Show enters and exits. Hide enters and exits.
| 11:21:51 | danlucraft | test |
| 11:25:08 | danlucraft | just checking logging is in order |
| 11:25:18 | danlucraft | wasn't sure if the slicehost downtime had affected us, but all seems ok |
| 16:53:33 | evan | morning. |
| 16:56:52 | danlucraft | good morning! |
| 16:57:57 | rue | Hi hi |
| 17:45:33 | boyscout | Revert varint changes to .rbc format. See below. - e59317e - Brian Ford |
| 17:45:33 | boyscout | Import SDC to c-ng dir. - 1c725e1 - Brian Ford |
| 17:45:33 | boyscout | Use 64bit method to get a 64bit value. - 73d1fcd - Brian Ford |
| 17:45:33 | boyscout | Disable find_lost_souls (segfaults iterating eden). - 457062e - Brian Ford |
| 17:45:33 | boyscout | Added pseudo profiler methods for GC runs. - 5599aa0 - Brian Ford |
| 17:52:06 | boyscout | CI: Build 5599aa0 failed. http://ci.rubini.us/rubinius/builds/5599aa007ca1ed2859b4a93707c9a1772979246a |
| 17:52:26 | evan | ARG |
| 17:52:29 | evan | why is that crashing?! |
| 17:53:00 | brixen | I think we need a crash reporter |
| 17:53:16 | evan | there actually is one on linux |
| 17:53:16 | brixen | want me to see how the jvm does it and add that? |
| 17:53:18 | evan | catchsegv |
| 17:53:33 | evan | nah, I know how to do it |
| 17:53:33 | brixen | the jvm outputs a file on segfaults |
| 17:53:35 | evan | i'll add it now |
| 17:53:40 | brixen | ok |
| 17:54:34 | evan | why is exitstatus nil? |
| 17:55:23 | brixen | someone forgot the exitstatus.decloak call? |
| 17:55:39 | evan | heh |
| 17:55:59 | evan | at this point, knowing if it stopped because of a signal or segfault would be useful. |
| 17:56:06 | brixen | yeah indeed |
| 17:56:48 | brixen | I'm thinking it's related to the handle segv I showed you |
| 17:57:07 | brixen | I get it much earlier under c-ng, but most of the specs that cause handle creation come late in the run |
| 18:02:56 | ddub | segfaults are good |
| 18:03:00 | ddub | when they are hit by you two :) |
| 18:28:06 | evan | ooh |
| 18:28:15 | evan | google has a BSD licensed crash reporter |
| 18:28:21 | evan | works on linux, darwin, solaris and windows |
| 18:28:29 | evan | going to see how hard to integrate it is |
| 18:32:27 | brixen | sweet |
| 18:38:44 | evan | heh |
| 18:38:51 | evan | i'm trying to figure out how you USE breakpad |
| 18:38:53 | evan | maybe you just link with it |
| 18:39:03 | evan | and it's initialization code sets up the signal handlers... |
| 18:40:56 | brixen | you send it notes "ohai ur cute, want to protect my segvs <wink>" |
| 18:41:51 | rue | Create an ExceptionHandler, RAII? |
| 18:42:07 | evan | ? |
| 18:44:56 | rue | 'S what the Client doc said |
| 18:45:08 | evan | where? |
| 18:45:44 | rue | http://code.google.com/p/google-breakpad/w/list |
| 18:46:02 | evan | where in there? |
| 18:46:03 | evan | what page? |
| 18:46:07 | evan | i looked through those |
| 18:47:00 | evan | oh, i see |
| 18:47:05 | evan | i missed it in ClientDesign |
| 18:55:53 | rue | They completely forgot DocumentationDesign |
| 19:10:36 | evan | man |
| 19:10:50 | evan | the linking story for this library o/~ suuuuucks! o/~ |
| 20:04:13 | boyscout | alias Module#name to Module#__name__ to make amqp work - e26aa1a - Ezra Zygmuntowicz |
| 20:04:13 | boyscout | Expect SIGTERM, so we don't output the term message - 654194a - Evan Phoenix |
| 20:04:13 | boyscout | Add signal handlers for segv and quit signals - a2298ac - Evan Phoenix |
| 20:04:25 | evan | ok, that should help. |
| 20:04:50 | evan | "QUIT" signals (TERM, HUP, USR1, USR2) now output a message and then quit |
| 20:04:59 | evan | unless there is a Signal.trap handler for them |
| 20:06:54 | boyscout | CI: Build a2298ac failed. http://ci.rubini.us/rubinius/builds/a2298acea70f34b7aeeff75beaca5e8fc5f34c29 |
| 20:07:36 | ddub | hmm, it would be interesting if the build bot used gist :) |
| 20:08:35 | evan | damnit! |
| 20:08:40 | evan | why did it get truncated... |
| 20:16:10 | boyscout | Try and improve error output flushing - f13e28d - Evan Phoenix |
| 20:19:14 | ddub | halloween I listened to rob zombie/white zombie, powerman 5000, marilyn manson, etc. |
| 20:19:35 | ddub | totally wrecked my electronica music trends |
| 20:19:56 | boyscout | CI: Build f13e28d failed. http://ci.rubini.us/rubinius/builds/f13e28d9939cf2c70d01074b9512bb24c0a8d5bc |
| 20:23:02 | boyscout | CI: f13e28d success. 3003 files, 11767 examples, 35776 expectations, 0 failures, 0 errors |
| 20:23:07 | ddub | yay |
| 20:24:07 | evan | ok, lunch. |
| 20:41:16 | brixen | building C exts is so much fun |
| 20:41:27 | brixen | mushrooms and golden spades and simple flowers all around |
| 20:41:40 | brixen | grasses and trees and wind and sunshine |
| 20:41:47 | brixen | such beauty |
| 20:42:03 | brixen | it's under-appreciated surely |
| 21:00:46 | brianmario | lol |
| 22:16:49 | brixen | Symbol not found: ___gxx_personality_v0 |
| 22:16:56 | evan | oh noes. |
| 22:16:56 | brixen | my C ext has no personality |
| 22:16:57 | evan | in what? |
| 22:17:10 | brixen | trying to get melbourne loading in MRI |
| 22:17:20 | evan | did you recompile it? |
| 22:17:25 | brixen | yeah |
| 22:17:26 | evan | wait, course you did. |
| 22:17:31 | evan | thats a C++ symbol |
| 22:17:38 | evan | there shouldn't really be any C++ symbols in your extension |
| 22:18:06 | brixen | I didn't really request any |
| 22:18:07 | scoopr | disable extensions or link stdc++ ? |
| 22:18:27 | scoopr | extensions -> exceptions |
| 22:18:48 | evan | brixen: h.m |
| 22:18:55 | evan | maybe sure you cleaned out all related .o's |
| 22:19:15 | evan | oh wait, does any of the parser stuff use C++ |
| 22:19:18 | brixen | I do make clean in the ext dir before building for MRI |
| 22:19:21 | brixen | yeah |
| 22:19:26 | evan | oh |
| 22:19:27 | evan | thats it. |
| 22:19:29 | evan | umm |
| 22:19:39 | evan | yeah, i guess try adding -lstdc++ in there |
| 22:19:41 | brixen | I have extern C around it |
| 22:19:45 | tmm1 | em does a bunch of stuff to link libstdc++ |
| 22:19:46 | evan | doesn't matter |
| 22:19:50 | evan | it doesn't work like that |
| 22:20:04 | evan | extern C just controls the names that the symbols are exported as |
| 22:20:07 | tmm1 | CONFIG['LDSHARED'] = "$(CXX) -shared" |
| 22:20:07 | evan | the bodies are still C++ |
| 22:20:12 | evan | and still need the C++ runtime. |
| 22:20:35 | brixen | tmm1: do you use extconf.rb? |
| 22:20:36 | brixen | evan: ok |
| 22:20:53 | tmm1 | yea, http://github.com/eventmachine/eventmachine/blob/master/ext/extconf.rb#L40-106 |
| 22:21:09 | brixen | tmm1: sweet, thanks |
| 22:25:27 | tmm1 | np |
| 22:26:11 | brixen | tmm1: I might borrow that whole case expr |
| 22:26:14 | brixen | :) |
| 22:27:46 | tmm1 | yea i would recommend it, lot of pain went into making it work on those platforms |
| 23:27:48 | brixen | evan: ok, I've got it building and linking against the eg vm/external_libs/libbstring/libbstring.a and the other libs |
| 23:27:58 | evan | k |
| 23:27:59 | brixen | and it loads fine, so all those symbols are resolved |
| 23:28:05 | brixen | but ld is giving me a warning |
| 23:28:13 | brixen | file is not of required architecture |
| 23:28:37 | evan | :/ |
| 23:29:41 | evan | gist the output? |
| 23:31:14 | brixen | http://gist.github.com/225579 |
| 23:32:38 | brixen | I should say, it loads and runs |
| 23:32:53 | rue | What if you leave out PPC? |
| 23:33:13 | brixen | might as well build with -arch ppc for the other libs then |
| 23:33:21 | brixen | so this works on ppc |
| 23:34:00 | brixen | should just remove all these libs |
| 23:34:39 | rue | The whole thing needs PPC (or not) if you plan on linking those |
| 23:36:16 | evan | brixen: get rid of all that |
| 23:36:20 | evan | delete all arch lines |
| 23:36:31 | evan | you don't specify the arch |
| 23:36:33 | evan | thats bad form. |
| 23:36:41 | evan | and likely the source of your troubles |
| 23:37:06 | evan | brixen: no no no |
| 23:37:08 | rue | (Unless you want to create a Universal) |
| 23:37:09 | evan | don't specify multiple archs |
| 23:37:16 | evan | you're not making universal stuff |
| 23:37:47 | brixen | evan: that's just mkmf, I didn't add it |
| 23:37:53 | evan | something did |
| 23:37:54 | brixen | but I'll figure out how to remove it |
| 23:37:56 | brixen | yeah yeah |
| 23:38:08 | evan | no really |
| 23:38:09 | evan | thats bad bad |
| 23:38:13 | evan | if mkmf is adding it |
| 23:38:16 | brixen | I got it |
| 23:38:21 | brixen | mkmf is really bad |
| 23:39:00 | evan | :/ |
| 23:39:03 | evan | just trying to help. |
| 23:39:08 | brixen | appreciated |
| 23:40:28 | evan | goes back into his grumpy fort |
| 23:43:50 | evan | i thought you didn't have to do |
| 23:43:54 | evan | gem "whatever" |
| 23:43:55 | evan | anymore |
| 23:44:27 | brixen | -arch is in rbconfig's CFLAGS |
| 23:45:49 | rue | You never have to do that, do you? Still need rrubygems |
| 23:47:03 | evan | *shrug* |
| 23:47:14 | evan | brixen: oh wait. |
| 23:47:23 | evan | what ruby are you using? |
| 23:47:25 | evan | the apple one? |
| 23:47:32 | brixen | yep |
| 23:47:36 | evan | ah thats why. |
| 23:47:41 | brixen | that's the point of what I'm building here |
| 23:47:44 | evan | can you just ignore those ld warnings |
| 23:47:47 | evan | right right |
| 23:47:52 | evan | i'd gotten confused. |
| 23:47:53 | brixen | ok |
| 23:48:10 | evan | you didn't build those other libs as universal probably |
| 23:48:21 | evan | but it's tried to build a universal .bundle anyway |
| 23:48:36 | brixen | well, I can build the libs as universal I suppose |
| 23:48:42 | brixen | if I could decode the damn makefiles |
| 23:49:19 | brixen | this is why one-offs never ever pay off in the end |
| 23:49:22 | rue | It is fine either way, so long as it is consistent |
| 23:49:29 | brixen | because someone has to fix the shit someday |
| 23:50:00 | evan | one off what? |
| 23:50:09 | slava | hi guys |
| 23:50:09 | brixen | these libs |
| 23:50:12 | evan | ah. |
| 23:50:15 | brixen | hi slava |
| 23:50:24 | evan | are those ld errors or warnings? |
| 23:50:27 | evan | can you just ignoret hem? |
| 23:50:34 | brixen | yeah, they are just warnings |
| 23:50:37 | brixen | the ext loads and runs |
| 23:50:38 | evan | just ignore them. |
| 23:50:39 | evan | go on |
| 23:50:44 | evan | don't go back. |
| 23:51:03 | brixen | you can never go back :) |