| *** amccabe has quit IRC | 00:26 | |
| *** medberry is now known as med_out | 00:29 | |
| *** heckj has quit IRC | 00:36 | |
| *** jmckenty_ has joined #openstack-dev | 00:49 | |
| *** jmckenty_ has quit IRC | 00:54 | |
| *** jakedahn has quit IRC | 00:56 | |
| *** mgius has quit IRC | 00:58 | |
| *** mwhooker has quit IRC | 01:04 | |
| *** HugoKuo_ has joined #openstack-dev | 01:09 | |
| *** HugoKuo has quit IRC | 01:09 | |
| *** negronjl has joined #openstack-dev | 01:20 | |
| *** jdurgin has quit IRC | 01:27 | |
| *** dragondm has quit IRC | 01:33 | |
| *** Tushar has quit IRC | 01:46 | |
| *** jakedahn has joined #openstack-dev | 01:47 | |
| *** mfer has joined #openstack-dev | 02:33 | |
| *** mfer has quit IRC | 02:34 | |
| *** mfer has joined #openstack-dev | 02:35 | |
| *** mfer has quit IRC | 03:05 | |
| *** negronjl` has joined #openstack-dev | 03:17 | |
| *** negronjl` has quit IRC | 03:17 | |
| *** martine_ has joined #openstack-dev | 03:29 | |
| *** martine_ has quit IRC | 03:43 | |
| *** negronjl` has joined #openstack-dev | 03:46 | |
| *** dolphm_ has joined #openstack-dev | 03:48 | |
| *** negronjl has quit IRC | 03:48 | |
| *** negronjl has joined #openstack-dev | 03:48 | |
| *** dolphm_ has quit IRC | 03:51 | |
| *** negronjl has quit IRC | 03:51 | |
| *** dolphm_ has joined #openstack-dev | 03:51 | |
| *** negronjl` is now known as negronjl | 03:51 | |
| *** dolphm_ has quit IRC | 03:52 | |
| *** negronjl has quit IRC | 03:54 | |
| *** negronjl has joined #openstack-dev | 03:54 | |
| *** negronjl has quit IRC | 04:25 | |
| *** negronjl has joined #openstack-dev | 04:33 | |
| *** negronjl has joined #openstack-dev | 04:35 | |
| *** negronjl has quit IRC | 04:36 | |
| *** negronjl has joined #openstack-dev | 04:36 | |
| *** negronjl has quit IRC | 04:37 | |
| *** negronjl has joined #openstack-dev | 04:37 | |
| *** zaitcev has quit IRC | 04:42 | |
| *** mwhooker has joined #openstack-dev | 05:07 | |
| *** RobertLaptop has joined #openstack-dev | 05:18 | |
| *** zedas has quit IRC | 06:01 | |
| *** termie has quit IRC | 06:01 | |
| *** zedas has joined #openstack-dev | 06:02 | |
| *** termie has joined #openstack-dev | 06:02 | |
| *** mwhooker has quit IRC | 06:07 | |
| HugoKuo_ | http://pastebin.com/HrZsgjWd while the additional compute node fire up instance , failed with this error in r1360 | 06:44 |
|---|---|---|
| *** reidrac has joined #openstack-dev | 07:11 | |
| *** mnour has quit IRC | 07:18 | |
| *** nickon has joined #openstack-dev | 07:57 | |
| *** mnour has joined #openstack-dev | 08:09 | |
| *** chemikadze has quit IRC | 08:13 | |
| *** chemikadze has joined #openstack-dev | 08:17 | |
| *** darraghb has joined #openstack-dev | 08:58 | |
| *** mnour has quit IRC | 09:06 | |
| *** mnour has joined #openstack-dev | 09:06 | |
| *** mnour has quit IRC | 09:20 | |
| *** mnour has joined #openstack-dev | 09:20 | |
| *** mnour has quit IRC | 09:34 | |
| *** mnour has joined #openstack-dev | 09:35 | |
| *** tudamp has joined #openstack-dev | 10:13 | |
| tudamp | hi all | 10:28 |
| tudamp | i'm doing some testing with the distribute scheduler | 10:28 |
| tudamp | i saw that using libvirt get_host_stats is not implemented | 10:28 |
| tudamp | todo? | 10:28 |
| tudamp | or i can find an initial implementation in some branch? | 10:30 |
| *** rods has joined #openstack-dev | 10:41 | |
| *** yamahata has joined #openstack-dev | 10:49 | |
| *** BK_man_ has joined #openstack-dev | 11:14 | |
| *** BK_man has quit IRC | 11:15 | |
| *** BK_man_ is now known as BK_man | 11:15 | |
| *** chemikadze has quit IRC | 11:16 | |
| *** markvoelker has joined #openstack-dev | 11:18 | |
| *** mfer has joined #openstack-dev | 11:21 | |
| blamar | tr3buchet: negative, not currently working on lp815957, I originally thought I knew what was up... dprince might be though | 11:52 |
| *** bcwaldon has joined #openstack-dev | 11:53 | |
| *** chuck_ has joined #openstack-dev | 12:05 | |
| *** chuck_ has quit IRC | 12:06 | |
| *** chuck_ has joined #openstack-dev | 12:06 | |
| *** zul has quit IRC | 12:07 | |
| *** chuck_ is now known as zul | 12:07 | |
| *** zul has joined #openstack-dev | 12:07 | |
| bcwaldon | mtaylor: tarmac doesn't appear to be merging branches :( | 12:08 |
| *** jkoelker has joined #openstack-dev | 12:32 | |
| *** yamahata has quit IRC | 12:36 | |
| *** AhmedSoliman has joined #openstack-dev | 12:38 | |
| *** sandywalsh has joined #openstack-dev | 12:40 | |
| *** bsza has joined #openstack-dev | 12:50 | |
| *** nmistry has joined #openstack-dev | 12:55 | |
| *** chemikadze has joined #openstack-dev | 12:57 | |
| chemikadze | Is there any method to get compute's host name where VM is running? | 13:00 |
| chemikadze | Through API of course. | 13:01 |
| *** nmistry has quit IRC | 13:06 | |
| bcwaldon | nope, you'll have to use 'nova-manage vm list' | 13:29 |
| bcwaldon | chemikadze: ^ | 13:29 |
| *** kbringard has joined #openstack-dev | 13:35 | |
| *** amccabe has joined #openstack-dev | 13:50 | |
| openstackjenkins | Project keystone build #62: SUCCESS in 55 sec: http://jenkins.openstack.org/job/keystone/62/ | 13:52 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Added support for versioned openstack MIME types https://review.openstack.org/120 | 13:52 |
| openstackjenkins | dolph.mathews: Added support for versioned openstack MIME types | 13:52 |
| *** dragondm has joined #openstack-dev | 14:04 | |
| soren | bcwaldon: It looks stuck. I'll kick it. | 14:22 |
| bcwaldon | soren: thanks | 14:22 |
| soren | bcwaldon: It's doing *something* now :) | 14:23 |
| *** alfred_mancx has joined #openstack-dev | 14:23 | |
| *** cp16net has joined #openstack-dev | 14:26 | |
| zykes- | what's a endpoint in keystone ? | 14:28 |
| *** dprince has joined #openstack-dev | 14:33 | |
| dolphm | zykes-: "Endpoint represents a service that a tenant could consume." -keystone dev guide | 14:37 |
| *** mnour has joined #openstack-dev | 14:38 | |
| zykes- | and diff for enpoint template vs endpoint ? | 14:40 |
| tudamp | i'm doing some testing with the distribute scheduler | 14:43 |
| tudamp | i saw that using libvirt get_host_stats is not implemented | 14:43 |
| tudamp | i can find an initial implementation in some branch? | 14:43 |
| dolphm | mtaylor: I'm not seeing a specific reason as to why this failed: https://jenkins.openstack.org/job/keystone-pylint/68/ | 14:47 |
| soren | https://jenkins.openstack.org/job/keystone-pylint/68/console | 14:48 |
| soren | dolphm: It seems to somewhere test whether 0 -ne 0. Which is false. | 14:48 |
| soren | dolphm: Do you know which script it runs? | 14:48 |
| dolphm | soren: no clue | 14:49 |
| *** nickon has quit IRC | 14:51 | |
| dolphm | zykes-: all endpoints have-a template, and basically there's one template per tenant... i don't know much more beyond that | 14:51 |
| kbringard | the thing I'd be curious to see, is some documentation explaining how Keystone "terms" map to Nova terms | 14:52 |
| kbringard | for instance, tenant is kind of like a project | 14:52 |
| soren | dolphm: Ah, I was barking up the wrong tree. | 14:52 |
| kbringard | but not exactly | 14:52 |
| soren | dolphm: It's pylint itself. Hang on. | 14:52 |
| dolphm | kbringard: that's been discussed quite a bit (esp on the mailing list) but I'm not sure i've seen anything officially documented | 14:53 |
| kbringard | yea… I suppose over time the keystone terms will just become the de-facto | 14:53 |
| annegentle | kbringard: dolphm: the keystone doc talks of it generically now, but certainly it could be updated: http://docs.openstack.org/incubation/identitydevguide/content/Tenant-d1e94.html | 14:54 |
| annegentle | since it really depends on the service provider's implementation | 14:54 |
| kbringard | annegentle: right, and thanks.. that's good info | 14:54 |
| dolphm | kbringard: i'm personally under the impression that "tenant" is a horrible term, because it confuses absolutely everyone coming from nova/swift... and everyone explains "tenant" as being some type of "resource" ... so why not just call it a "resource" instead? | 14:54 |
| kbringard | dolphm: yea... | 14:55 |
| kbringard | I dunno, I generally just find that I bang my head against it until I figure out how it works then sort of create terms in my own mind | 14:55 |
| kbringard | which isn't a slam on how OpenStack does things, it's just how I learn best | 14:55 |
| soren | dolphm: Ah, I know what's going on. | 14:56 |
| dolphm | soren: explain! i'm curious | 14:56 |
| soren | dolphm: pylint only returns succesfully if the code it examined was completely clean. | 14:57 |
| dolphm | soren: which is definitely not the case | 14:57 |
| *** rnirmal has joined #openstack-dev | 14:57 | |
| soren | dolphm: Nope :) | 14:58 |
| dolphm | soren: i know pylint returns a status code that you have to interpret bitwise | 14:58 |
| soren | dolphm: I think I fixed it. It's running again now. | 14:58 |
| dolphm | cool | 14:58 |
| soren | dolphm: Exactly. It returns 31 here. | 14:58 |
| soren | 32 means usage error. | 14:58 |
| soren | So anything below that just means it found something to complain about. Which is a success in this context. | 14:58 |
| soren | So if $? < 32, I just exit 0 instead. | 14:59 |
| dolphm | soren: 31 == 0b11111 | 14:59 |
| soren | Yeah, it's all of them :) | 14:59 |
| dolphm | so basically failures across the board | 14:59 |
| soren | erk.. https://jenkins.openstack.org/job/keystone-pylint/69/console | 15:00 |
| soren | mtaylor: https://jenkins.openstack.org/job/keystone-pylint/69/console <--- This job doesn't work if you do a "build now". Is that expected? | 15:00 |
| dolphm | soren: i think it needs more internets | 15:01 |
| *** mnour has quit IRC | 15:01 | |
| *** reidrac has quit IRC | 15:04 | |
| soren | dolphm: Nah, it seems to try to access '$GERRIT_REFSPEC' | 15:05 |
| soren | Rather than whatever the variable called GERRIT_REFSPEC contains. | 15:05 |
| * soren is on a unit testing frenzy | 15:06 | |
| dolphm | soren: https://jenkins.openstack.org/job/keystone-pylint/68/parameters/? | 15:06 |
| soren | dolphm: Right. | 15:06 |
| soren | dolphm: It probably gets all of those when it's triggered by Gerrit. | 15:07 |
| dolphm | soren: ah | 15:07 |
| soren | dolphm: ...but it wasn't thig time. I just clicked "Build now". | 15:07 |
| soren | ...which became build 69. | 15:07 |
| dolphm | can we trigger it from gerrit? | 15:07 |
| soren | No clue. | 15:07 |
| mtaylor | soren: you need to re-trigger it | 15:07 |
| soren | mtaylor: I don't know how. | 15:08 |
| * soren hasn't used Gerrit at all yet. | 15:08 | |
| openstackgerrit | Verification of a change to openstack/keystone failed: Added support for versioned openstack MIME types https://review.openstack.org/120 | 15:08 |
| mtaylor | soren: if you go to build 68, and click retrigger (which I just did) it will re-do it with appropriate version info | 15:08 |
| soren | Oh! | 15:08 |
| soren | And blow away the existing data? | 15:08 |
| dolphm | mtaylor: i assume i don't have that option, right? | 15:08 |
| mtaylor | soren: nope. it does a whole new thing | 15:08 |
| soren | mtaylor: Oh, lovely. | 15:09 |
| soren | mtaylor: I though "re-trigger" was a Gerrit operation. | 15:09 |
| soren | thought, even. | 15:09 |
| mtaylor | soren: in the future, build now should also work and should build against trunk. at the moment, we have a build step which is working around a bug in the git plugin | 15:09 |
| soren | mtaylor: Hm... Does it also reuse the same build steps? | 15:09 |
| soren | Er.. | 15:10 |
| soren | My bad. | 15:10 |
| mtaylor | soren: it should use the current build steps | 15:10 |
| soren | I misspelled 32. | 15:10 |
| mtaylor | hehe | 15:10 |
| kbringard | does keystone have it's own repo? or are we not building packages for it yet? | 15:11 |
| openstackgerrit | A change was merged to openstack/keystone: Added support for versioned openstack MIME types https://review.openstack.org/120 | 15:11 |
| dolphm | kbringard: https://github.com/openstack/keystone and no | 15:12 |
| kbringard | cool, yea, I have the code checked out, just didn't know if there were packages yet, thanks! | 15:12 |
| mtaylor | soren: yay! neat. thanks! | 15:12 |
| soren | mtaylor: No problem. The same fix probably needs to be applied in a number of other places, I presume? | 15:12 |
| dolphm | mtaylor: soren: thanks! | 15:13 |
| mtaylor | soren: eventually, yeah - keystone is the only one using a jenkins pylint job to prevent merges so far though | 15:13 |
| dolphm | mtaylor: can you revise our pylint threshold when you get a chance? (just broke 800.... @ 791) | 15:13 |
| soren | mtaylor: Ok. | 15:13 |
| mtaylor | dolphm: yes! and good job! | 15:13 |
| mtaylor | dolphm: done. it's now at 791 | 15:14 |
| soren | Good grief... | 15:15 |
| * soren is looking at the API exposed by virt drivers in Nova. | 15:15 | |
| * soren is not impressed | 15:15 | |
| Daviey | Are bzr branches still being reviewed? | 15:16 |
| soren | For nova? | 15:16 |
| Daviey | yah | 15:16 |
| soren | Yes. | 15:16 |
| Daviey | Coulda fooled me :) | 15:16 |
| soren | Nova hasn't gone git yet. | 15:16 |
| soren | Daviey: *chuckle* Which mp? | 15:16 |
| Daviey | damn it.. i just checked and it has | 15:17 |
| Daviey | egg on MY face. | 15:17 |
| mtaylor | mmm. egg | 15:18 |
| kbringard | speaking of which, I want breakfast | 15:18 |
| mtaylor | right? | 15:18 |
| kbringard | totally | 15:18 |
| mtaylor | dammit Daviey | 15:18 |
| kbringard | just killed productivity for everyone in a TZ where it's morning | 15:19 |
| kbringard | I know someone else was just asking, but I have a quick question about the endpoint_templates table in the keystone.db | 15:20 |
| kbringard | from what I can tell, the only thing it seems to care about is the internal_url | 15:20 |
| kbringard | how does the public_url and admin_url come into play there? | 15:20 |
| kbringard | actually, I lied, I see how the admin_url is used | 15:23 |
| kbringard | I don't entirely get the difference between the internal_url and the public_url | 15:23 |
| *** mwhooker has joined #openstack-dev | 15:23 | |
| *** zaitcev has joined #openstack-dev | 15:37 | |
| tr3buchet | dprince: are you working on lp815957 | 15:39 |
| dprince | mtaylor: Hey man. I is tarmac working again? | 15:39 |
| mtaylor | dprince: yes. | 15:40 |
| dprince | tr3buchet: I will if we need me to. | 15:40 |
| creiht | Is nova going to move to using unittest2? | 15:40 |
| dprince | tr3buchet: I was able to resolve the issue by using --flat_injected=false | 15:40 |
| ttx | vishy: I'm working on improving privilege separation in Nova, see https://blueprints.launchpad.net/nova/+spec/improve-privsep | 15:40 |
| ttx | vishy: feel free to target to d4/Medium of something | 15:40 |
| ttx | s/of/or/ | 15:40 |
| dprince | tr3buchet: We use the openstack guest agent to inject the network right? | 15:41 |
| tr3buchet | dprince: correct | 15:42 |
| dprince | tr3buchet: In any case it appears that network injection is hosed. I don't use that feature since I use the agent w/ XenServer. | 15:42 |
| tr3buchet | you are talking about the mounting vhd type of network injection? | 15:42 |
| dprince | tr3buchet: Yes. | 15:43 |
| tr3buchet | yikes, what's broken about it? | 15:43 |
| westmaas | you guys were talking about different bugs :) | 15:43 |
| westmaas | bug 815957 | 15:43 |
| uvirtbot | Launchpad bug 815957 in nova "lazy load operation of attribute 'network' cannot proceed" [Medium,Confirmed] https://launchpad.net/bugs/815957 | 15:43 |
| westmaas | bug 817866 | 15:44 |
| uvirtbot | Launchpad bug 817866 in nova "xenserver instance build timeout with xvdd" [Undecided,Invalid] https://launchpad.net/bugs/817866 | 15:44 |
| tr3buchet | ha, well i guess that settles the issue | 15:44 |
| dprince | tr3buchet: Yeah. Sorry man. I was talking about the same bug in multiple windows and got confused. | 15:44 |
| tr3buchet | no problems | 15:44 |
| tr3buchet | easier this way since we aren't both working on it | 15:44 |
| tr3buchet | i can take a look at that one after i get this settled if you like | 15:45 |
| *** alfred_mancx has quit IRC | 15:45 | |
| dolphm | kbringard: just got back to my desk after looking everywhere for eggs... i could only find deviled eggs | 15:46 |
| kbringard | hmmmm, deviled eggs are good | 15:46 |
| openstackgerrit | Monty Taylor proposed a change to openstack/openstack-ci: Add script to fix the email problem on gerrit. https://review.openstack.org/136 | 15:46 |
| tr3buchet | they are delicious | 15:46 |
| tr3buchet | i could eat a dozen deviled eggs | 15:46 |
| tr3buchet | right now | 15:46 |
| dolphm | kbringard: yeah but i didn't know how long they had been sitting out lol | 15:46 |
| tr3buchet | actually | 15:46 |
| kbringard | haha | 15:46 |
| tr3buchet | damn these blueberries | 15:47 |
| kbringard | lol at tr3buchet | 15:47 |
| openstackgerrit | A change was merged to openstack/openstack-ci: Add script to fix the email problem on gerrit. https://review.openstack.org/136 | 15:47 |
| kbringard | dolphm: ah, the old deviled egg gamble | 15:47 |
| dprince | tr3buchet: I'd like to look at it ... but probably not today. | 15:48 |
| *** heckj has joined #openstack-dev | 15:49 | |
| dprince | mtaylor: Is this hung: https://jenkins.openstack.org/job/nova-tarmac/109707/console | 15:50 |
| mtaylor | dprince: oh for crying out loud | 15:50 |
| *** nati has joined #openstack-dev | 15:51 | |
| tr3buchet | dprince: good i'd like to play with gerrit today anyhow | 15:51 |
| mtaylor | dprince: that seems like the same problem. I'm going to cry | 15:51 |
| dprince | mtaylor: enighten me. Can I help fix this? | 15:52 |
| mtaylor | dprince: no. so - I'm verifying with the lp admins - but a few days ago, launchpad installed an haproxy in front of the bzr servers and set a connection timeout on it | 15:52 |
| mtaylor | dprince: tarmac is holding open an idle bzr connection while running the test, and bzrlib is not handling the disconnect properly ... so yesterday they increased the timeout until the real bugfix could be done | 15:53 |
| mtaylor | dprince: this is either that resurfacing, or it's something else - but I'm poking the lp admins about it already | 15:54 |
| dprince | mtaylor: Cool. Thanks man. So it really is just hung then. Good to know. | 15:54 |
| mtaylor | dprince: totally. gotta say - running gerrit on a machine we control certainly makes it nice from a "I can fix it when it breaks" perspective :) | 15:54 |
| creiht | any keystone devs around? | 15:55 |
| mtaylor | creiht: no, they're all off looking for eggs to eat after Daviey mentioned eggs earlier this morning :) | 15:56 |
| kbringard | lol | 15:56 |
| creiht | hah | 15:56 |
| kbringard | I made some scrambled eggs… cook it up with some cottage cheese…. mmmm | 15:56 |
| * mtaylor bangs head against desk | 15:56 | |
| kbringard | haha, sorry | 15:56 |
| kbringard | I'd offer you some, but I'm guessing you're nowhere near Colorado Springs | 15:57 |
| dolphm | creiht: <-- | 15:57 |
| dolphm | i'm back, i couldn't find any eggs | 15:57 |
| creiht | dolphm: howdy | 15:57 |
| dolphm | well, trustworthy eggs | 15:58 |
| dolphm | what's up? | 15:58 |
| creiht | there were bagles in the break room this morning :) | 15:58 |
| dolphm | which one? | 15:58 |
| creiht | cloud break room | 15:58 |
| creiht | I think they are gone now | 15:58 |
| dolphm | south station? | 15:58 |
| creiht | main street? | 15:58 |
| creiht | I think | 15:58 |
| dolphm | cloud is all spread out lol | 15:58 |
| creiht | hah | 15:59 |
| creiht | true | 15:59 |
| creiht | dolphm: so back to keystone :) | 15:59 |
| creiht | the unit testing seems a bit akward to me | 16:00 |
| creiht | awkward | 16:00 |
| dolphm | creiht: you speak truth | 16:01 |
| creiht | hehe | 16:01 |
| dolphm | many are not actually unit tests | 16:01 |
| dolphm | most of them depend on the server to be running | 16:01 |
| * ttx goes to check the neighbour hens to see if any eggs are left | 16:01 | |
| creiht | well and you run all the tests, but then have to scroll back to look if everything succeeded | 16:01 |
| dolphm | some unit tests aren't even being used | 16:01 |
| * creiht loves it when his neighbor brings over fresh eggs | 16:02 | |
| creiht | dolphm: oh so yeah, that is really bad :) | 16:02 |
| dolphm | test_keystone.py needs to be deleted, and replaced with a test suite | 16:02 |
| pvo | dolphm: qq, not sure if you know the answer thought… with nova integration with keystone, what client is being used to test? | 16:02 |
| pvo | thought/though | 16:02 |
| dolphm | client of what? keystone? | 16:03 |
| pvo | client of nova | 16:03 |
| creiht | and for example, test_tenants.py fails | 16:03 |
| pvo | keystone doesn't have a specific client, I assume. It integrates with the service apis. | 16:03 |
| creiht | but doesn't seem to get run in the overall test run | 16:03 |
| creiht | gotta run to lunch | 16:03 |
| dolphm | pvo: right | 16:03 |
| *** yogirackspace has joined #openstack-dev | 16:04 | |
| dolphm | creiht: these are problems i would love to understand and then make go away! | 16:04 |
| pvo | so, to test with nova, swift, etc… are the clients being modified to use keystone? if so, which clients? | 16:04 |
| dolphm | pvo: that's not something i'm directly aware of, but i believe nova is being updated to use keystone as we speak (?) | 16:05 |
| pvo | well, there are 2 pieces of that puzzle. nova will have to verify the token but the client will have to know how to talk to keystone to to the token exchange initally as well. | 16:06 |
| dolphm | pvo: that's a fair statement | 16:08 |
| Vek | my brainstorm on how to address this: http://wiki.openstack.org/ClientAuthenticationPlugin | 16:10 |
| sandywalsh | I have the nova middleware integration working with keystone now, but I'd like to have a suitable client to use and do something meaningful | 16:10 |
| sandywalsh | (based on the keystone integration docs) | 16:10 |
| sandywalsh | I can hack on novaclient ... but only if no one else is already gone down that path | 16:11 |
| *** TimR has joined #openstack-dev | 16:11 | |
| sandywalsh | dolphm ^^ | 16:12 |
| dolphm | so, what would help you? | 16:12 |
| openstackgerrit | Yogeshwar Srikrishnan proposed a change to openstack/keystone: Changes to support services. https://review.openstack.org/137 | 16:13 |
| sandywalsh | hmm, perhaps not a lot, the examples are pretty good ... just need to make sure that I'm not treading on any toes | 16:13 |
| sandywalsh | that said, I don't know a lot on the problem yet :) | 16:14 |
| sandywalsh | s/on/about/ | 16:14 |
| *** tudamp has left #openstack-dev | 16:15 | |
| *** tudamp has joined #openstack-dev | 16:15 | |
| dolphm | i feel like a nova dev could probably point you in the right direction better than i | 16:16 |
| sandywalsh | dolphm, k, I was really just wondering if someone on the keystone team had already earmarked that story (novaclient/keystone integration) ... and if I should hold off or not | 16:17 |
| Vek | I actually posted an email about this subject to the list, but only got one (not actually very helpful) response... | 16:17 |
| *** AhmedSoliman has quit IRC | 16:19 | |
| openstackgerrit | Monty Taylor proposed a change to openstack/openstack-ci: Changed the default gerrit username to match reality. https://review.openstack.org/138 | 16:22 |
| dolphm | sandywalsh: from my perspective, i'd say go for it... but that's just my perspective | 16:25 |
| sandywalsh | dolphm, :) works for me | 16:25 |
| tudamp | Vek: I think what is missing only someone who says, "I begin to do" :) | 16:33 |
| openstackgerrit | Monty Taylor proposed a change to openstack/openstack-ci: Update the update script to set email correctly. https://review.openstack.org/139 | 16:33 |
| *** mwhooker has quit IRC | 16:33 | |
| tr3buchet | anyone else seeing this in trunk: | 16:34 |
| tr3buchet | Running pep8 ... | 16:34 |
| tr3buchet | nova/api/openstack/create_instance_helper.py:95:1: W293 blank line contains whitespace | 16:34 |
| dolphm | Open question: On nova/swift/etc, are tests analyzed by pylint? | 16:35 |
| *** adambergstein has joined #openstack-dev | 16:35 | |
| Vek | dolphm, question: isn't 2 core approves sufficient to approve a keystone patch? Or am I mistaken about which of my reviewers is core? :) | 16:37 |
| mtaylor | dolphm: pylint --rcfile=pylintrc -f parseable bin/* nova | 16:38 |
| openstackgerrit | A change to openstack/openstack-ci has been rejected: Update the update script to set email correctly. https://review.openstack.org/139 | 16:38 |
| mtaylor | Vek: the only people in the keystone core team are jesse and ziad | 16:39 |
| Vek | so I *am* mistaken about which of my reviewers is core :) Thanks :) | 16:40 |
| mtaylor | Vek: If additional people should be added to that group, you guys just let me know :) | 16:40 |
| Vek | I'm sure they will; I'm not really deep enough into any of the openstack projects yet for me to call myself anything approaching "core" | 16:41 |
| mtaylor | :) | 16:42 |
| openstackjenkins | Project keystone build #63: SUCCESS in 49 sec: http://jenkins.openstack.org/job/keystone/63/ | 16:42 |
| openstackjenkins | kevin.mitchell: Determine is_admin based on 'Admin' role; remove dead project_ref code; | 16:42 |
| *** jdurgin has joined #openstack-dev | 16:48 | |
| *** tudamp_ has joined #openstack-dev | 16:51 | |
| *** tudamp has left #openstack-dev | 16:51 | |
| openstackgerrit | Monty Taylor proposed a change to openstack/openstack-ci: Update the update script to set email correctly. https://review.openstack.org/139 | 16:52 |
| *** mnour has joined #openstack-dev | 16:52 | |
| dolphm | mtaylor: yogi wants pylint to ignore everything in the keystone.test module ... not sure how i feel about that | 16:53 |
| mtaylor | dolphm: I'm also not sure how I feel about that | 16:54 |
| mtaylor | dolphm: actually, I am - I don't like it - but I'm not sure how _strongly_ I feel about that | 16:54 |
| dolphm | mtaylor: same | 16:54 |
| dolphm | mtaylor: https://review.openstack.org/#patch,sidebyside,137,1,.pylintrc | 16:54 |
| *** BK_man has quit IRC | 16:54 | |
| zykes- | what's pylint vs pep8 ? | 16:55 |
| mtaylor | dolphm: I think I'm fine with that for now perhaps? I'd like to see the test code be clean as well in the future | 16:55 |
| mtaylor | zykes-: two different linting tools - pep8 checks for compliance with the python coding style guidelines (PEP8) ... pylint checks the code for things that are common errors or things that could be confusing and cause errors | 16:56 |
| zykes- | mtaylor: like ? | 16:56 |
| zykes- | for pylint | 16:56 |
| dolphm | unused variables, referencing a class variable that hasn't been initialized properly | 16:57 |
| zykes- | ah ok | 16:57 |
| jeblair | mtaylor, dolphm, yogirackspace: what could be a problem is getting pylint to 0 with the test module commented out. and then bumping up the limit again to work on that. just somethig to keep in mind. | 17:02 |
| jk0 | anyone else noticing LP taking forever to update diffs on MPs? | 17:04 |
| dolphm | jeblair: i agree... we're going to fix any issues in there eventually anyway, why not do it along the way, and track it? | 17:05 |
| dolphm | personally, if `ignore=test` is added to our pylint rc, i'm just going to remove it locally | 17:06 |
| *** Tushar has joined #openstack-dev | 17:06 | |
| *** mattray has joined #openstack-dev | 17:06 | |
| Tushar | sleepsonthefloor: Are you there? | 17:08 |
| sleepsonthefloor | Hey Tushar | 17:08 |
| Tushar | sleepsonthefloor:I want to change the format for security group rules, every rule will always have a single group or cidr. so there is no need to use plurals and shouldn't be a list.what do you think? | 17:09 |
| sleepsonthefloor | Tushar - that makes sense | 17:10 |
| *** jakedahn has quit IRC | 17:11 | |
| Tushar | sleepsonthefloor: Ok, thanks. | 17:11 |
| jk0 | mtaylor: LP still pretty backed up? | 17:15 |
| jk0 | or jenkins rather | 17:15 |
| *** mgius has joined #openstack-dev | 17:16 | |
| mtaylor | jk0: there is some weirdness with timeouts that I'm looking in to | 17:16 |
| jk0 | I'm noticing merge diffs aren't updating either | 17:17 |
| *** darraghb has quit IRC | 17:20 | |
| mtaylor | GREAT | 17:21 |
| openstackgerrit | A change was merged to openstack/openstack-ci: Update the update script to set email correctly. https://review.openstack.org/139 | 17:22 |
| openstackgerrit | A change was merged to openstack/openstack-ci: Changed the default gerrit username to match reality. https://review.openstack.org/138 | 17:22 |
| *** dprince has quit IRC | 17:22 | |
| * Vek slowly breaks Gerrit | 17:25 | |
| *** heckj has quit IRC | 17:25 | |
| Vek | actually, more accurately, I'm slowly breaking jenkins, or so it would seem. | 17:26 |
| Vek | mtaylor: see https://review.openstack.org/#change,131 , particularly the pylint build failure and my last comment on the patch. | 17:26 |
| * Vek makes the assumption that mtaylor is a person to point this out to... | 17:27 | |
| mtaylor | Vek: ah yes! we should install depends on your build host | 17:28 |
| mtaylor | Vek: can you tell me which things need installed? | 17:28 |
| Vek | definitely nova and glance and their respective dependencies. | 17:35 |
| Vek | I'm not familiar with swift, but swift_auth presumably imports something. | 17:35 |
| openstackgerrit | Monty Taylor proposed a change to openstack/openstack-ci-puppet: Silence unneeded output. https://review.openstack.org/140 | 17:35 |
| Vek | it does. We should also make sure to install webob | 17:35 |
| Vek | that's what I know of off the top of my head... | 17:36 |
| *** alfred_mancx has joined #openstack-dev | 17:37 | |
| mtaylor | hrm. ok. this may take a few minutes to sort | 17:38 |
| Vek | figured as much :) | 17:38 |
| mtaylor | :) | 17:38 |
| mtaylor | jeblair: ^^^ | 17:38 |
| mtaylor | jeblair: apparently keystone needs a bunch of modules installed on their slave :) | 17:39 |
| jeblair | mtaylor: okay, looking into it | 17:39 |
| *** mnour has quit IRC | 17:43 | |
| dolphm | mtaylor: jeblair: what else? | 17:44 |
| dolphm | (other modules) | 17:44 |
| jeblair | dolphm, mtaylor: "unable to import" pylint errors: http://paste.openstack.org/show/2032/ | 17:48 |
| dolphm | jeblair: none of those are installed? | 17:49 |
| mtaylor | what's dtest? | 17:49 |
| dolphm | mtaylor: a test framework that we have a sample test for, but which we're not using | 17:49 |
| mtaylor | dolphm: no. we don't have packages yet, so we didn't have a build-dep list for you | 17:49 |
| Vek | that's the test framework I wrote for integrated testing... | 17:49 |
| Vek | someone migrated the tests away, then, because I contributed several tests using dtest | 17:50 |
| jeblair | i'll update our puppet config for that host and see if i can get these installed | 17:51 |
| *** jakedahn has joined #openstack-dev | 17:51 | |
| mtaylor | so, the overall policy is that we require any deps to be installable via apt | 17:51 |
| mtaylor | in the case of dtest - I think we should decide: if we're keeping it, we need to package it, if we're not, then we should make sure it doesn't wind up being a dep | 17:52 |
| zykes- | why is "Tokens" stored in a alternate db then the rest of the stuff in KeyStone ? | 17:52 |
| Vek | dtest is on pypi, but I'm not aware of anyone having packaged it. | 17:52 |
| kbringard | zykes-: that is a good question | 17:52 |
| kbringard | that I was wondering myself | 17:52 |
| Vek | zykes: keystone may not be able to modify the user store. | 17:52 |
| dolphm | zykes-: great question... i can sort of answer | 17:52 |
| Vek | (that's my guess) | 17:53 |
| zykes- | why shouldn't it be able to modify it ? | 17:53 |
| mtaylor | Vek: well, packaging isn't a problem - I've actually got to package another thing for glance today... it's just a matter of deciding if that's what needs to be done (I'm just saying that because there seems to be differing opinions about whether or not it's being used) | 17:53 |
| zykes- | like in what cases is that ? | 17:53 |
| dolphm | zykes-: kbringard: yogi wanted to "experiment" with storing a single model in a different database, and ended up committing his experiment | 17:53 |
| kbringard | lol | 17:53 |
| zykes- | dolphm: not cleaned up afterwards ? | 17:53 |
| kbringard | so it'll eventually go to a single DB model? | 17:54 |
| *** cp16net_ has joined #openstack-dev | 17:54 | |
| dolphm | zykes-: kbringard: i have an open review which removes the "alterdb" stuff from keystone.conf, and all introduces three alternative backend configurations, which are cycled through during testing | 17:54 |
| kbringard | I was thinking it maybe had something to do with keeping the "static" admin tokens logically separate from everything else | 17:54 |
| Vek | well, what I'm thinking is, what if you get your authentication information out of, say, an LDAP store, but don't want the keystone server to be able to modify it | 17:54 |
| *** alfred_mancx has quit IRC | 17:55 | |
| Vek | in that case, it might be easier to keep the tokens in a different store. | 17:55 |
| dolphm | kbringard: nothing like that | 17:55 |
| kbringard | yea | 17:55 |
| dolphm | kbringard: or rather, nothing that ambitious :) | 17:55 |
| kbringard | that was purely a guess | 17:55 |
| kbringard | hehe | 17:55 |
| zykes- | cycles how dolphm ? | 17:55 |
| dolphm | zykes-: https://review.openstack.org/#change,132 | 17:56 |
| *** cp16net has quit IRC | 17:56 | |
| dolphm | zykes-: my next goal is to apply automatic test discovery to all of keystone.test, which means all tests will execute in a single "row" like a test suite, which means we can delete "keystone/test/unit/test_keystone.py"... the catch is that we have a lot of unreferenced failing tests | 17:57 |
| *** alfred_mancx has joined #openstack-dev | 17:58 | |
| *** rnirmal has quit IRC | 18:00 | |
| *** tudamp has joined #openstack-dev | 18:01 | |
| *** rnirmal_ has joined #openstack-dev | 18:01 | |
| *** mwhooker has joined #openstack-dev | 18:02 | |
| jeblair | Vek, mtaylor, dolphm: everything but dtest is installed on the keystone build slave now | 18:03 |
| openstackgerrit | James E. Blair proposed a change to openstack/openstack-ci-puppet: Add keystone dependencies to keystone jenkins slave. https://review.openstack.org/141 | 18:05 |
| openstackgerrit | A change was merged to openstack/keystone: Determine is_admin based on 'Admin' role; remove dead project_ref code; pass auth_token into request context; pass user_id/project_id into request context instead of their refs https://review.openstack.org/131 | 18:06 |
| Vek | that did it for me, thanks :) | 18:07 |
| jeblair | you're welcome! | 18:07 |
| dolphm | Vek: why doesn't DTestCase extend unittest.TestCase? | 18:08 |
| Vek | dolphm: because unittest and dtest are incompatible, largely. | 18:10 |
| dolphm | how so? | 18:11 |
| Vek | dtest approaches its basic testing and test running in a completely different manner from unittest, because it has the constraint of doing dependencies and auto-threading of tests. | 18:12 |
| dolphm | can you not accomplish the same constraint by starting from unittest? | 18:12 |
| Vek | I looked at extending unittest and/or nose, and decided it would be easiest to start from scratch | 18:13 |
| Vek | largely because of the need for thread safety. | 18:13 |
| zykes- | dolphm: difference on the admin api and the legacy one ? | 18:14 |
| dolphm | by legacy do you mean 'service api'? | 18:15 |
| zykes- | yes | 18:15 |
| zykes- | something rackspace stuff? | 18:15 |
| dolphm | admin is currently a superset of the service api, but is not defined as such | 18:15 |
| dolphm | service api is sort of the public api for clients to authenticate with | 18:15 |
| zykes- | but does it need the legacy auth stuff ? | 18:16 |
| dolphm | while the admin api is for, say, nova itself to manage keystone through, and validate client tokens, etc | 18:16 |
| dolphm | legacy auth is a reimplementation of "auth 1.0 / 1.1" ... rackspace-specific stuff | 18:16 |
| dolphm | 1.0 and 1.1 really don't have anything to do with 'keystone' ... 'keystone' starts at 2.0 | 18:16 |
| zykes- | remove code ? M) | 18:17 |
| openstackjenkins | Project swift build #301: SUCCESS in 30 sec: http://jenkins.openstack.org/job/swift/301/ | 18:17 |
| *** jakedahn_ has joined #openstack-dev | 18:17 | |
| openstackjenkins | Tarmac: Bug #820185: Intermittent TypeError NoneType in atexit handler of unit tests | 18:17 |
| dolphm | which code? | 18:17 |
| openstackjenkins | Work around what appears to be a Python standard library bug by silencing | 18:17 |
| uvirtbot | Launchpad bug 820185 in swift "Intermittent TypeError NoneType in atexit handler of unit tests" [Undecided,Fix committed] https://launchpad.net/bugs/820185 | 18:17 |
| openstackjenkins | exceptions in the logging library. | 18:17 |
| zykes- | "'keystone-legacy-auth', options, args) | 18:18 |
| zykes- | from bin/keystone | 18:18 |
| *** jakedahn has quit IRC | 18:19 | |
| *** jakedahn_ is now known as jakedahn | 18:19 | |
| *** mszilagyi has joined #openstack-dev | 18:19 | |
| zykes- | do you use vi dolphm ? | 18:23 |
| dolphm | zykes-: nope | 18:26 |
| zykes- | there's alot of spaces that aren't used on empty lines.. | 18:27 |
| dolphm | zykes-: just random trailing whitespace? | 18:27 |
| zykes- | no, in general, you see alot of "^........." spaces | 18:28 |
| zykes- | in NormalizingFilter under "self.conf = conf" | 18:29 |
| zykes- | there's some | 18:29 |
| dolphm | zykes-: just tested it.. my ide (pydev) is set to auto-remove that | 18:30 |
| zykes- | wonder who's doing that then ;p | 18:30 |
| dolphm | zykes-: i'm not saying it not me... i have too many dev environments | 18:32 |
| zykes- | other public projects ? | 18:32 |
| *** jakedahn has quit IRC | 18:33 | |
| dolphm | i mean, i run pydev on os x and ubuntu on this laptop, i run pydev on os x and ubuntu on another laptop, and on debian at home | 18:33 |
| zykes- | ah | 18:34 |
| dolphm | i also fall back on nano/gedit/whatever sometimes, i used komodo about a month ago | 18:34 |
| dolphm | i'm sure they're not all exactly the same | 18:34 |
| dolphm | more importantly, i don't view whitespace on any of them :) but i do find useless whitespace annoying when i stumble on it | 18:35 |
| *** KnuckleSangwich has joined #openstack-dev | 18:36 | |
| creiht | dolphm: so I'm supposed to evaluate the whole git/gerrit workflow | 18:37 |
| creiht | The first thing that stood out to me that could use some help for keystone were the unit tests | 18:38 |
| creiht | but that seems like it needs more than just a couple of quick fixes | 18:38 |
| *** jakedahn has joined #openstack-dev | 18:39 | |
| creiht | though I may still be able to find some small things that can help | 18:40 |
| *** tudamp has quit IRC | 18:40 | |
| creiht | I need to dig further, but alas I have meetings most of the rest of the day :/ | 18:40 |
| *** tudamp has joined #openstack-dev | 18:44 | |
| annegentle | mtaylor: around? | 18:45 |
| dolphm | creiht: i totally agree | 18:46 |
| dolphm | creiht: working on tests as we speak! | 18:46 |
| creiht | hehe | 18:46 |
| *** jakedahn_ has joined #openstack-dev | 18:47 | |
| *** jakedahn has quit IRC | 18:48 | |
| *** jakedahn_ is now known as jakedahn | 18:48 | |
| openstackgerrit | John Dickinson proposed a change to openstack/keystone: updated README with more accurate swift info https://review.openstack.org/142 | 18:49 |
| *** cp16net has joined #openstack-dev | 18:49 | |
| *** alfred_mancx has quit IRC | 18:50 | |
| notmyname | dolphm: (or other keystone devs): so now what? I've proposed a merge to keystone, pushed it to gerrit, ....? | 18:51 |
| creiht | lol | 18:53 |
| notmyname | just want to understand the tools if I'm going to be expected to use them | 18:54 |
| dolphm | notmyname: two or more people have to review it, and if all is good, jenkins will run tests on it, check it for pylint issues, and then merge it into openstack/keystone | 18:54 |
| notmyname | dolphm: who do I talk to about configuring those things (2 reviews, pylint, etc)? | 18:55 |
| *** nati has quit IRC | 18:55 | |
| dolphm | notmyname: i can either help or point you in the right direction | 18:55 |
| notmyname | jeblair: ^? | 18:55 |
| dolphm | notmyname: the right direction usually being jeblair, mtaylor, or soren | 18:56 |
| notmyname | ok | 18:56 |
| *** nati has joined #openstack-dev | 18:56 | |
| dolphm | notmyname: for pylint, we have a '.pylintrc' in the root of the project | 18:56 |
| notmyname | jeblair: mtaylor: soren: is the idea that all projects using git+gerrit will use the same rules? is there any way to override those rules? (eg can a core dev merge in a "small" change without 2 reviews?) | 18:57 |
| jaypipes | notmyname: no. | 19:00 |
| notmyname | to which? ;-) | 19:00 |
| *** alfred_mancx has joined #openstack-dev | 19:00 | |
| jaypipes | notmyname: projects can decide what policies they wish to enforce. the only policy we have is that there is no human-gatekeeper. | 19:00 |
| notmyname | ok | 19:00 |
| jaypipes | notmyname: just email jeblair after talking with the swift contribs what your preference is. | 19:01 |
| jaypipes | notmyname: you can have mtaylor and jeblair put as few or as many barriers as you want (pep8, pylint, whatever you want). | 19:01 |
| *** rnirmal_ has quit IRC | 19:02 | |
| *** rnirmal has joined #openstack-dev | 19:02 | |
| *** bsza has quit IRC | 19:03 | |
| *** bsza has joined #openstack-dev | 19:03 | |
| jaypipes | notmyname: also, to be clear, the keystone code review process has some automated barriers (currently pylint checks), but for small, non-contentious pull requests, any keystone-core dev can set the review to +2 to "send it on" to Jenkins and the automated builders. As long as nobody has done a -1 review on the patch, that's enough to send it on. | 19:05 |
| jaypipes | notmyname: but, as mentioned, you can set up whatever policies fit the swift team best. | 19:05 |
| jaypipes | notmyname: and thx for trying it out. appreciate the feedback and questions. | 19:05 |
| *** cp16net has quit IRC | 19:07 | |
| *** mdomsch has joined #openstack-dev | 19:11 | |
| *** TimR has quit IRC | 19:13 | |
| *** TimR has joined #openstack-dev | 19:14 | |
| openstackgerrit | Yogeshwar Srikrishnan proposed a change to openstack/keystone: Changes to support services. https://review.openstack.org/137 | 19:20 |
| *** mnour has joined #openstack-dev | 19:21 | |
| *** deshantm_laptop has joined #openstack-dev | 19:21 | |
| *** adambergstein has quit IRC | 19:22 | |
| notmyname | jaypipes: jeblair: will all reviews for all projects be listed for everyone? (eg https://review.openstack.org/#q,status:open,n,z) | 19:23 |
| jaypipes | notmyname: good question. I'm not sure. jeblair or mtaylor? | 19:23 |
| *** nati has quit IRC | 19:23 | |
| *** TimR has quit IRC | 19:32 | |
| *** TimR has joined #openstack-dev | 19:33 | |
| *** adambergstein has joined #openstack-dev | 19:36 | |
| openstackgerrit | Yogeshwar Srikrishnan proposed a change to openstack/keystone: Changes to support services. https://review.openstack.org/137 | 19:37 |
| *** tudamp has quit IRC | 19:45 | |
| *** cp16net has joined #openstack-dev | 19:46 | |
| openstackjenkins | Project swift build #302: SUCCESS in 29 sec: http://jenkins.openstack.org/job/swift/302/ | 19:47 |
| openstackjenkins | Tarmac: added Ewan to authors file | 19:47 |
| notmyname | jaypipes: jeblair: mtaylor: also, with gerrit forces updates to reviewed code be submitted essentially a new review submission (although it has the same gerrit id). since each review is one commit (squashed before review) it becomes difficult for the reviewer to see what's changed between the two submissions | 19:49 |
| jaypipes | comstud, s1rp: https://blueprints.launchpad.net/glance/+spec/split-unified-api. Targeted to Essex. | 19:49 |
| comstud | cools | 19:49 |
| *** nati_ has joined #openstack-dev | 19:50 | |
| jaypipes | notmyname: Diff from Previous Set button should show what you want... | 19:51 |
| notmyname | ok | 19:51 |
| notmyname | jaypipes: got any examples of this? | 19:57 |
| jaypipes | notmyname: *blush* I confused what ttx had said earlier about this feature as already being in Gerrit.. apparently, it's not done yet: http://code.google.com/p/gerrit/issues/detail?id=379 | 19:58 |
| jaypipes | notmyname: sorry man :( | 19:58 |
| jaypipes | notmyname: I'm hoping mtaylor or jeblair have a solution or a suggestion on this. I think they're both afk right now. | 19:59 |
| *** bcwaldon has quit IRC | 20:04 | |
| *** bcwaldon has joined #openstack-dev | 20:06 | |
| notmyname | jaypipes: jeblair: mtaylor: seems to me that using gerrit removes almost all of the reasons for using github (github's pull request/issue management being one of them). so why use github at all and not something simple like gitweb or gitorious or something at git.openstack.org? | 20:13 |
| notmyname | ie the gerrit workflow replaces github | 20:14 |
| jaypipes | notmyname: do you really want to get into this conversation again? | 20:15 |
| notmyname | jaypipes: I'm offering my thoughts as requested. also, I don't think there was ever a conversation on gerrit | 20:15 |
| notmyname | I'm not arguing git vs bzr | 20:15 |
| termie | (git) | 20:16 |
| zykes- | dolphm: what's the difference of the stuff under controllers and IdentityService class ? | 20:16 |
| jaypipes | ugh. | 20:16 |
| dolphm | zykes-: looking.. | 20:16 |
| jaypipes | notmyname: we've had the "why can't we just use GitHub as is" conversation already a hundred times. | 20:17 |
| termie | funny that people would keep asking for that | 20:17 |
| jaypipes | termie: seriously, dude, not today. | 20:17 |
| dolphm | zykes-: don't know! | 20:18 |
| zykes- | :| | 20:19 |
| zykes- | is there any plans to "commonize" stuff ? like wsgi.py under projects... | 20:19 |
| zykes- | config.py | 20:19 |
| yogirackspace | controllers deal with contents coneversion and if any needed any endpoint related stuff | 20:19 |
| dolphm | zykes-: it looks like hte controllers all use config.SERVICE which *is* IdentityService | 20:19 |
| zykes- | what's the "class IdentityService" then under service.py dolphm ? | 20:20 |
| yogirackspace | where else services are designed not to deal with domain models | 20:20 |
| dolphm | zykes-: ask yogi ^ | 20:20 |
| yogirackspace | that was the reason for seperation | 20:20 |
| kbringard | oh, hey Yogi | 20:21 |
| zykes- | yogirackspace: ? | 20:21 |
| yogirackspace | sorry services are designed to deal with domain models | 20:21 |
| kbringard | I may need to read back if you answered this already, but I was wondering if you could explain what the public_url is for in keystone | 20:21 |
| dolphm | yogirackspace: these things should be documented at the module level, because it's definitely not clear | 20:21 |
| *** cp16net has quit IRC | 20:22 | |
| zykes- | domain models is what yogirackspace ? | 20:22 |
| yogirackspace | public_url are urls that the end users of a service could actualluy use to consume the service | 20:22 |
| kbringard | so it's similar to a vhost of sorts? | 20:23 |
| *** cp16net has joined #openstack-dev | 20:23 | |
| kbringard | because, in theory I could put some external IP as the internal_url and have that be the public_url too, right? | 20:23 |
| yogirackspace | you could | 20:24 |
| yogirackspace | internal ip is for services to use among themselfes | 20:24 |
| yogirackspace | they all might be under same network | 20:24 |
| zykes- | yogirackspace: what is the diff on controllers stuff and the identityservice stuff ? | 20:24 |
| yogirackspace | controllers were to do deal with endpoint conversions and stuff like that | 20:26 |
| zykes- | and the identity service ? | 20:26 |
| yogirackspace | sxample a req object might not be there for services at all | 20:26 |
| kbringard | so the public_url is more for the software to know what people will be coming in from, and not necessarily something for keystone to use to talk on, since it's going to use the admin_url and the internal_url | 20:27 |
| yogirackspace | it was supposed to deal with domain level operations without worrying about the data conversions | 20:27 |
| kbringard | ah, OK | 20:27 |
| yogirackspace | its more for actual users to consume respective services | 20:28 |
| yogirackspace | they wud use the public url along with the tokens to perform operations on individual service | 20:29 |
| kbringard | ah, OK | 20:30 |
| kbringard | got it | 20:30 |
| kbringard | thanks :-) | 20:30 |
| yogirackspace | cool | 20:30 |
| kbringard | if you have a moment, I have a couple other questions too :-D | 20:30 |
| yogirackspace | sure | 20:30 |
| kbringard | but I don't want to derail you from real work | 20:30 |
| yogirackspace | u cud email me or any one from the team | 20:30 |
| yogirackspace | u cud email openstack list we ll answer | 20:31 |
| jaypipes | notmyname: I've noted your feedback on Gerrit (summary: it sucks, we should really just be using GitHub anyway). I'll include it in the email creiht asked me to send out about Gerrit. | 20:31 |
| kbringard | ok, dolphm gave me your address so I'll just email you, if you don't mind | 20:31 |
| yogirackspace | sure | 20:32 |
| yogirackspace | :) | 20:32 |
| dolphm | mtaylor: problems! you there? | 20:34 |
| notmyname | jaypipes: is that really what I said? now having actually used gerrit in a limited way (ie one commit), I'm asking questions that I have based on my experience with our workflow | 20:36 |
| jaypipes | notmyname: it was very clear to me what you were saying. | 20:37 |
| dolphm | jaypipes: the new Change-Id required-for-every-commit is borked | 20:37 |
| jaypipes | dolphm: could you explain, pls? | 20:37 |
| *** sandywalsh has quit IRC | 20:38 | |
| notmyname | jaypipes: I never once offered a value judgement on gerrit | 20:38 |
| zykes- | what's the reason for having "wsgi.py" and those files independant in each project? | 20:38 |
| dolphm | jaypipes: i have the hook installed, i pulled from openstack/keystone which did an auto-merge (and thus a commit with no change-id), i did my own (normal) commit with a change-id, and then tried to git review... the merge is blocking my review | 20:38 |
| dolphm | zykes-: excellent question - i'm sort of hoping they get swallowed by openstack-common | 20:39 |
| zykes- | is there any such project ? | 20:39 |
| *** bcwaldon has quit IRC | 20:40 | |
| jaypipes | notmyname: I'm not sure anyone would read your comment above and not think that it was value judgment. | 20:40 |
| dolphm | zykes-: i think it was just started | 20:40 |
| dolphm | zykes-: https://github.com/openstack/openstack-skeleton | 20:41 |
| jaypipes | notmyname: not talking about the questions regarding diff and about listing all projects in the review screen. you know what comment I am referring to. | 20:41 |
| dolphm | zykes-: i guess skeleton != common ... doesn't solve duped code | 20:42 |
| jaypipes | dolphm: skeleton is the baking ground for common, which is why skeleton includes an openstack/ directory (that is openstack-common). | 20:42 |
| jaypipes | dolphm: since nobody can seem to agree on even the simplest of things between the projects, openstack-common is basically dead on arrival until someone says "this is a skeleton project with an openstack lib that works." | 20:43 |
| dolphm | jaypipes: well anyway, i'm looking forward to importing it and dumping all the out-of-scope code in keystone :) | 20:44 |
| jaypipes | dolphm: you and me both. | 20:45 |
| eday | jaypipes: to be fair, I think notmyname was only asking if there was a reason we were still using github for code hosting if we're not using issues or pull requests. I think the answer is so we don't have to host our own code hosting too, and you can still use GH for branch/fork management | 20:47 |
| jaypipes | dolphm: so, re: the changeId hook, gotta link? | 20:47 |
| notmyname | eday: correct. but with gerrit you don't even use the branches in github (since gerrit is the one managing them) | 20:48 |
| dolphm | jaypipes: http://paste.openstack.org/show/2034/ | 20:49 |
| jaypipes | notmyname: what do you mean by "you don't even use the branches in github"? | 20:49 |
| termie | jaypipes, notmyname: i assume meaning that we push to gerrit rather than pushing to github and issuing pull requests | 20:50 |
| notmyname | jaypipes: all my branch management is local with gerrit since gerrit handles all of the merging | 20:50 |
| dolphm | i use my own github fork to sync between my own workstations.. | 20:50 |
| jaypipes | termie: right. because GitHub's pull requests are inadequate and it's not open source, so we can't edit it for our needs. | 20:51 |
| termie | jaypipes: i didn't bring that up | 20:51 |
| termie | jaypipes: :D | 20:51 |
| jaypipes | notmyname: ok, well, I'll take your word that your question was an honest question and not just yet another jab at not using GitHub for everything. eday answered your question: so we don't have to do code hosting ourselves. | 20:52 |
| termie | tangentially, i've written some stuff that allows easy migration of issues, milestones and blueprints from lp to gh, in case anybody else is interested | 20:54 |
| jaypipes | no | 20:54 |
| jaypipes | termie: for bugs/blueprints/milestones, personally for Glance I'm not interested in giving up lots of functionality to get rounded corners in the UI. | 20:56 |
| jaypipes | dolphm: ok, so reading through your link there, it looks like the commit hook should recognize a merge commit and not error on the Change-Id being missing. correct? | 20:57 |
| *** alfred_mancx has quit IRC | 20:57 | |
| dolphm | jaypipes: yessir | 20:57 |
| jaypipes | dolphm: k, lemme grab jim and chat when he gets back. thx for your patience. | 20:58 |
| termie | jaypipes: rounded corners AND gradients | 20:58 |
| dolphm | jaypipes: np | 20:58 |
| jaypipes | termie: ah, sorry, forgot about that. | 20:58 |
| mtaylor | dolphm, jaypipes hey, whazzup? | 20:58 |
| jaypipes | termie: and a groovy 404 image that moves with your mouse pointer. glad they spent the time to get that done! | 20:59 |
| dolphm | yo | 20:59 |
| termie | jaypipes: gravatar integration! that stuff is hot | 20:59 |
| dolphm | mtaylor: relevant: http://paste.openstack.org/show/2034/ | 20:59 |
| jaypipes | termie: lol | 20:59 |
| dolphm | termie: but can they do rounded corners on my gravatar? | 20:59 |
| jaypipes | dolphm: I'm sure if you found them and bought enough micro-brews they'd do it. | 21:00 |
| dolphm | jaypipes: i would actually hate that... my gravatar is square and fuck you if you try and make it shiny | 21:00 |
| termie | dolphm: sounds like a chrome extension to me, i have one to remove most of the icons and extra content on the launchpad pages | 21:00 |
| jaypipes | dolphm: might even prioritize it over smaller features like an email interface... | 21:01 |
| eday | we could just use browser plugins to style LP to look like GH :) | 21:01 |
| termie | eday: i wish it were that easy | 21:01 |
| jaypipes | eday: I would die. that would be awesome. :) | 21:01 |
| dolphm | lol | 21:01 |
| dolphm | mtaylor: so, after this change-id guard-ignoring-merges issue, can you add a zero-tolerance pep8 blocker on keystone? | 21:02 |
| mtaylor | notmyname, jaypipes: hey! I'm only half-reading the scrollback ... we're replicating to github because it's an easier place for people to do the whole "fork/pull" thing than asking them to clone directly from gerrit - and because people seem to like that as an easy to understand place to go to find a copy of the code | 21:02 |
| creiht | jaypipes: I'm not trying to stir the hornet's nest, but I think you were a bit unfair and mis-characterized his question | 21:03 |
| termie | mtaylor: but we're also completely supplanting the usual flows people are used to when using github | 21:03 |
| mtaylor | dolphm: I think jeblair is actually the person to look at the Change-ID thing - but I know in the past that if the Change-ID has a space before it or is not on the final line, things get wonky. jim should be done with lunch soon | 21:03 |
| creiht | I still think it is a valid question, if we are only using github for code hosting, and not any other features, why do we not host it ourselves? | 21:03 |
| creiht | we are already hosting the gerrit stuff | 21:03 |
| mtaylor | sure. well, we _do_ host it ourselves | 21:03 |
| termie | mtaylor: people are used to seeing a project, forking and issuing pull requests, we don't want them to do anything but look at the code? | 21:04 |
| creiht | and there are some decent tools that allow for the git hosting (like gitorious) | 21:04 |
| mtaylor | github is a mirror for conveniece for people who would like to fork code there | 21:04 |
| dolphm | mtaylor: the issue is that the Change-Id hook doesn't add ID's on merges, and then gerrit is blocking | 21:04 |
| mtaylor | creiht: totally - and I have no objections to also running a gitorious mirror or something of that nature | 21:04 |
| mtaylor | dolphm: hrm. yes. we need to loop in jeblair | 21:05 |
| dolphm | mtaylor: should i just email him? | 21:05 |
| mtaylor | dolphm: he's just at lunch - should be back soon | 21:06 |
| mtaylor | dolphm: but yeah, you can also email him | 21:06 |
| creiht | mtaylor: then I vote that the canonical location should be something like git.openstack.org, and we keep the github as a mirror | 21:06 |
| jeblair | hiya | 21:06 |
| creiht | to avoid confusion | 21:06 |
| jeblair | i'm caught up on scrollback. :) | 21:06 |
| dolphm | jeblair: hola | 21:06 |
| *** jjm has quit IRC | 21:06 | |
| creiht | otherwise if github is the cannonical location for the source, you are going to have a lot of people trying to contribute the github way | 21:07 |
| termie | creiht, mtaylor: and maybe get some bugzilla love in there to remove lp completely | 21:07 |
| *** _cerberus_ has quit IRC | 21:07 | |
| mtaylor | termie: you REALLY hate launchpad don't you | 21:07 |
| *** mnaser has quit IRC | 21:08 | |
| mtaylor | creiht: I'd be fine with something like that ... shall we look in to what that looks like and make some suggestions on how it could work? | 21:08 |
| * creiht shrugs | 21:08 | |
| jeblair | this is fun: i'm researching the change-id on merge commit problem, and my own documentation on the openstack wiki just came up as result #3 in google | 21:08 |
| creiht | I'm just making observations | 21:08 |
| mtaylor | :) | 21:08 |
| dolphm | jeblair: win! | 21:08 |
| *** mdomsch has quit IRC | 21:09 | |
| termie | mtaylor: yup, but not in a particularly emotional way, it is more like it is a roadblock between me and a pleasurable working environment | 21:09 |
| termie | our gerrit has a particularly lovely color scheme, btw, which is why i thought of bugzilla | 21:10 |
| mtaylor | termie: fair enough. I mean - I feel that way about bugzilla, so I know what you're talking about at least :) | 21:10 |
| notmyname | mtaylor: I think it could be valuable to see what other git hosting options look like if we aren't using the particular featureset that github offers | 21:10 |
| *** _cerberus_ has joined #openstack-dev | 21:10 | |
| mtaylor | notmyname: there are some examples already of folks doing a git+gerrit+gitorious setup ... jeblair, who was that project we were looking at who did that? | 21:10 |
| *** jaypipes has quit IRC | 21:11 | |
| jeblair | mtaylor: mahara | 21:11 |
| notmyname | mtaylor: I tend to agree with termie that using github but not any of the standard github "ways" will probably cause much confusion | 21:11 |
| jeblair | https://wiki.mahara.org/index.php/Developer_Area/Developer_Tools | 21:11 |
| mtaylor | notmyname, creiht: definitely would be happy to look at those | 21:11 |
| jeblair | gitorious does support openid; that could make cross-tool SSO nicer | 21:12 |
| mtaylor | ++ | 21:12 |
| mtaylor | jeblair: we can run our own gitorious, right? so we could actually manage the SSO stack in a sensible manner? (I'm imagining people would like it if they didn't have to have a lp account to do things) | 21:13 |
| mtaylor | (not saying we need to - just thinking out loud) | 21:13 |
| jeblair | and yes, self-hosting git repos (including turning on gitweb support in gerrit) are other options we can consider. i have deferred experimenting with gitweb and gerrit since we've been considering github as the public facing code hosting, but i do know that gerrit has goot integration with it | 21:13 |
| jeblair | mtaylor: yes, i'm agnostic about what the sso stack _is_, but do think we should have one. | 21:14 |
| jk0 | I might have missed this (and don't want to stir anyting up) but why aren't we just using github for hosting and code reviews? | 21:14 |
| jk0 | instead of gerret | 21:14 |
| mtaylor | jk0: the main reason is lack of a workflow state in github pull requests | 21:14 |
| mtaylor | jk0: which then makes jenkins integration pre-merge a bit difficult to model in a consistent manner | 21:15 |
| jeblair | dolphm: regarding merge commits, apparently using "git pull --rebase" can help avoid them | 21:15 |
| jk0 | it's probably not ideal but we could just come up with our own way to give thumbs up or down | 21:15 |
| mtaylor | jk0: yeah, we went down that road for a bit | 21:15 |
| jk0 | right now on LP you could still approve w/o two +1s, couldn't you? | 21:15 |
| jeblair | dolphm: i'm still leaning toward "merge commits don't fit well in our recommended workflow", so my first idea would be to try to avoid them. | 21:15 |
| jk0 | or does jenkins know better | 21:15 |
| mtaylor | jk0: different thing | 21:16 |
| mtaylor | jk0: it's not about enforcing vote vs. approve policy | 21:16 |
| dolphm | jeblair: it's not really a merge.. it was just a pull | 21:16 |
| *** mnaser has joined #openstack-dev | 21:16 | |
| termie | mtaylor, jk0: there is still a person involved clicking 'merge this'... | 21:16 |
| mtaylor | jk0: it's about the difference between tracking votes and holding the state of the request itself | 21:16 |
| dolphm | jeblair: which, yes, has a merge, but it's not like i'm merging in someone else's commits that haven't already been through gerrit/jenkins | 21:16 |
| jeblair | dolphm: pulls can be either ff or merge depending on what has happened to the remote | 21:16 |
| jeblair | dolphm: do you have a github location where you can push your repo? | 21:17 |
| jk0 | mtaylor: yeah, it would be a pain, but I think we could get used to it | 21:17 |
| dolphm | jeblair: yep | 21:17 |
| termie | mtaylor, jk0: were i discussing this i would probably claim that much of our process comes from a lp-centric image of the world, while other people have indeed successfully used other processes | 21:17 |
| jeblair | notmyname: 12:23 <+notmyname> jaypipes: jeblair: will all reviews for all projects be listed for everyone? (eg https://review.openstack.org/#q,status:open,n,z) | 21:18 |
| jk0 | termie / mtaylor: that's my point.. why not consider a new process? | 21:18 |
| mtaylor | termie, jk0: people TOTALLY use other processes - the actual driving factor here is the jenkins-runs-pre-merge workflow to produce a gated trunk | 21:18 |
| jk0 | or just a workaround (manually doing stuff) until we can find something better | 21:18 |
| jeblair | notmyname: yes, they'll all be there. However, this page: https://review.openstack.org/#q,is:watched+status:open,n,z | 21:18 |
| mtaylor | jk0: because there is nothing broken in the process? | 21:18 |
| *** jjm has joined #openstack-dev | 21:19 | |
| mtaylor | jk0: I'm not saying others wouldn't also work | 21:19 |
| jeblair | notmyname: (My -> watched changes) should show you changes that are interesting to you based on what project groups you are a member of | 21:19 |
| notmyname | jeblair: good | 21:19 |
| jk0 | mtaylor: not saying anything's broken, but like termie said we're sticking to the LP style of merges which led to gerret | 21:19 |
| jeblair | notmyname: 12:49 <+notmyname> jaypipes: jeblair: mtaylor: also, with gerrit forces updates to reviewed code be submitted essentially a new review submission (although it has the same gerrit id). since each review is one commit (squashed before review) it becomes difficult for the reviewer to see what's changed between the two submissions | 21:20 |
| creiht | jk0:++ | 21:20 |
| *** jaypipes has joined #openstack-dev | 21:20 | |
| mtaylor | jk0: not really - it's not really about lp or gh style merges - it's about serving a jenkins shotgun on the trunk | 21:20 |
| jeblair | notmyname: i _think_ that should be handled by selecting "Patch history" while viewing a diff | 21:21 |
| mtaylor | jk0: gerrit happens to feed the ability to have code reviews trigger testing and then flow in to merging to trunk - and is already tightly integrated with jenkins | 21:21 |
| termie | mtaylor, jk0: my hypothetical statement wasn't referring to merges specifically, but embedded process in general | 21:21 |
| jeblair | notmyname: where you can thes select whether to diff based on base, or any submitted patchset | 21:21 |
| notmyname | jeblair: do you have any examples? it's something that is fairly common | 21:21 |
| jeblair | notmyname: here is a simple change that has 2 patchsets: https://review.openstack.org/#change,139 | 21:21 |
| termie | mtaylor, jk0: an example of which might be having a bot merge your code for you | 21:22 |
| jk0 | termie / mtaylor: sure, that's why I think we should expore a few more options before making gerret official (or is it too late?) | 21:22 |
| notmyname | jeblair: so how do I see the diff between the patchsets? | 21:22 |
| mtaylor | jk0: so - here's the thing ... at the moment, we have the current process, which has a fairly well designed set of requirements | 21:23 |
| dolphm | notmyname: select a patch set under "Old Version History" | 21:23 |
| jaypipes | notmyname: https://review.openstack.org/#patch,sidebyside,139,2,update_gerrit_users.py | 21:23 |
| jaypipes | notmyname: select patch set 1 under Old Version History, then hit side by side on the file... | 21:23 |
| mtaylor | jk0: and we have thoughts that we could do something else, but of those myriad possible something elses (and there are honestly many different ways we could go here) they are all essentially just possibilities that don't have sets of requirements | 21:24 |
| mtaylor | jk0: so without a concrete problem to solve, I'm pretty sure we'd start "exploring" things, but with nothing other than personal preference or things people happen to have done before guiding us | 21:24 |
| jaypipes | creiht: sorry, got disconnected.. reading back in log. | 21:25 |
| *** dolphm has quit IRC | 21:25 | |
| jeblair | notmyname: yes, using "old version history" will change the file list on the main page for the change. then when you are doing a side by side diff of a file, "patch history" will let you change what the diff is based on | 21:25 |
| jk0 | mtaylor: that's fair enough. at the very least, I'd really like to see a new UI for gerret (hopefully nokia will not let us down) | 21:26 |
| notmyname | jeblair: ya, trying that out now. | 21:26 |
| mtaylor | jk0: I mean, gerrit wasn't my first choice, but as I've worked with it, I've become pretty pleased with its possibilities and extra pleased with how comprehensively it's integreated with jenkins | 21:26 |
| mtaylor | jk0: ++ | 21:26 |
| mtaylor | jk0: although the nice thing is that if one of us gets enterprising, we can certainly patch the UI | 21:26 |
| termie | lulz | 21:27 |
| jk0 | true, and I'm sure that will have to happen eventually | 21:27 |
| mtaylor | yes. I agree | 21:27 |
| jeblair | jk0: i don't know if nokia is working on a whole new ui, but they are making some usability changes around patch viewing | 21:27 |
| jk0 | because that's one of the worst UIs I've ever seen, no offense to anyone | 21:27 |
| mtaylor | not first on my list of things to do - but it's on there | 21:27 |
| termie | jk0: bugzilla! | 21:27 |
| mtaylor | jk0: what termie said | 21:27 |
| mtaylor | :) | 21:27 |
| jk0 | haha | 21:27 |
| jeblair | gerrit is based on google web toolkit. | 21:27 |
| termie | jeblair: aka nobody is ever going to touch that ui | 21:28 |
| mtaylor | termie: I dunno - it might piss of steve jobs one day and he might have some rounded corners and some brushed metal added | 21:28 |
| termie | mtaylor: ? | 21:29 |
| mtaylor | jk0: there's also an ssh/command-line interface if you're in to that sort of thing | 21:29 |
| termie | mtaylor: brushed metal is so 2005 | 21:29 |
| mtaylor | termie: I can't keep up with apple fashion trends. what are we on now? white? that's passe too, right? | 21:30 |
| jk0 | mtaylor: very much so | 21:30 |
| jeblair | i realize this is the opposite of rounded corners, but yeah, there is a command line interface for many actions as mtaylor mentioned, and if you hit '?' in the web ui, you'll get a keyboard shortcut reference | 21:30 |
| jk0 | the more things I can do out of my screen shell, the better | 21:30 |
| mtaylor | jk0: I think you can do a LOT of things via command line | 21:31 |
| mwhooker | what's the status of doctests within openstack. I see there are none. would people be amenable to adding them, or would they prefer the tests go in separate modules? | 21:31 |
| jk0 | I'll have to look more into it when I get time | 21:31 |
| jeblair | jk0: https://review.openstack.org/Documentation/cmd-index.html | 21:31 |
| jk0 | nice, thanks jeblair | 21:31 |
| jeblair | np | 21:31 |
| termie | mtaylor, jk0: btw, i'm fine with gerrit, i just think we will also get pull requests on gh, and i think it is disappointing to also use lp | 21:32 |
| jk0 | yeah, I'm not sure we should even have a GH mirror | 21:32 |
| jk0 | no point in it if we're hosting our own repo | 21:32 |
| mtaylor | termie: definitely agree on the first, totally understand the second ... let's keep talking about how we can deal with the lp thing in a way that doesn't kill our release management side | 21:32 |
| *** markvoelker has quit IRC | 21:33 | |
| *** Tushar has quit IRC | 21:33 | |
| termie | whoa dudes, bugzilla has rounded corners now: https://bugzilla.mozilla.org/show_bug.cgi?id=309330 | 21:35 |
| uvirtbot | Mozilla bug 309330 in Location Bar "unable to force the download of a location bar address" [Enhancement,Unconfirmed: ] | 21:35 |
| mtaylor | jk0, termie: jeblair and I will put together some thoughts on alternatives to the github mirror - I think that's a really good point | 21:35 |
| mtaylor | termie: CRAZY | 21:36 |
| notmyname | mtaylor: thanks | 21:36 |
| mtaylor | notmyname: my pleasure! I really do want our infrastructure to make sense for everyone - and I think at some point we'll get there! :) | 21:37 |
| notmyname | so one more question: I think this is something that is being evaluated as one of a "vetted set" of options, right? the existing bzr/lp process will still be available? or is this designed to be _the_ way openstack dev is done? | 21:39 |
| *** jkoelker has quit IRC | 21:41 | |
| termie | the latter, as i understand it | 21:41 |
| eday | well, according to the last PPB vote, it was a tie (options vs one way) | 21:43 |
| termie | exciting | 21:43 |
| mtaylor | notmyname: I have my opinions there- and I sort of want the second | 21:43 |
| eday | notmyname: so I think we'll need to revisit that once jbryce has the autonomy doc done | 21:44 |
| mtaylor | notmyname: s/sort of/completely/ | 21:44 |
| notmyname | good to know | 21:45 |
| jeblair | mtaylor: i think dolphm is still gone, but i'm leaning toward rebase being the right answer for merge commits | 21:46 |
| *** bsza has quit IRC | 21:46 | |
| jeblair | ie, rebase and don't submit merge commits to gerrit | 21:46 |
| eday | jeblair: would the rebase happen pre or post review? | 21:47 |
| *** sandywalsh has joined #openstack-dev | 21:47 | |
| jeblair | eday: pre -- dolphm's problem is that he ended up with a merge commit when updating his working branch | 21:48 |
| jeblair | eday: so i think the suggestion would be "if you update your working branch from upstream, use git pull --rebase" | 21:48 |
| eday | jeblair: I guess I should ask which which use of rebase as well. basically, I'd hate to loose the commit history during review if there was one to do it in chunks | 21:49 |
| jeblair | so with gerrit, each commit becomes a change (which is one reason we don't want merge commits; they generate extra changes) | 21:51 |
| jeblair | so if there is a meaningful commit history where commits should be individually reviewed (or reviewed in sequence), that makes sense to upload several commits | 21:51 |
| *** deshantm_laptop has quit IRC | 21:51 | |
| jeblair | but we recommend squashing what we're calling "checkpoint commits" before uploading them for review | 21:52 |
| jeblair | ie, if you work on a change for a long time and check in your work frequently, but each of those commits shouldn't be reviewed, only the result, then squash them | 21:52 |
| jeblair | (devs can always keep their private history on another branch if they want) | 21:53 |
| eday | Ok. I guess I was thinking it would be squashed post-review right before trunk (could do ff-only then for trunk maybe?) | 21:57 |
| eday | But what you say makes sense, if it's big enough to want to review different commits, should probably be different reviews :) | 21:57 |
| jeblair | right now we have gerrit set to ff or merge when necessary, which gives it a little more flexibility in merging patches | 21:58 |
| jeblair | so gerrit itself can create merge commits to facilitate dev B's change getting merged after dev A's, without dev B having to rebase | 21:58 |
| jeblair | otherwise, devs would likely need to rebase after any change to trunk, which would be...annoying. :) | 21:59 |
| mwhooker | if I wanted to get the discussion about doctests going, where would be the best place to pose it? | 21:59 |
| termie | it has already begun | 21:59 |
| termie | (because you just started it) | 22:00 |
| mwhooker | boom | 22:00 |
| mwhooker | what if I were to commit some doctests and modify run_tests to run nose with --with-doctests | 22:00 |
| mwhooker | checking, of course, that it doesn't slow down the tests too much | 22:00 |
| *** joar has joined #openstack-dev | 22:01 | |
| joar | Does Swift fully support the python-cloudfiles bindings? | 22:01 |
| mwhooker | I'm yak shaving the nova/api/openstack/common.py file and notice there are essentially doctests describing behavior of, eg, get_id_from_href | 22:02 |
| mwhooker | so it would be nice to formalize those | 22:02 |
| *** nati_ has quit IRC | 22:12 | |
| *** TimR has quit IRC | 22:13 | |
| *** TimR has joined #openstack-dev | 22:14 | |
| creiht | joar: the python-cloudfiles bindings should work with swift | 22:16 |
| *** RobertLaptop has quit IRC | 22:17 | |
| *** j05h has joined #openstack-dev | 22:18 | |
| *** j05h has left #openstack-dev | 22:18 | |
| *** RobertLaptop has joined #openstack-dev | 22:19 | |
| joar | thanks creiht! | 22:22 |
| joar | Is there any service out there using openstack currently? | 22:22 |
| creiht | joar: swift is what backs cloud files | 22:24 |
| creiht | at rackspace | 22:24 |
| joar | Ok, that's great. | 22:24 |
| creiht | there are several other deployments of swift | 22:25 |
| creiht | Internap, Nephoscale, Korea Telecom off the top of my head | 22:25 |
| creiht | Can't speak for Nova | 22:26 |
| creiht | Wikimedia was experimenting with it to store images for wikipedia, but I never heard if they turned it all the way on | 22:27 |
| joar | creiht: I was actually wondering because I needed to confirm that my development will also work with Swift setups. | 22:27 |
| creiht | ahh cool | 22:28 |
| creiht | it should then | 22:28 |
| joar | But if Cloud Files are powered by Swift, I've already done so. | 22:28 |
| joar | :) | 22:28 |
| creiht | cool, may I ask which tool? | 22:28 |
| joar | MediaGoblin | 22:28 |
| joar | or GNU MediaGoblin | 22:29 |
| creiht | The only main difference between the two, is to make the auth endpoint url configurable | 22:29 |
| joar | Yes, that's implemented :) | 22:29 |
| joar | also servicenet. | 22:29 |
| creiht | cool | 22:29 |
| joar | Works perfect, the images load 4.5 times as fast as the CSS ;) | 22:29 |
| creiht | haha | 22:29 |
| joar | To the defense of the CSS it's hosted via pagekite. | 22:30 |
| joar | http://mg.jwandborg.pagekite.me/u/joar/m/4e39b15fd9d7834655000000-balloons-3/ | 22:30 |
| creiht | that's pretty cool | 22:30 |
| openstackgerrit | Monty Taylor proposed a change to openstack/openstack-ci-puppet: Silence unneeded output. https://review.openstack.org/140 | 22:32 |
| joar | creiht: Yeah, it is :) | 22:34 |
| Daviey | 23:36 < pakaran> 5*50/30k you mean | 22:37 |
| Daviey | hmm, not sure what happend there | 22:38 |
| *** kbringard has quit IRC | 22:40 | |
| *** mattray has quit IRC | 22:43 | |
| openstackgerrit | Monty Taylor proposed a change to openstack/openstack-ci-puppet: Silence unneeded output. https://review.openstack.org/140 | 22:45 |
| openstackgerrit | A change was merged to openstack/openstack-ci-puppet: Silence unneeded output. https://review.openstack.org/140 | 22:46 |
| *** amccabe has quit IRC | 22:51 | |
| openstackgerrit | Vish Ishaya proposed a change to openstack/keystone: Adds support for authenticating via ec2 signatures https://review.openstack.org/127 | 23:03 |
| *** yogirackspace has left #openstack-dev | 23:05 | |
| *** jakedahn has quit IRC | 23:19 | |
| *** rnirmal has quit IRC | 23:42 | |
| *** jc_smith has joined #openstack-dev | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!