|
00:00
vrurg joined,
melezhik left
01:35
librasteve_ left
06:12
melezhik3 joined
06:13
melezhik3 left
|
|||
| disbot6 | <melezhik.> [Coke]: I guess you stopped agent at some point yesterday , but the test looks good, reports from agent were coming steadily and tests worked correctly. I guess we hit a case when one job stuck , so I suppose you either terminated it or it finally went through , anyways we can think about sort of timeouts on agents side for jobs which is absolutely possible from sparky perspective | 06:57 | |
|
09:30
melezhik joined
10:59
librasteve_ joined
11:31
melezhik left
15:06
moray joined
15:53
melezhik joined
|
|||
| melezhik | Ok. 2000 modules for Rakudo 2025.11 brownie report - brw.sparrowhub.io/file/brw-orch/dbm...ummary.txt | 15:54 | |
| 0 sec timing means modules were installed as dependencies so can’t measure time | |||
| Rakudo::Perl6::Format 50663 sec looks weird but probably some stuck job finally made it, I don’t know | 15:58 | ||
| Or maybe docker container slept for awhile and the started that resulted in such a long time | 16:56 | ||
| [Coke] | I'm on a laptop, so that could happen (but my laptop rebooted due to a problem earlier, didn't auto-restart any docker stuff) | 16:57 | |
|
18:08
melezhik_ joined
|
|||
| melezhik_ | yeah, no worries , imho timing does not makes a sense at all bearing distributed natures of tests on brownie, as it depends on randomness of agent state at the time a module is installed ( e.g. how many decencies of a module are already installed ?) | 18:10 | |
| [Coke] | is the job server smart enough to send out a job that wasn't returned by agent 1 to another agent? | 18:11 | |
| melezhik_ | yes, it's already so | 18:14 | |
| actually I call this mixing, 010r sends random that have not been reported from agents to across all agents in random lists | 18:16 | ||
| sends random -> sends modules | |||
| also if agent does not report it status after 6 minutes it is taken out of rotation | 18:18 | ||
| so modules are distributed only to active ( aka online ) agents | |||
| to across all agents -> to one of online agents | 18:19 | ||
| [Coke]: I guess I am ready to say that brownie is production ready, at least I hopefully fixed all important bug in it , it'd be interesting to run 2000 modules against rakudo 2025.10 and then run compare report, and see if we get any discrepancies , after all this is what brownie/blin is for ... | 18:22 | ||
|
18:31
melezhik_ left
|
|||
| [Coke] | Yup, let's give it a try for this release (I'll also run blin) | 18:35 | |
| (er, for 2025.12) - we can also re-run against a previous one for comparisons, etc. | |||
| disbot6 | <melezhik.> 👍 | 18:51 | |
|
19:03
melezhik left
19:08
patrickb left,
patrickb joined
19:25
vrurg_ joined,
linkable6 left,
notable6 left,
linkable6 joined
19:26
bloatable6 left,
benchable6 left,
tellable6 left
19:27
vrurg left
19:29
notable6 joined
22:59
moray left
|
|||