| wgrant | Just make sure lucid-lp-base isn't running while the clones are. | 00:00 |
|---|---|---|
| wgrant | Doesn't matter wth the btrfs variant, but for aufs it does. | 00:00 |
| wgrant | Also, it needs to have your SSH key in authorized_keys, as you may be able to tell. | 00:00 |
| wgrant | There's probably some huge piece I've completely forgotten, but hunting for that can happen tomorrow. | 00:03 |
| LPCIBot | Project db-devel build #676: FAILURE in 5 hr 48 min: https://lpci.wedontsleep.org/job/db-devel/676/ | 00:09 |
| lifeless | wgrant: well, I am probably going to keep 100GB or so of windows around Just In Case | 00:09 |
| lifeless | I'm assuming dmraid will handle the dell raid setup | 00:10 |
| wgrant | Bah. | 00:10 |
| wgrant | Possibly. | 00:10 |
| === jkakar_ is now known as jkakar | ||
| lifeless | mwhudson: you have a sandybridge machine right ? | 03:21 |
| mwhudson | lifeless: i do | 03:38 |
| lifeless | did you have any issues with the natty desktop cd ? | 03:50 |
| lifeless | I get the fine grey crosshatch and nothing more | 03:50 |
| mwhudson | lifeless: no, installation went absurdly smoothly | 03:53 |
| LPCIBot | Yippie, build fixed! | 04:09 |
| LPCIBot | Project devel build #846: FIXED in 5 hr 29 min: https://lpci.wedontsleep.org/job/devel/846/ | 04:09 |
| lifeless | mwhudson: interestingly, its happy with lcd, but it hated my analogue monitor. | 05:42 |
| mwhudson | lifeless: heh | 05:42 |
| lifeless | 50% of my desktop backed up | 05:43 |
| mwhudson | lifeless: i've tested mine with a projector over vga and that worked | 05:43 |
| mwhudson | (post-install) | 05:43 |
| lifeless | mwhudson: this has a non-intel graphics card in it, for additional fun | 05:43 |
| lifeless | being a desktop and all | 05:43 |
| mwhudson | ah ok | 05:44 |
| mwhudson | well i can't help you with that then :) | 05:44 |
| lifeless | :) | 05:45 |
| LPCIBot | Yippie, build fixed! | 06:00 |
| LPCIBot | Project db-devel build #677: FIXED in 5 hr 51 min: https://lpci.wedontsleep.org/job/db-devel/677/ | 06:00 |
| lifeless | jml: https://dev.launchpad.net/ArchitectureGuide/ServicesRequirements#preview | 06:26 |
| === almaisan-away is now known as al-maisan | ||
| === al-maisan is now known as almaisan-away | ||
| wgrant | bigjools: Any objections to my comments on bug #797599? | 09:06 |
| _mup_ | Bug #797599: Copying archivepermissions cross-distro is wrong. <derivation> <qa-untestable> <Launchpad itself:Fix Committed by rvb> < https://launchpad.net/bugs/797599 > | 09:07 |
| bigjools | wgrant: I don't see any comments from you | 09:07 |
| wgrant | bigjools: Packagesets are per series. | 09:09 |
| wgrant | bigjools: So after duplicating the sets we need to regrant permissions. | 09:09 |
| StevenK | Why did that just get removed? It's trivial to fix. | 09:09 |
| wgrant | launchpad_dev=# \d packageset | 09:09 |
| wgrant | [...] | 09:09 |
| wgrant | distroseries | integer | not null | 09:09 |
| wgrant | abentley: You has QA. | 09:12 |
| bigjools | wgrant: RARGH | 09:12 |
| StevenK | Haha | 09:12 |
| StevenK | That's my line. | 09:12 |
| wgrant | bbiRargh? | 09:12 |
| wgrant | Er. | 09:12 |
| wgrant | bigjools: Rargh? | 09:12 |
| bigjools | having to re-init *archive* level permissions on a new series is crack | 09:13 |
| wgrant | Well. | 09:13 |
| bigjools | the model is broken | 09:13 |
| wgrant | Yes. | 09:13 |
| wgrant | Sort of. | 09:13 |
| wgrant | Packagesets have to be per-series. | 09:13 |
| bigjools | I know | 09:13 |
| wgrant | So permissions based on them have to be. | 09:13 |
| bigjools | the FK is wrong | 09:13 |
| wgrant | How? | 09:13 |
| wgrant | Well, how would it be improved? | 09:13 |
| bigjools | NFI ATM | 09:13 |
| wgrant | Besides nuking from orbit. | 09:13 |
| wgrant | Right. | 09:13 |
| wgrant | Exactly. | 09:14 |
| bigjools | need to think about it | 09:14 |
| bigjools | but not in a plenary! | 09:14 |
| wgrant | True. | 09:14 |
| wgrant | I won't roll it back for now, but will file a critical regression bug. | 09:14 |
| bigjools | that's fine, we can fix it | 09:14 |
| bigjools | next week | 09:14 |
| lifeless | morning d'blin | 09:16 |
| wgrant | Hi lifeless. | 09:16 |
| lifeless | apparently dmraid doth not split IO across underlying devices | 09:17 |
| lifeless | this saddens me | 09:17 |
| wgrant | Yay! | 09:17 |
| wgrant | Fakeraid must die. | 09:17 |
| StevenK | wgrant: Including MD? | 09:17 |
| lifeless | wgrant: its not substantially different to mdraid other than needing to match BIOS supported layouts, *and* being able to boot sanely. | 09:18 |
| StevenK | And wgrant has the deployment report open. Who would have thunk it. | 09:18 |
| lifeless | wgrant: i like being able to boot sanely. | 09:18 |
| StevenK | lifeless: Er? GRUB2 can boot off RAID | 09:19 |
| lifeless | StevenK: last I saw it requires the device to be on a single disk, because its not actually bringing up all the layers | 09:19 |
| lifeless | its essentially running in degraded form | 09:19 |
| StevenK | lifeless: My fileserver boots off RAID1 just fine | 09:20 |
| lifeless | this means it won't work for anything other than a vanilla mirror | 09:20 |
| StevenK | But I can't check, since my home DSL is down. | 09:20 |
| lifeless | StevenK: e.g. vanilla mirror | 09:20 |
| StevenK | You're closer, so go check. :-P | 09:20 |
| lifeless | StevenK: ring your wife :) | 09:20 |
| wgrant | StevenK: MD is not fakeraid. | 09:20 |
| wgrant | It's software RAID. | 09:20 |
| lifeless | wgrant: so is fakeraid | 09:20 |
| wgrant | It is, but with an extra layer of proprietary crap. | 09:21 |
| lifeless | wgrant: its really not that proprietary. | 09:21 |
| lifeless | crappy docs though | 09:21 |
| wgrant | Stupid and pointless and non-standard, then. | 09:21 |
| lifeless | wgrant: what would make it a standard? | 09:21 |
| wgrant | Not being chipset-dependent. | 09:21 |
| StevenK | It mainly exists because versions of Windows didn't do RAID, right? | 09:21 |
| lifeless | StevenK: *booting* from RAID requires either fragile gymnastics or a BIOS (or lower) support for the raid layout | 09:22 |
| lifeless | StevenK: the ICH (and similar) raid implementations exist because controller cards that have ATAPI/SCSI on the topside and N disks on the bottom side are hugely expensive for entry-level environments, and *other* than such cards ... you had to do gymnastics | 09:23 |
| lifeless | so fakeraid does *exactly* what the early boot BIOS interrupt calls do for SCSI and IDE disks, and run just long enough to let the OS load its drivers. | 09:24 |
| lifeless | if fakeraid is stupid, so is being able to boot off of IDE and SCSI disks. | 09:25 |
| gmb | wgrant: Is it possible to set append-revisions-only on an existing branch? I can only find references to it on bzr init. | 09:30 |
| wgrant | gmb: You probably have to edit .bzr/branch/branch.conf using something like hitchhiker. | 09:31 |
| wgrant | bzr reconfigure doesn't seem to have an option. | 09:31 |
| gmb | wgrant: Ah, right, thanks. | 09:32 |
| wgrant | Just add a line with 'append_revisions_only = True' to branch.conf | 09:32 |
| maxb | You can use bzr config now to set append_revisions_only | 09:37 |
| wgrant | Ahh, true. | 09:37 |
| wgrant | Forgot about the new shiny. | 09:37 |
| StevenK | LIFELESS! | 09:43 |
| StevenK | RARGH | 09:43 |
| wgrant | StevenK: Oh? | 09:43 |
| StevenK | parellel-test #80 has been building for *13* hours | 09:44 |
| wgrant | Heh. | 09:45 |
| StevenK | hudson 25142 0.0 0.0 0 0 ? Z Jun28 0:14 [/usr/bin/python] <defunct> | 09:46 |
| StevenK | ORSUM | 09:46 |
| StevenK | wgrant: Shall I kill it with fire? | 09:47 |
| lifeless | jml: hi | 09:54 |
| lifeless | jml: I've put some prose into the servicesrequirements page | 09:54 |
| jml | lifeless: yeah, I saw, thanks. | 09:54 |
| jml | lifeless: will look soon. | 09:54 |
| lifeless | I think it captures all that we spoke about, and includes sufficient rationale | 09:54 |
| lifeless | jml: if it doesn't, please fix-or-tell-me | 09:54 |
| lifeless | ? | 09:54 |
| nigelb | StevenK: heh, weren't you about to kill all of windmill with fire? :) | 09:56 |
| nigelb | *windmill tests | 09:56 |
| LPCIBot | Project parallel-test build #80: ABORTED in 12 hr: https://lpci.wedontsleep.org/job/parallel-test/80/ | 09:57 |
| StevenK | DIE, FIEND | 09:57 |
| nigelb | lol | 09:58 |
| jml | lifeless: ok, will do. | 09:58 |
| StevenK | nigelb: parallel-test has nothing to do with Windmill. | 09:59 |
| nigelb | StevenK: ah! | 09:59 |
| nigelb | One of my merge requests has become sort of stale. Before I work on that branch, I want to "rebase" it wth master | 09:59 |
| nigelb | is there a bzr equivalent of that? | 09:59 |
| nigelb | (this is an LP merge proporsal that I need to finish working on) | 10:00 |
| StevenK | Don't rebase, just bzr merge devel in | 10:00 |
| jml | lifeless: any ideas for useful LP (or related) hacking for me this week? | 10:01 |
| wgrant | lifeless: :( resetting rabbit is hard. | 10:01 |
| lifeless | jml: just about anything is useful | 10:02 |
| StevenK | wgrant: Do you think there is any point to keeping the windmill-{db-,}devel jobs on Jenkins, or shall I delete them? | 10:02 |
| wgrant | StevenK: I'd keep them for now, just in case. It is cheap... | 10:02 |
| lifeless | jml: one thing that you and I probably have most experience around is getting the test suite to stop stomping on stdout when something goes wrong | 10:02 |
| jml | lifeless: yeah, I know. that's largely the problem :D | 10:02 |
| lifeless | jml: that messes up subunit | 10:02 |
| nigelb | er, I did "bzr merge ../devel" and now it seems to have just gotten stuck | 10:02 |
| jml | lifeless: yeah, ok, I could try with that. | 10:03 |
| StevenK | nigelb: Is devel up-to-date? | 10:03 |
| nigelb | doesn't merge give me feedback on stdout on what's going on? | 10:03 |
| nigelb | StevenK: yeah | 10:03 |
| nigelb | oh | 10:03 |
| nigelb | done :D | 10:03 |
| StevenK | Heh | 10:03 |
| stub | I always wanted to make tests fail that spat stuff out to stdout or stderr | 10:03 |
| StevenK | More patience required | 10:03 |
| stub | Not sure the most bulletproof method of doing that though | 10:03 |
| jml | stub: interestingly, the way we patch the login object practically guarantees that tests will | 10:03 |
| nigelb | Sadly, I can't apt-get install patience :) | 10:04 |
| StevenK | I'd like to see an option to bin/test that will stop spewing the librarian log out on test failure | 10:04 |
| lifeless | deprecation warnings and bzr's __del__ methods are particularly hard to get at to fix this | 10:04 |
| lifeless | StevenK: its an attachment to the test | 10:04 |
| lifeless | StevenK: something you could do that is better | 10:04 |
| lifeless | StevenK: would be to teach the layer to only attach the *per-test* component of the log | 10:04 |
| lifeless | StevenK: rather than the log-since-fixture-start. | 10:04 |
| lifeless | StevenK: e.g. truncate the log or record the position or whatever in the layer per test setup | 10:05 |
| lifeless | StevenK: if you use tribunal though,t he librarian log won't be in your face anyway | 10:05 |
| spiv | lifeless: most of bzr's __del__ methods are being deleted. | 10:05 |
| lifeless | spiv: I know :) | 10:05 |
| lifeless | spiv: (and I'm glad) | 10:05 |
| StevenK | lifeless: Bleh, more reason to learn tribunal | 10:05 |
| nigelb | wow, I updated my devel again. | 10:06 |
| nigelb | Lots of deletions \o/ | 10:06 |
| StevenK | % apt-cache search tribunal | wc -l | 10:07 |
| StevenK | 0 | 10:07 |
| nigelb | I guess that's the windmill being burned with fire. | 10:07 |
| StevenK | jml killed lots of line too | 10:07 |
| lifeless | StevenK: poolie: can get you going with tribunal | 10:07 |
| nigelb | I now realzie I'm going to fix the same bug 4 times. | 10:07 |
| _mup_ | Bug #4: Importing finished po doesn't change progressbar <lp-translations> <Launchpad itself:Fix Released by carlos> <Ubuntu:Invalid> < https://launchpad.net/bugs/4 > | 10:08 |
| lifeless | StevenK: it has testr integration, so its basically just tribunal . | 10:08 |
| StevenK | I don't have testr either ... | 10:08 |
| lifeless | StevenK: that one is packaged; shiniest version is in the testing cabal ppa, but the one in natty is eminently usable | 10:08 |
| wgrant | lifeless: It seems there is no way to list all rabbitmq queues without using rabbitmqctl, which is 100-200ms :( | 10:11 |
| wgrant | lifeless: Which makes fixture reset difficult. | 10:11 |
| lifeless | wgrant: :( | 10:13 |
| lifeless | wgrant: there are some alternatives | 10:13 |
| wgrant | Creating a new vhost is also 100-200ms. | 10:14 |
| wgrant | Seems to be erlang startup. | 10:14 |
| lifeless | wgrant: such as, use the in python api to note queue access and connect-and-flush-all-those at the end of the test | 10:14 |
| wgrant | lifeless: Yeah, possibly... | 10:15 |
| wgrant | lifeless: But that becomes really bad when we have multiple consumers. | 10:15 |
| wgrant | That we don't control. | 10:15 |
| lifeless | wgrant: for an LP test run, all consumers should be test fakes | 10:15 |
| lifeless | wgrant: which means we have a testing interface to them, and can tell them to do their own reset dance | 10:16 |
| wgrant | "should" | 10:16 |
| lifeless | wgrant: the page has been changed. | 10:16 |
| wgrant | Also, I am talking about now. | 10:16 |
| wgrant | Not in two years. | 10:16 |
| lifeless | wgrant: *must*, or it can't go live. Existing things like the librarian are excluded | 10:16 |
| wgrant | Hmm. | 10:16 |
| lifeless | I'm flexible on this, but I think that being waffly would make it harder for folk to design well | 10:17 |
| lifeless | so I've stopped waffling on this aspect. | 10:17 |
| lifeless | by flexible I mean I am willing to be wrong and discuss; however your objection seemed to be that there was no requirement so we would have trouble in the future. | 10:19 |
| StevenK | abentley: Can haz QA? And not submitting branches with [r=me] ? | 10:20 |
| wgrant | Well, I told him to verbally, but I didn't think he'd take it literally :) | 10:20 |
| lifeless | StevenK: do you mean 'please use r=<ownname> rather than r=me when self reviewing' ? | 10:21 |
| lifeless | StevenK: if so, +1 | 10:21 |
| StevenK | r=me or rs=me is so pointless it may as well not be there. | 10:22 |
| StevenK | Since loggerhead on devel won't tell us who commited it | 10:22 |
| StevenK | Which means you either need to know, or ask buildbot. And asking buildbot makes me sad. | 10:23 |
| wgrant | Or check the committer of the merge revs. | 10:23 |
| wgrant | lifeless: Hmm, FixtureResource doesn't use reset()? | 10:27 |
| wgrant | The rabbitmq management plugin does what we want, but the rabbitmq plugin build system seems to be designed primarily to make people want to die. | 10:32 |
| lifeless | ha. headdesk. ha. | 10:32 |
| StevenK | 103 files mention 'unseen' -- some would be false-positives | 10:32 |
| wgrant | Perhaps we should use carrot. | 10:34 |
| wgrant | (it has a memory-based backend, as well as a rabbit one) | 10:35 |
| wgrant | Which would allow us to do unit tests with a nice fake. | 10:36 |
| lifeless | wgrant: that only helps if we're not talking to a microservice | 10:54 |
| Riddell | simple review needed https://code.launchpad.net/~jr/launchpad/797688-packaging-branches-label | 10:59 |
| StevenK | Riddell: I'd like to see some test changes, to be honest. | 11:04 |
| benji | someone asked about this yesterday and I wan't yet aware that it existed: http://ricostacruz.com/js2coffee/ | 11:06 |
| lifeless | benji: wicked | 11:07 |
| benji | poolie: this just came up on my feeds and wondered if you'd seen it: https://www.mergebox.net/ | 11:10 |
| benji | they bill themselves as the GitHub of bzr | 11:10 |
| lifeless | benji: they announced themselves on the bzr dev list | 11:11 |
| lifeless | benji: very interesting project | 11:11 |
| benji | ahh, I should have figured they would engauge in some way | 11:12 |
| lifeless | not at all | 11:12 |
| lifeless | its good to raise interesting things | 11:12 |
| wgrant | lifeless: Sure. | 11:12 |
| LPCIBot | Project parallel-test build #81: STILL FAILING in 1 hr 11 min: https://lpci.wedontsleep.org/job/parallel-test/81/ | 11:23 |
| lifeless | usb2 is so slow! | 11:33 |
| wgrant | bigjools: It is the 'ppa' name. | 11:33 |
| wgrant | bigjools: Not 'Test' | 11:33 |
| wgrant | So it's not quite so easy. | 11:33 |
| bigjools | yeah just saw | 11:34 |
| LPCIBot | Project db-devel build #678: FAILURE in 5 hr 51 min: https://lpci.wedontsleep.org/job/db-devel/678/ | 11:51 |
| === almaisan-away is now known as al-maisan | ||
| LPCIBot | Project parallel-test build #82: STILL FAILING in 37 min: https://lpci.wedontsleep.org/job/parallel-test/82/ | 12:01 |
| LPCIBot | Launchpad Patch Queue Manager: [r=jml][no-qa] Remove syncUpdate. | 12:01 |
| poolie | benji, lifeless, has anyone actually got a login yet? | 12:07 |
| lifeless | I haven't; I can't decide if I should, or should not :) | 12:07 |
| benji | heh | 12:07 |
| nigelb | poolie: maco did get a login for it if you're looking for feedback | 12:12 |
| nigelb | (it being mergebox) | 12:12 |
| lifeless | bigjools: dsds are shaping up really quite nice | 12:25 |
| bigjools | lifeless: hellyeah! | 12:25 |
| lifeless | bigjools: I hit a small usability thing today, not sure if its a bug or not. | 12:25 |
| lifeless | bigjools: I wanted to answer, for a friend 'whats different between debian and ubuntu for virt-manager and libvirt-bin' | 12:26 |
| bigjools | that is not enough info to properly ascertain an answer | 12:26 |
| lifeless | bigjools: localpackagediffs doesn't seem to have a unfiltered view | 12:27 |
| bigjools | it's always series-dependent | 12:27 |
| bigjools | lifeless: ! | 12:27 |
| bigjools | yes it does | 12:27 |
| lifeless | like there is unignored, ignored, ignored-and-newer, and resolved | 12:27 |
| bigjools | oh that | 12:27 |
| bigjools | there's a bug about tha | 12:27 |
| bigjools | t | 12:27 |
| lifeless | kk | 12:27 |
| lifeless | I shall not file one then | 12:27 |
| lifeless | did I mention usb2 is painfully slow ? | 12:28 |
| bigjools | lifeless: thx for feedback though | 12:28 |
| lifeless | de nada | 12:28 |
| lifeless | my friend uses Debian and then whinges that stuff doesn't Just Work :) | 12:29 |
| StevenK | lifeless: USB 2 sucks, use something better. | 12:29 |
| bigjools | there's an answer to that | 12:29 |
| lifeless | StevenK: would-if-I-could | 12:29 |
| StevenK | lifeless: There is this thing, called eSATA? | 12:29 |
| lifeless | StevenK: external disk only having usb2 | 12:30 |
| lifeless | StevenK: I just backed up everything from my old desktop onto it, and it got 30MB/s sustained... restoring at the same rate :( | 12:30 |
| StevenK | lifeless: The last time I copied stuff over USB 2, I copied 800GiB at an average of 3MB/s, and it took 4 days. | 12:32 |
| lifeless | StevenK: -ouch- | 12:34 |
| lifeless | StevenK: I'm only doing 400GB and still its painfule | 12:34 |
| huwshimi | StevenK: https://launchpad.net/launchpad/+bugs?field.tag=easy+ui&field.tags_combinator=ALL | 12:36 |
| === al-maisan is now known as almaisan-away | ||
| huwshimi | StevenK: I forgot to mention, if you fix a bug then tag it with "dublin" so we can keep track of the bugs fixed here | 14:26 |
| === almaisan-away is now known as al-maisan | ||
| huwshimi | benji: I have updated this branch, just wanted to check the fix looks ok: https://code.launchpad.net/~huwshimi/launchpad/table-headings-728187/+merge/66008 | 14:34 |
| benji | huwshimi: I made positive sounds on the MP. | 14:44 |
| huwshimi | benji: Thank you! | 14:44 |
| === Ursinha_ is now known as Ursinha | ||
| Riddell_ | StevenK: are you able to point me in the direction of writing tests for my 797688-packaging-branches-label branch at some point? | 15:00 |
| === Riddell_ is now known as Riddell | ||
| StevenK | benji: https://code.launchpad.net/~stevenk/launchpad/archive-picker-value/+merge/66318 | 15:56 |
| jtv | bigjools, StevenK, wgrant: any objections to a df app restart? | 15:56 |
| StevenK | jtv: No, doit. | 15:56 |
| wgrant | jtv: None. | 15:56 |
| jtv | StevenK: who are you calling a doit? | 15:56 |
| jtv | …or whommm? | 15:57 |
| StevenK | jtv: You, duh. | 15:57 |
| jtv | oh, I see. | 15:57 |
| benji | StevenK: done | 15:58 |
| === danilo_ is now known as danilos | ||
| danilos | jtv, hey, I've got a question about the expander widget: how do I make it start out expanded? | 16:02 |
| jtv | danilos: disable JS. :) | 16:03 |
| danilos | jtv, ok, so that's a feature we are missing: if "expanded" flag is on it on load, we should probably not fold it | 16:03 |
| jtv | Sure. | 16:03 |
| danilos | jtv, it's a case I hit on a page | 16:03 |
| jtv | I mentioned this yesterday as "nice for later"; I guess "later" came early | 16:03 |
| danilos | jtv, ok, I'll assume that works and just set the "expanded" class on the content node | 16:03 |
| jtv | BTW where is that flag exactly? Is it a CSS class, a config arg..? | 16:04 |
| danilos | jtv, yeah, I was not sure if it was implemented or not, and it's needed to be able to replace the existing collapsible stuff | 16:04 |
| jtv | Ah | 16:04 |
| danilos | jtv, a CSS class state_marker.expanded | 16:04 |
| jtv | OK | 16:04 |
| Riddell | StevenK: https://code.launchpad.net/~jr/launchpad/797688-packaging-branches-label/+merge/66146 | 16:07 |
| bigjools | jtv: nup | 16:09 |
| jtv | ? | 16:10 |
| Riddell | StevenK: voila | 16:13 |
| Riddell | hmm, wait | 16:13 |
| StevenK | benji: https://code.launchpad.net/~mbp/lazr.restfulclient/789369-getattr/+merge/62743 | 16:31 |
| matsubara | anyone available to review a branch that attempts to fix bug 404279? https://code.launchpad.net/~matsubara/launchpad/404279-private-ppa-link/+merge/66335 | 16:41 |
| _mup_ | Bug #404279: apt urls on private ppa pages are incorrect (missing credentials) <bugjam2010> <easy> <lp-soyuz> <p3a> <ppa> <ui> <Launchpad itself:In Progress by matsubara> < https://launchpad.net/bugs/404279 > | 16:41 |
| poolie | i will try | 17:03 |
| bigjools | stub: addAfterCommitHook doesn't work.... | 17:10 |
| bigjools | he no callee backee | 17:10 |
| stub | You said it worked! | 17:10 |
| bigjools | I LIED | 17:10 |
| stub | commit is being called? | 17:11 |
| bigjools | yes, we just stepped in pdb | 17:11 |
| === al-maisan is now known as almaisan-away | ||
| stub | bigjools: So we know the concept is sound, or else PostgreSQL connections would never be committed. | 17:12 |
| stub | bigjools: Oh.... aftercommithook... not sure if they are used anywhere. | 17:13 |
| bigjools | indeed | 17:13 |
| stub | Maybe just a standard commit hook? | 17:13 |
| bigjools | define "standard" ? | 17:13 |
| stub | We can worry about ordering of rabbit or pg first later | 17:13 |
| bigjools | the other one I can see in the doc is beforeCommitHook() | 17:13 |
| bigjools | stub: maybe we need registerSynch | 17:14 |
| stub | Now I can't find the documentation... | 17:15 |
| bigjools | stub: http://readthedocs.org/docs/zodb/en/latest/transactions.html | 17:15 |
| wgrant | It's not a synchronizer. | 17:15 |
| stub | You can register a class with the prepare, commit, rollback hooks like a pg database connection does | 17:15 |
| wgrant | It's not a beforeCommitHook. | 17:15 |
| wgrant | It's also not an IDataManager | 17:15 |
| bigjools | stub: how do you register that? | 17:16 |
| * stub wonders if our transaction package is out of date with those docs | 17:17 | |
| bigjools | and buggy | 17:17 |
| stub | Writing it as a full blown datamanager should work | 17:17 |
| stub | And not too traumatic since most of the methods will be noops | 17:17 |
| bigjools | ok | 17:17 |
| bigjools | stub: got an example we can see? | 17:18 |
| bigjools | preferably in our code :) | 17:18 |
| stub | bigjools: Those docs have one. Otherwise ZStorm has one. Nothing in the lp tree though. | 17:19 |
| stub | The one in the docs will be simpler, and that api has been stable for over 10 years (apart from doom which we added) | 17:20 |
| stub | Hmm... but I wonder if something else is going on. Given the aftercommit api exists, I would sort of expect it to do what it says on the tin | 17:23 |
| stub | bigjools: Your hook has the right signature? | 17:25 |
| stub | foo(success, args, kwargs) | 17:25 |
| stub | kws | 17:25 |
| stub | finished here - popping down | 17:26 |
| LPCIBot | Yippie, build fixed! | 17:44 |
| LPCIBot | Project db-devel build #679: FIXED in 5 hr 52 min: https://lpci.wedontsleep.org/job/db-devel/679/ | 17:44 |
| === salgado is now known as salgado-lunch | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!