 | abc123 | Defiler: It is the same errror. Maybe it has something to do with readline? |
| 17:23:19 | headius_ enters the room. | |
| 17:23:34 | jinjing enters the room. | |
| 17:24:24 | abc123 | Defiler: Is it possible to corner this error by running some tests/specs? |
| 17:24:43 | Defiler | abc123: Yeah. Run bin/mspec ci |
| 17:24:58 | Defiler | In the directory where you built rubinius |
| 17:25:16 | brixen | I've got too |
| 17:25:23 | abc123 | OK. Is running... |
| 17:25:25 | brixen | bin/mspec ci runs fine |
| 17:25:51 | brixen | it's something to do with my change of kind_of instruction versus send kind_of? |
| 17:25:55 | brixen | it would appear |
| 17:26:48 | binary42 leaves the room. | |
| 17:26:53 | Defiler | So this is a stable vs. latest problem? |
| 17:27:01 | Arjen_ leaves the room. | |
| 17:27:09 | abc123 | @Defiler: the result of bin/mspec ci: 1845 files, 6268 examples, 22358 expectations, 0 failures, 0 errors |
| 17:27:13 | binary42 enters the room. | |
| 17:27:42 | brixen | shotgun/rubinius -e 'puts "hello"' |
| 17:27:44 | brixen | works fine |
| 17:27:56 | brixen | Defiler: this is an irb problem on head, yes |
| 17:27:57 | headius leaves the room. | |
| 17:28:10 | Defiler | gawtchya |
| 17:28:29 | brixen | Defiler: take a look at 42f0b52cd |
| 17:28:57 | brixen | the issue was that if you rebuilt stables: http://pastie.org/174005 |
| 17:29:09 | abc123 | @Defiler: OK. shotgun/rubinius -e 'puts "hello"' works. If I put those commands into a rb file everthing works, too. |
| 17:29:10 | wycats_ enters the room. | |
| 17:29:30 | brixen | however, it would now appear that something in irb source does not have the right obj->klass perhaps |
| 17:30:05 | Defiler | Why does the kind_of instruction use a different mechanism than the kind_of? method? |
| 17:31:02 | jinjing leaves the room. | |
| 17:31:14 | jinjing enters the room. | |
| 17:32:07 | mutle leaves the room. | |
| 17:32:15 | brixen | and HISTORY.push(l) is the problem in lib/irb/input-method.rb:98 |
| 17:35:28 | brixen | where the heck is HISTORY even assigned? |
| 17:35:51 | abc123 | and in line 110 there seems to be a typo, too: def readable_atfer_eof? |
| 17:36:13 | jtoy leaves the room. | |
| 17:37:48 | brixen | Defiler: looking at def kind_of? it would appear to be the same |
| 17:38:28 | brixen | kind_of instruction uses ISA, but that's defined to be object_kind_of_p |
| 17:38:35 | brixen | which is what opt_kind_of primitive uses |
| 17:39:24 | brixen | however, if GUARD(CLASS_P(t1) || MODULE_P(t1)); is causing the prim to fail, then the Ruby.asm code is running in #kind_of? |
| 17:39:55 | brixen | since CLASS_P and MODULE_P use RTYPE, that GUARD could very well be failing |
| 17:40:16 | Defiler | How could it fail if it ends up running the same code? |
| 17:40:25 | brixen | e.g. module Rubinius will fail that guard, even though it is technically a module |
| 17:40:47 | brixen | it wouldn't end up running object_kind_of_p because of that GUARD |
| 17:41:08 | brixen | well, I take that back |
| 17:41:28 | Defiler | Aha |
| 17:41:29 | brixen | since, the primitive then runs the kind_of instruction with Rubinius.asm |
| 17:41:31 | brixen | hmm |
| 17:41:37 | brixen | wth |
| 17:42:05 | brixen | le'me change this guard |
| 17:42:25 | abc123 | is this still related to the problem I asked about? |
| 17:42:30 | brixen | I spend hours on this in 29cc22f2c1 |
| 17:42:34 | brixen | spent* |
| 17:42:39 | brixen | abc123: yep |
| 17:43:03 | wycats leaves the room. | |
| 17:43:57 | brixen | Defiler: the main problem is that we do not uniformly set obj->obj_type |
| 17:43:59 | Defiler | We're totally sure those macros have enough parens to compose like that? |
| 17:44:10 | Defiler | Why don't we uniformly set obj_type? |
| 17:44:16 | brixen | dunno |
| 17:44:28 | Defiler | Let me rephrase that.. |
| 17:44:32 | Defiler | shouldn't we fix that? |
| 17:45:18 | brixen | well, it's a shortcut, first of all |
| 17:45:26 | brixen | doesn't take into consideration inheritance |
| 17:45:40 | brixen | I'm waiting for evan to either clarify it or show us the new vm code |
| 17:46:30 | headius | hello again |
| 17:47:00 | brixen | headius: dang! I was just about to strike you from our channel count :P |
| 17:47:07 | brixen | thought you were just lurking :) |
| 17:47:55 | KirinDave enters the room. | |
| 17:48:11 | headius | no new vm code pushed yet? |
| 17:48:14 | brixen | Defiler: I've changed the GUARD in opt_kind_of to use RISA instead of RTYPE, same exception |
| 17:48:22 | brixen | headius: not yet |
| 17:48:28 | brixen | we're all on the edge of our seats |
| 17:48:37 | Defiler | wacky |
| 17:48:47 | brixen | I saw evan made an appearance, he's conveniently disappeared :) |
| 17:49:31 | brixen | Defiler: can you see where HISTORY is even assigned? |
| 17:49:59 | Defiler | readline defines that |
| 17:50:37 | Defiler | rb_define_const(mReadline, "HISTORY", history); |
| 17:50:46 | brixen | got it |
| 17:51:00 | headius | I'm also keen to see the new stuff |
| 17:51:00 | headius | I've never heard of C++ making a codebase *more* manageable, but miracles can happen |
| 17:51:10 | brixen | so, rb_define_const perhaps |
| 17:51:22 | headius | using STL will at least help eliminate bugs in low-level data structs |
| 17:51:49 | scoopr | headius, or introduce more evil ones! ;) |
| 17: |
