Da
Show enters and exits. Hide enters and exits.
| 01:51:38 | boyscout | Specs for unpack BbHh. - 0e6ca2f - Brian Ford |
| 01:51:39 | boyscout | Benchmarks for unpack BbHh. - 262ef72 - Brian Ford |
| 01:51:39 | boyscout | Added unpack BbHh. - 01384cd - Brian Ford |
| 01:51:39 | boyscout | implement rb_hash in the C API - 5f9d5fa - Duncan Mac-Vicar P |
| 01:51:39 | boyscout | spec for rb_hash implementation in the C API - 5de99ad - Duncan Mac-Vicar P |
| 01:51:39 | boyscout | expose rb_hash in ruby C api - e40249a - Duncan Mac-Vicar P |
| 01:51:39 | boyscout | Fixed C-API specs to run with 1.9.x. - 11de338 - Brian Ford |
| 01:51:40 | boyscout | Fixed rb_hash specs. - fd8659b - Brian Ford |
| 01:51:40 | boyscout | Fixed rb_hash implementation. Closes #446. - 52d6d81 - Brian Ford |
| 01:51:41 | boyscout | Fixed some NORETURN C-API functions. Closes #448. - abc6282 - Brian Ford |
| 01:51:41 | boyscout | Added define for RB_NUM_COERCE_FUNCS_NEED_OPID. Closes #447. - d833d57 - Brian Ford |
| 02:02:16 | boyscout | CI: rubinius: d833d57 successful: 3522 files, 15296 examples, 43150 expectations, 0 failures, 0 errors |
| 03:23:23 | brixen | jakedouglas: did you see my comment on #411 ? |
| 03:23:39 | brixen | jakedouglas: we would really like to fix the issue but we need more info |
| 03:24:36 | jakedouglas | i did not see it. ill take a look. |
| 03:24:44 | brixen | kk |
| 03:25:41 | jakedouglas | i think the reason i used caller() is because i wasnt getting a backtrace.. |
| 03:26:03 | brixen | that would be... odd |
| 03:26:12 | jakedouglas | that top piece that shows the console |
| 03:26:16 | jakedouglas | is literally all i saw |
| 03:26:21 | brixen | you'll need to get it from gdb then |
| 03:26:40 | jakedouglas | ok. how will i do that? |
| 03:26:56 | brixen | sec.. |
| 03:28:00 | jakedouglas | yea i think it just spat out that one line error, and then left me at the console. which actually worked fine. |
| 03:28:09 | jakedouglas | i wouldnt have used caller if there was a bt |
| 03:29:53 | jakedouglas | iirc dbussink also had this issue and experienced the same thing |
| 03:29:54 | brixen | jakedouglas: try setting a breakpoint on rubinius::Exception::type_error |
| 03:30:04 | brixen | did dbussink have repro code? |
| 03:30:16 | brixen | if I had repro code, I could fix it |
| 03:30:23 | brixen | I can't guess what your app does :P |
| 03:30:33 | jakedouglas | i understand that |
| 03:30:48 | brixen | do you have time to try this in gdb now? |
| 03:31:16 | jakedouglas | i investigated the issue as far as i could and i gave you all the information i knew how to get. i think dbussink only made it as far as i did. |
| 03:31:59 | jakedouglas | uhmmmm..yea i guess let me see if thats all setup still.. |
| 03:40:43 | jakedouglas | the breakpoint did not hit |
| 03:41:27 | jakedouglas | i enabled bp on all 3 signatures of rubinius::Exception::type_error |
| 03:43:17 | jakedouglas | btw, i am using a checkout from whenever the last time i was hacking on rbx, like a month ago or something. just fyi. |
| 03:47:07 | brixen | jakedouglas: you get the TypeError but you br is not hit, yes? |
| 03:47:28 | brixen | your br* |
| 03:48:36 | brixen | jakedouglas: set a br on rubinius::TypeError::raise too |
| 03:49:06 | brixen | also, use rbx master, please |
| 03:53:14 | jakedouglas | does rvm have the ability to pull and do a partial rebuild/install of rbx yet? last time i was doing this it was all fucked up and i had to completely remove and reinstall |
| 03:58:13 | brixen | you will save yourself untold hours of agony by just cloning rbx |
| 03:58:31 | brixen | clone; ./configure; rake build; run |
| 03:58:58 | jakedouglas | yea but dont i have to set a bazillion paths? what happens when i start invoking scripts that invoke ruby again etc, where does it find/install gems, etc |
| 03:59:40 | brixen | rbx gem install |
| 04:01:29 | brixen | you can try with rvm, I don't know if the reinstall bug was fixed or not |
| 04:01:45 | brixen | since 1.0 was released and wayne knew about it, I'm guessing it may be fixed |
| 04:02:35 | jakedouglas | yea i've found that its generally just easier to just let it rebuild on rvm, than trying to figure out what script is shelling out to ruby again and how to fix it to use rbx. unfortunately some codebases have stuff like that and it makes it painful |
| 04:03:15 | brixen | yeah |
| 04:03:20 | jakedouglas | it's annoying to rebuild but at least i can go do something else instead of wasting my time tracking shit down |
| 04:37:28 | jakedouglas | brixen: ok, the breakpoint hit. what now |
| 05:06:44 | brixen | jakedouglas: here's an example http://gist.github.com/555984 |
| 05:06:57 | brixen | so, on line 18 of the gist, bt 10 |
| 05:07:09 | brixen | that will give you 10 frames of the C stack |
| 05:07:23 | brixen | find one that is ::interperter |
| 05:07:32 | brixen | like on line 28 of the gist |
| 05:07:50 | brixen | fr N where N is the frame number |
| 05:08:00 | brixen | like line 30 of the gist |
| 05:08:21 | brixen | then run the command at line 33 of the gist |
| 05:08:31 | brixen | p call_frame->print_backtrace(state) |
| 05:08:42 | jakedouglas | ah, i remember this now. |
| 05:16:42 | jakedouglas | Cannot access memory at address 0x0 |
| 05:16:54 | brixen | arg |
| 05:17:07 | brixen | I need to track that down :( |
| 05:17:12 | brixen | le sigh |
| 05:17:31 | brixen | could you gist me bt |
| 05:17:36 | jakedouglas | ok |
| 05:18:27 | jakedouglas | http://gist.github.com/555994 |
| 05:19:15 | brixen | ok, that's better than nothing :) |
| 05:19:30 | brixen | any chance you can try to extract some code? |
| 05:19:44 | jakedouglas | how so |
| 05:19:58 | brixen | well, is this code in the controller, view, model? |
| 05:20:33 | brixen | I'm assuming you are hitting a URL? start slicing it up... |
| 05:21:13 | brixen | could you rp obj ? |
| 05:21:18 | jakedouglas | hitting a URL? did you read the original ticket at all? |
| 05:21:50 | jakedouglas | this is when script/console loads |
| 05:22:04 | brixen | ah yes |
| 05:22:18 | brixen | it's a vanilla rails app? |
| 05:22:26 | brixen | if not, then some of your code is running |
| 05:22:35 | brixen | I don't know what that is |
| 05:22:59 | brixen | I'm fixing other bugs, sorry ;) |
| 05:23:09 | jakedouglas | ok. you asked me to help you, so i tried. |
| 05:24:03 | brixen | yeah, remote debugging is a PITA |
| 05:24:06 | jakedouglas | yep |
| 05:24:42 | brixen | are you still in gdb? |
| 05:25:30 | jakedouglas | no, but i could be |
| 05:26:37 | jakedouglas | im more than happy to let you guys do it on skype or similar, most of these things end up being very obvious to you guys given the right information. the problem is that it takes 25 back-and-forths with me to get the right information, but you could do it easily if the problem was in front of you |
| 05:27:19 | jakedouglas | the same issue occurs on a fresh rails 2.1.2 app. |
| 05:28:02 | brixen | ok! then that's easy |
| 05:28:24 | brixen | can you gist me exactly what you did to get it? |
| 05:29:40 | brixen | ugh gem uninstall... |
| 05:30:15 | jakedouglas | gem install rails -v=2.1.2; rails new_app; cd new_app; script/console |
| 05:30:51 | brixen | k, I installed 2.1.2 but need to clean out the rest to make sure it's isolated |
| 07:11:32 | boyscout | Specs for RBIGNUM_SIGN. - e1071b5 - Brian Ford |
| 07:11:33 | boyscout | Added RBIGNUM_SIGN. - b611152 - Brian Ford |
| 07:22:10 | boyscout | CI: rubinius: b611152 successful: 3522 files, 15298 examples, 43152 expectations, 0 failures, 0 errors |
| 19:39:06 | ezmobius | http://mozakai.blogspot.com/2010/08/emscripten-llvm-to-javascript-compiler.html rubinius in the browser someday? |
| 19:47:46 | dbussink | ezmobius: that would be quite a lot of steps :P |
| 19:59:23 | duncanmv | There was an error loading the compiler. |
| 19:59:25 | duncanmv | It appears that your compiler is out of date with the VM. |
| 19:59:26 | duncanmv | Please use 'rbx --rebuild-compiler' or 'rake [instal]' to |
| 19:59:28 | duncanmv | what is that? |
| 19:59:40 | duncanmv | I mean, how rake clean does not fix it? |
| 21:13:53 | brixen | duncanmv: context please, are you using rvm? trying to install rbx and moving stuff? editing files? ... |
| 21:25:35 | duncanmv | I am using rvm, 1.8.7 |
| 21:25:42 | duncanmv | no, I am not installing, just building |
| 21:25:56 | duncanmv | yes, I have modified files, but I guess nothing relevant |
| 21:26:12 | duncanmv | http://gist.github.com/556700 |
| 21:27:56 | brixen | if you're working on rbx, you should do it in a direct clone |
| 21:28:21 | brixen | you've modified something in the parser, compiler chain and you've not rebuilt |
| 21:28:46 | brixen | so rbx tries to load and discovers that the .rbc signature for the bytecode compiler is not the signature the vm expects |
| 21:38:24 | duncanmv | I haven touched the compiler |
| 21:38:37 | duncanmv | and I try rake clean to build from scratch |
| 21:44:37 | brixen | well, do it in a clone |
| 21:44:57 | duncanmv | what can change the signature of the bytecode? |
| 21:46:02 | duncanmv | ** Invoke nkf-utf8/nkf.c (first_time, not_needed) |
| 21:46:04 | duncanmv | ** Invoke nkf-utf8/utf8tbl.c (first_time, not_needed) |
| 21:46:05 | duncanmv | ** Invoke nkf.c (not_needed) |
| 21:46:06 | duncanmv | ** Execute build |
| 21:46:08 | duncanmv | ** Invoke extensions:openssl (first_time) |
| 21:46:09 | duncanmv | ** Execute extensions:openssl |
| 21:46:11 | duncanmv | /space/git/rubinius/bin/rbx-build extconf.rb |
| 21:46:13 | duncanmv | There was an error loading the compiler. |
| 21:46:14 | duncanmv | It appears that your compiler is out of date with the VM. |
| 21:46:15 | duncanmv | with --trace |
| 21:46:53 | brixen | rakelib/kernel.rake:105 |
| 21:47:44 | brixen | you're probably trying to build with the rbx that you have installed |
| 21:47:47 | brixen | you can't do that |
| 21:47:58 | brixen | the rbx you have installed and have changed |
| 21:48:13 | brixen | rbx must be built completely by a separate ruby instance |
| 21:48:53 | duncanmv | I haven't installed rbx at all |
| 21:49:04 | brixen | if you are using rvm, you have |
| 21:49:18 | brixen | in whatever way rvm works |
| 21:49:21 | duncanmv | but it is not accessible for the build or? |
| 21:49:39 | brixen | honestly, I don't know |
| 21:50:09 | brixen | I can tell you what I know: you must bootstrap rbx |
| 21:50:09 | duncanmv | interesting |
| 21:50:15 | duncanmv | I removed rvm 's rbx |
| 21:50:19 | duncanmv | and now it went further |
| 21:50:22 | duncanmv | that is a bug |
| 21:50:29 | duncanmv | wayneeseguin: ? |
| 21:50:35 | brixen | ie, you must use a separate, complete ruby instance to bootstrap rbx |
| 21:51:28 | brixen | rvm is a tool to use various rubies, I don't think it's a dev tool |
| 21:52:44 | duncanmv | that is the point |
| 21:52:52 | duncanmv | I use rvm ruby 1.8.7 to build rubinius |
| 21:53:04 | duncanmv | there is a rbx in rvm too, but it is not accessible until I do rvm use rbx |
| 21:53:18 | brixen | maybe? |
| 21:53:23 | brixen | in a perfect world? |
| 21:53:41 | brixen | if you dev rbx in a clone with a normally installed ruby, you will have no issue |
| 21:53:51 | brixen | because I have 1000s of hrs doing that |
| 21:54:09 | brixen | if you use rvm, you are probably making some assumptions that we don't know about |
| 21:54:17 | brixen | and that complicates things greatly |
| 21:56:31 | duncanmv | I use also rvm to test rb |
| 21:56:33 | duncanmv | rbx |
| 21:56:40 | duncanmv | I build rbx-head from my git repo |
| 21:56:44 | duncanmv | so I can test gems |
| 21:56:56 | duncanmv | I don't know how to create an environment with my current git rbx |
| 21:57:00 | duncanmv | without "installing" it |
| 21:57:37 | brixen | well, there's a couple ideas here... |
| 21:57:51 | brixen | rbx will run perfectly fine in it's build dir |
| 21:58:19 | brixen | the only problem with that is software (like rails) that expects to exec 'ruby' to run the ruby exe |
| 21:58:34 | brixen | those are all unequivocal bugs and need to be fixed |
| 21:58:54 | brixen | most gems you can run from just rbx gem install foo; rbx -S foo_exe |
| 21:58:59 | brixen | from the build dir directly |
| 21:59:31 | brixen | what would be cool with rvm (and maybe this exists), is to be able to tell rvm to use path/to/rbx/git/clone when it runs rvm install rbx |
| 22:00:08 | brixen | then you could dev in a git clone, and plop that into an rvm instance install to test as a generic "ruby" |
| 22:01:52 | duncanmv | that is what I do |
| 22:02:24 | duncanmv | duncan@linux-llxy:/space/git/rubinius> cat ~/.rvm/config/user |
| 22:02:26 | duncanmv | # Users settings file, overrides db file settings and persists across installs. |
| 22:02:27 | duncanmv | rubinius_repo_url=git://github.com/dmacvicar/rubinius.git |
| 22:02:49 | duncanmv | but I did not understand why rbx build system found the installed rbx |
| 22:02:55 | duncanmv | it is not in path |
| 22:03:02 | brixen | then the only thing you probably have to worry about is having rbx activated and trying to build with that |
| 22:03:03 | duncanmv | unless I do rvm use rbx-head |
| 22:03:18 | duncanmv | I alwyas build with rvm use 1.8.7 |
| 22:03:50 | brixen | must be a rvm bug then |
| 22:04:18 | duncanmv | actually now that I removed rbx from rvm |
| 22:04:28 | duncanmv | rubinius build system cant build the capi specs |
| 22:04:30 | duncanmv | An exception occurred during: loading /space/git/rubinius/spec/ruby/optional/capi/time_spec.rb ERROR |
| 22:04:32 | duncanmv | Errno::ENOENT: No such file or directory - /home/duncan/.rvm/rubies/rbx-head/include/ruby.h |
| 22:05:19 | duncanmv | ok, for some reason , I was on rvm rbx-head :-/ |
| 22:05:22 | duncanmv | mmmm |
| 22:05:25 | brixen | haha |
| 22:05:28 | brixen | oh man |
| 22:06:15 | duncanmv | but with 1.8.7 happens the same, I need to reconfigure from scratch |
| 22:06:18 | brixen | give someone two ways to do something, one of which will result in catastrophic system failure, and that will be the way chosen |
| 22:06:21 | duncanmv | rake clean is the most clean-clean-clean? |
| 22:06:42 | brixen | rake clean is a "build" clean |
| 22:06:49 | brixen | ie, it cleans rbx artifacts |
| 22:06:53 | duncanmv | how can I remove the cfg cache? |
| 22:07:02 | brixen | rake distclean cleans rbx and external libs artifacts |
| 22:07:12 | brixen | I don't know what the cfg cache is |
| 22:43:16 | boyscout | Specs for Float#divmod bug, issue #451 - 70a3f04 - Ivan Evtukhovich |
| 22:43:16 | boyscout | Fixes Float#divmod bug. Closes #451. - 0141e58 - Ivan Evtukhovich |
| 22:53:01 | boyscout | CI: rubinius: 0141e58 successful: 3522 files, 15298 examples, 43156 expectations, 0 failures, 0 errors |