| 2011-12-12 09:52:40 utc | phaeron | jmettraux: epiphany (webkit) just leaked 1G of memory trying to draw that process ... |
| 2011-12-12 09:55:30 utc | phaeron | and the new ruote-kit is quite fragile when handling some processes. |
| 2011-12-12 09:55:36 utc | phaeron | I'll open some bugs about that |
| 2011-12-12 10:06:39 utc | jmettraux | phaeron: what's the new ruote-kit ? |
| 2011-12-12 10:08:53 utc | phaeron | well new for me. I just update it |
| 2011-12-12 10:09:21 utc | phaeron | jmettraux: I get a lot of NoMethodError at /_ruote/processes undefined method `state' for nil:NilClass |
| 2011-12-12 10:09:32 utc | phaeron | trying to find out what breaks it |
| 2011-12-12 10:09:44 utc | jmettraux | what else did you update ? |
| 2011-12-12 10:10:12 utc | phaeron | its dependencies. I can get you a list |
| 2011-12-12 10:10:53 utc | phaeron | I am suspecting the haml gem |
| 2011-12-12 10:11:58 utc | jmettraux | waiting for a detailed issue report |
| 2011-12-12 10:12:05 utc | phaeron | ok will do later |
| 2011-12-12 10:12:06 utc | phaeron | thanks |
| 2011-12-12 10:13:42 utc | jmettraux | quick note, #state is used in ruote-kit against the root expression of a process, NoMethodError 'state' would indicate the root expression of a process is gone, it could indicate you have corrupted processes |
| 2011-12-12 10:14:15 utc | jmettraux | one thing I can do is harden ruote-kit to survive such issues |
| 2011-12-12 10:14:37 utc | jmettraux | but it would be good to identify them as such |
| 2011-12-12 10:15:35 utc | phaeron | yes I agreee, they seem to appear only when I delete or cancel a process |
| 2011-12-12 10:15:43 utc | jmettraux | is it intermittent ? |
| 2011-12-12 10:15:51 utc | jmettraux | does the issue go if you hit reload ? |
| 2011-12-12 10:15:56 utc | phaeron | no |
| 2011-12-12 10:16:16 utc | phaeron | I find some "floating" expressions left in the fs database |
| 2011-12-12 10:16:35 utc | jmettraux | no errors ? Nothing in the application log ? |
| 2011-12-12 10:18:41 utc | phaeron | there's a backtrace in ruote-kit http://pastie.org/3004441 |
| 2011-12-12 10:19:20 utc | jmettraux | that's the backtrace of the symptom, we need to know what caused those expressions to get orphaned |
| 2011-12-12 10:19:40 utc | jmettraux | question: are you seeing this as you trying to cancel a fleeting expression ? |
| 2011-12-12 10:20:31 utc | phaeron | yes I _killed_ an expression and then this appeared |
| 2011-12-12 10:21:09 utc | phaeron | the process is no longer there but there are some expressions and an error left (I can see them in the webui and on the fs ) |
| 2011-12-12 10:21:30 utc | jmettraux | what does the error say ? |
| 2011-12-12 10:21:33 utc | phaeron | it was a flanked expression if that makes a difference |
| 2011-12-12 10:21:58 utc | phaeron | it's an error relayed from a remote participant , not related to ruote |
| 2011-12-12 10:22:30 utc | jmettraux | ok |
| 2011-12-12 10:23:15 utc | phaeron | anyway I need to go somewhere |
| 2011-12-12 10:23:23 utc | phaeron | will come back with more detailed information |
| 2011-12-12 10:24:00 utc | jmettraux | I'll probably be away, an email to the mailing list or an issue would be welcome, a way to reproduce would be golden |
| 2011-12-12 10:24:30 utc | phaeron | I'll try , thanks |
| 2011-12-12 10:24:38 utc | jmettraux | the code about expression.state got added to ruote-kit in June this year, I don't know if this is new for you guys, the flanking code is recent too |
| 2011-12-12 11:49:30 utc | jmettraux | phaeron: I pushed a commit that removes the NoMethodError on root_expression.state: https://github.com/kennethkalmer/ruote-kit/commit/95d7bd34142c8bd2a8da4f543bd8ba1e39e2c3aa |
| 2011-12-12 11:49:43 utc | jmettraux | it should help you in your work isolating the underlying cause |
| 2011-12-12 11:56:26 utc | phaeron | jmettraux: thanks |
| 2011-12-12 11:56:33 utc | phaeron | I'll apply it and see what happens |
| 2011-12-12 11:56:49 utc | phaeron | maybe we can then insert some debugging code to find the real issue |
| 2011-12-12 11:58:08 utc | jmettraux | if you can find a way to reproduce the "fleeting expression" issue, that'd be great |
| 2011-12-12 11:59:21 utc | jmettraux | if you can write a test that exhibits the issue |
| 2011-12-12 12:01:22 utc | jmettraux | back in one or two hours |
| 2011-12-12 12:02:03 utc | phaeron | ok |
| 2011-12-12 12:02:05 utc | phaeron | will try |
| 2011-12-12 22:28:44 utc | phaeron | jmettraux: I had to at least see the id of the process causing the error , so we fixed the process listing page first |
| 2011-12-12 22:29:07 utc | jmettraux | hello, ok |
| 2011-12-12 22:43:25 utc | jmettraux | phaeron: if the root_expression is nil issue and it makes ruote-kit choke is fixed, does it solve all your current issues ? |
| 2011-12-12 22:44:17 utc | phaeron | yeah it's fine for now that the process listing page works. some other pages need fixing as well but it is less important. I can fix them on a need to basis and send patches. |
| 2011-12-12 22:45:18 utc | jmettraux | ok, please don't duplicate patches then, it's confusing |
| 2011-12-12 22:45:45 utc | phaeron | hmm ? but I didn't :) |
| 2011-12-12 22:46:23 utc | jmettraux | https://github.com/MeeGoIntegration/ruote-kit/commit/af36afad7124309d849d7da03984f1f3f760370a vs https://github.com/kennethkalmer/ruote-kit/commit/95d7bd34142c8bd2a8da4f543bd8ba1e39e2c3aa |
| 2011-12-12 22:46:46 utc | jmettraux | ah argh |
| 2011-12-12 22:47:03 utc | jmettraux | sorry |
| 2011-12-12 22:47:29 utc | phaeron | no problem :) |
| 2011-12-12 22:47:53 utc | jmettraux | thanks a lot ! |
| 2011-12-12 22:49:57 utc | phaeron | welcome |
| 2011-12-12 23:14:55 utc | jmettraux | phaeron: I will look at issue #13 later (in around 7 hours), and smooth any issue triggered by your flanking error handler |
| 2011-12-12 23:15:23 utc | phaeron | sure, take your time :) |