| olli | hi | 01:56 |
|---|---|---|
| rick_h | howdy | 01:57 |
| nigelb | The private team blogpost is so confusing! | 02:21 |
| rick_h | heh, all the privacy stuff tends to lean that way I think. It's a hard thing to solve/get right | 02:24 |
| nigelb | I read twice and it took me 5 minutes to understand what the blog post meant. | 02:25 |
| nigelb | I'm still not confident yet. | 02:25 |
| jml | lifeless: a day later than promised: https://code.launchpad.net/~jml/testtools/assert-raises-lambda/+merge/90574 | 13:50 |
| * jml gone | 13:58 | |
| james_w | lifeless, around by any chance? | 21:24 |
| lifeless | fsvo | 22:56 |
| james_w | hi lifeless | 23:01 |
| james_w | I'm a bit confused by oops-tools currently | 23:01 |
| james_w | I can trigger an oops at http://ec2-107-22-26-52.compute-1.amazonaws.com/pkgme/+oops and I can see it go across amqp and it's apparently inserted in to the db at http://ec2-107-20-69-109.compute-1.amazonaws.com/?oopsid=OOPS-884555d8407b0fe9fc62ce0f34094370 but it says there's no matching oopses | 23:02 |
| lifeless | select * from oops_oops where oopsid = ' | 23:03 |
| lifeless | OOPS-884555d8407b0fe9fc62ce0f34094370' | 23:03 |
| lifeless | ; | 23:03 |
| lifeless | (e.g. have a direct poke at the DB) | 23:03 |
| lifeless | james_w: what consumer are you using to pull it off of amqp ? | 23:03 |
| james_w | amqp2disk | 23:03 |
| james_w | hmm, maybe it's a setup issue actually | 23:04 |
| james_w | (1 row) | 23:06 |
| lifeless | ok, so if a literal lookup matches, in the DB, it is inserted | 23:13 |
| lifeless | and the view should always be searching on the literal string regardless of any heuristics | 23:13 |
| james_w | ah, I think it was running amqp2disk with a relative pathname | 23:18 |
| james_w | the exists() check was failing | 23:19 |
| james_w | http://ec2-107-20-69-109.compute-1.amazonaws.com/?oopsid=OOPS-c470f2678fd5ff65b48a522ae1a35816 | 23:21 |
| lifeless | yes | 23:22 |
| lifeless | nice | 23:22 |
| lifeless | shiny shiny | 23:23 |
| james_w | it seems like timeline/oops-tools need to be generalised at bit? | 23:23 |
| james_w | oops-tools assumes that everything is sql? | 23:23 |
| lifeless | well | 23:24 |
| lifeless | the UI certainly claims it does | 23:24 |
| lifeless | and it does the %s %d substitution on everything | 23:24 |
| lifeless | OTOH thats working quite well for LP for e.g. librarian downloads and so on - mapping different librarian requests to one pattern for aggregation | 23:25 |
| lifeless | if you would like to make it better, would definitely love to see that | 23:25 |
| james_w | I need to work out why the sql hooks aren't working here | 23:29 |
| james_w | http://ec2-107-20-69-109.compute-1.amazonaws.com/?oopsid=OOPS-304f3e703162fa09f30065c7d582c890 is from when they were | 23:30 |
| lifeless | it seems like you're finding this a bit tricky to debug / sort out | 23:32 |
| lifeless | if you have suggestions for making that process easier, +1 | 23:32 |
| james_w | well, it's complicated by the fact that my internet is crappy today and I'm doing this remotely | 23:37 |
| james_w | plus timeline_django is pretty damn hacky, so it's going to be fragile | 23:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!