| *** thiagop_ has quit IRC | 00:02 | |
| *** dims has quit IRC | 00:02 | |
| *** vhoward has joined #openstack-keystone | 00:04 | |
| *** dolphm has joined #openstack-keystone | 00:08 | |
| *** zhiyan` has joined #openstack-keystone | 00:08 | |
| *** Daviey has joined #openstack-keystone | 00:08 | |
| *** harlowja has joined #openstack-keystone | 00:08 | |
| *** marcoemorais1 has joined #openstack-keystone | 00:08 | |
| *** rwsu_ has joined #openstack-keystone | 00:08 | |
| *** YorikSar has joined #openstack-keystone | 00:08 | |
| *** sudorandom has joined #openstack-keystone | 00:08 | |
| *** kfox1111 has joined #openstack-keystone | 00:08 | |
| *** amcrn has joined #openstack-keystone | 00:08 | |
| *** gyee has joined #openstack-keystone | 00:08 | |
| *** jimbaker has joined #openstack-keystone | 00:08 | |
| *** vhoward- has joined #openstack-keystone | 00:08 | |
| *** dvorak has joined #openstack-keystone | 00:08 | |
| *** marekd|away has joined #openstack-keystone | 00:08 | |
| *** wchrisj has joined #openstack-keystone | 00:08 | |
| *** mberlin has joined #openstack-keystone | 00:08 | |
| *** lbragstad has joined #openstack-keystone | 00:08 | |
| *** haneef_ has joined #openstack-keystone | 00:08 | |
| *** bobt has joined #openstack-keystone | 00:08 | |
| *** flaper87|afk has joined #openstack-keystone | 00:08 | |
| *** bknudson has joined #openstack-keystone | 00:08 | |
| *** jordant has joined #openstack-keystone | 00:08 | |
| *** zigo has joined #openstack-keystone | 00:08 | |
| *** tellesnobrega has joined #openstack-keystone | 00:08 | |
| *** jraim has joined #openstack-keystone | 00:08 | |
| *** huats has joined #openstack-keystone | 00:08 | |
| *** zhiyan_ has joined #openstack-keystone | 00:08 | |
| *** mhu has joined #openstack-keystone | 00:08 | |
| *** ayoung has joined #openstack-keystone | 00:08 | |
| *** chmouel has joined #openstack-keystone | 00:08 | |
| *** mfisch has joined #openstack-keystone | 00:08 | |
| *** jaypipes has joined #openstack-keystone | 00:08 | |
| *** jamielennox|away has joined #openstack-keystone | 00:08 | |
| *** koolhead17 has joined #openstack-keystone | 00:08 | |
| *** luisbg has joined #openstack-keystone | 00:08 | |
| *** anteaya has joined #openstack-keystone | 00:08 | |
| *** dtroyer has joined #openstack-keystone | 00:08 | |
| *** morganfainberg has joined #openstack-keystone | 00:08 | |
| *** ChanServ has joined #openstack-keystone | 00:08 | |
| *** dickson.freenode.net sets mode: +o ChanServ | 00:08 | |
| *** dstanek_afk has quit IRC | 00:18 | |
| *** openstack has joined #openstack-keystone | 00:21 | |
| *** openstackstatus has quit IRC | 00:25 | |
| *** openstack has quit IRC | 00:26 | |
| *** openstack has joined #openstack-keystone | 00:31 | |
| *** dickson.freenode.net sets mode: +ns | 00:31 | |
| -dickson.freenode.net- [freenode-info] channel trolls and no channel staff around to help? please check with freenode support: http://freenode.net/faq.shtml#gettinghelp | 00:31 | |
| *** dickson.freenode.net sets mode: -o openstack | 00:36 | |
| -dickson.freenode.net- *** Notice -- TS for #openstack-keystone changed from 1394843518 to 1377384024 | 00:36 | |
| *** dickson.freenode.net sets mode: +ct-s | 00:36 | |
| *** derek_c has joined #openstack-keystone | 00:36 | |
| *** flashgordon has joined #openstack-keystone | 00:36 | |
| *** vhoward- has joined #openstack-keystone | 00:36 | |
| *** tellesnobrega has joined #openstack-keystone | 00:36 | |
| *** jamielennox|away has joined #openstack-keystone | 00:36 | |
| *** koolhead17 has joined #openstack-keystone | 00:36 | |
| *** luisbg has joined #openstack-keystone | 00:36 | |
| *** anteaya has joined #openstack-keystone | 00:36 | |
| *** dtroyer has joined #openstack-keystone | 00:36 | |
| *** morganfainberg has joined #openstack-keystone | 00:36 | |
| *** ChanServ has joined #openstack-keystone | 00:36 | |
| *** jaypipes has joined #openstack-keystone | 00:36 | |
| *** mfisch has joined #openstack-keystone | 00:36 | |
| *** chmouel has joined #openstack-keystone | 00:36 | |
| *** ayoung has joined #openstack-keystone | 00:36 | |
| *** mhu has joined #openstack-keystone | 00:36 | |
| *** huats has joined #openstack-keystone | 00:36 | |
| *** jraim has joined #openstack-keystone | 00:36 | |
| *** zigo has joined #openstack-keystone | 00:36 | |
| *** jordant has joined #openstack-keystone | 00:36 | |
| *** bknudson has joined #openstack-keystone | 00:36 | |
| *** flaper87|afk has joined #openstack-keystone | 00:36 | |
| *** haneef_ has joined #openstack-keystone | 00:36 | |
| *** lbragstad has joined #openstack-keystone | 00:36 | |
| *** mberlin has joined #openstack-keystone | 00:36 | |
| *** wchrisj has joined #openstack-keystone | 00:36 | |
| *** marekd|away has joined #openstack-keystone | 00:36 | |
| *** dvorak has joined #openstack-keystone | 00:36 | |
| *** jimbaker has joined #openstack-keystone | 00:36 | |
| *** gyee has joined #openstack-keystone | 00:36 | |
| *** amcrn has joined #openstack-keystone | 00:36 | |
| *** kfox1111 has joined #openstack-keystone | 00:36 | |
| *** sudorandom has joined #openstack-keystone | 00:36 | |
| *** YorikSar has joined #openstack-keystone | 00:36 | |
| *** rwsu_ has joined #openstack-keystone | 00:36 | |
| *** marcoemorais1 has joined #openstack-keystone | 00:36 | |
| *** harlowja has joined #openstack-keystone | 00:36 | |
| *** Daviey has joined #openstack-keystone | 00:36 | |
| *** zhiyan` has joined #openstack-keystone | 00:36 | |
| *** dolphm has joined #openstack-keystone | 00:36 | |
| *** openstackgerrit has joined #openstack-keystone | 00:36 | |
| *** thiagop_ has joined #openstack-keystone | 00:36 | |
| *** dims has joined #openstack-keystone | 00:36 | |
| *** dickson.freenode.net sets mode: +o ChanServ | 00:36 | |
| *** dickson.freenode.net changes topic to "[ Icehouse RC blockers https://launchpad.net/keystone/+milestone/icehouse-rc1 ][ Icehouse RC Target Date: March 27th, 2014 ]" | 00:36 | |
| morganfainberg | jamielennox|away, bknudson, ayoung, gyee, https://review.openstack.org/#/c/80726/ could use a 2nd +2, i'll baby sit it though and get it the +A when it's though hte basic stuff | 00:36 |
|---|---|---|
| morganfainberg | and/or fix anything else that comes up | 00:36 |
| flashgordon | thanks and have agood weekend! | 00:37 |
| morganfainberg | flashgordon, same to you | 00:37 |
| *** vhoward- has quit IRC | 00:44 | |
| *** tellesnobrega has quit IRC | 00:44 | |
| *** vhoward has joined #openstack-keystone | 00:44 | |
| *** derek_c has quit IRC | 00:48 | |
| *** tellesnobrega has joined #openstack-keystone | 00:48 | |
| *** rwsu_ has quit IRC | 00:53 | |
| ayoung | morganfainberg, ACKed | 00:55 |
| ayoung | flashgordon, +Aed | 00:56 |
| morganfainberg | ayoung, cool | 00:56 |
| *** derek_c has joined #openstack-keystone | 01:09 | |
| flashgordon | thanks guys | 01:16 |
| *** marcoemorais1 has quit IRC | 01:25 | |
| ayoung | ┻━┻ ︵ヽ(`Д´)ﾉ︵ ┻━┻ | 01:39 |
| ayoung | morganfainberg, I got it! | 01:39 |
| morganfainberg | ayoung, gratz | 01:43 |
| ayoung | morganfainberg, I still have failing tests, but I have the lazy activation working | 01:43 |
| *** openstackgerrit has quit IRC | 01:45 | |
| *** wchrisj has quit IRC | 01:49 | |
| *** mberlin has quit IRC | 02:05 | |
| ayoung | morganfainberg, OK, so it looks like the cached timeout is killing me on the unit tests | 02:09 |
| morganfainberg | ayoung, how so? | 02:10 |
| ayoung | I changed to using a KVS cached for the tree, but I need to give it no timeout for the tests | 02:10 |
| ayoung | I need to perform something that emits a revoke event, and immediately see the result | 02:10 |
| ayoung | no cache | 02:10 |
| morganfainberg | ok so when you emit the event you need to invalidate? | 02:11 |
| *** harlowja has quit IRC | 02:11 | |
| ayoung | well, not in the real world | 02:11 |
| *** harlowja has joined #openstack-keystone | 02:11 | |
| morganfainberg | i'm confused then | 02:11 |
| ayoung | in the real world, a second or two is OK | 02:11 |
| ayoung | its just for unit tests | 02:11 |
| morganfainberg | usually you want to invalidate on event | 02:11 |
| ayoung | OK...let me do that for now | 02:11 |
| morganfainberg | oh you're doing a timed cache | 02:11 |
| ayoung | can always yank if there is a performance issue | 02:12 |
| morganfainberg | yeah for now invalidate immidiately | 02:12 |
| morganfainberg | ideally you will only invalidate sometimes not on every call | 02:12 |
| morganfainberg | so the perfomance gain is still there when you cache | 02:12 |
| morganfainberg | now.... | 02:12 |
| morganfainberg | by default keystone doesn't _actually_ enable caching | 02:12 |
| morganfainberg | you have to go out of your way to do so because in-memory is icky | 02:12 |
| morganfainberg | and i don't want to "require" memcache or redis for a default install | 02:13 |
| ayoung | morganfainberg, sorry, I was coding, and dancing to the Gypsy King version of Hotel California | 02:17 |
| ayoung | as you might guess...It worked | 02:17 |
| morganfainberg | yeah | 02:17 |
| ayoung | so, it is interesting that, for testing we need to imports that trigger this pep8 error | 02:18 |
| ayoung | ./keystone/tests/core.py:433:1: F401 'oauth1' imported but unused | 02:18 |
| ayoung | I can flake8 noqa those | 02:18 |
| ayoung | but that is what registers the extension now | 02:18 |
| *** amcrn has quit IRC | 02:19 | |
| morganfainberg | would it make sense to use importutils so you fon't need to flake8 noqa? (don't mind either way) | 02:19 |
| *** dstanek_afk has joined #openstack-keystone | 02:20 | |
| *** mberlin has joined #openstack-keystone | 02:20 | |
| ayoung | actually, I don;t need them. And that kindof scares me | 02:22 |
| ayoung | I really hope everything I does here is kosher. | 02:23 |
| ayoung | morganfainberg, so now it looks like it is working in PyCharm and with run_tests.sh but not tox | 02:28 |
| *** flashgordon is now known as jogo | 02:37 | |
| ayoung | morganfainberg, so...assuming I get this patch in. Question is what to do about Horizon | 02:37 |
| morganfainberg | ayoung, getting this patch in will solve horizon's issue | 02:38 |
| ayoung | I'm tempted to say "revoking a token is neither necessary nor sufficient, so don't do it" | 02:38 |
| ayoung | yeah, but in the future... | 02:38 |
| morganfainberg | ayoung, we can work on the other part in Juno | 02:38 |
| ayoung | I can revoke a token by user and issued_at | 02:38 |
| morganfainberg | whether that is "don't revoke" or making it so we can revoke on issued_at as well | 02:38 |
| ayoung | but I don't want to | 02:38 |
| morganfainberg | i think that is a Juno topic | 02:38 |
| morganfainberg | we have a biut more breathing room on that front | 02:38 |
| ayoung | yeah. But turning off the revocations in Horizon means it can get tested sooner | 02:39 |
| morganfainberg | wont happen in Icehouse | 02:39 |
| morganfainberg | lets bring it up with them once RC is cut | 02:39 |
| ayoung | Well, I can get QA on it in house | 02:39 |
| morganfainberg | invade the channel. | 02:39 |
| morganfainberg | :) | 02:39 |
| ayoung | if I user a token t1 to create t2 and t2 to create t3, then I revoke t2, by the horizon logic, t3 would still be out there | 02:40 |
| ayoung | either re-auth, or don't revoke tokens | 02:40 |
| morganfainberg | correct | 02:40 |
| ayoung | make the expiry really short | 02:40 |
| morganfainberg | and since we've supported that model up until now, we need to evaluate if we can change that | 02:41 |
| ayoung | and the revoke need goes away | 02:41 |
| morganfainberg | i agree the revoke is silly | 02:41 |
| morganfainberg | but i don't know if we can "change" that behavior w/o calling it API incompatible | 02:41 |
| ayoung | My goal is to get kerberos and s4u2proxy working to kerberize Horizon and have horizon use a service ticket to fetch a token. THen they can revoke away to their hearts content. | 02:43 |
| morganfainberg | ++ | 02:43 |
| ayoung | x509, OTOH, does not provide a nice delegation mechanism | 02:43 |
| ayoung | Feh, Horizon is silly anyway. I mean, who does server side scripting anymore? | 02:44 |
| *** kfox1111 has quit IRC | 02:52 | |
| ayoung | morganfainberg, here it is https://review.openstack.org/#/c/80441/ | 02:54 |
| ayoung | and I am going to sign off | 02:55 |
| *** david-lyle has joined #openstack-keystone | 03:06 | |
| *** david-lyle has quit IRC | 03:10 | |
| *** david-lyle has joined #openstack-keystone | 03:13 | |
| *** harlowja is now known as harlowja_away | 03:22 | |
| *** stevemar has joined #openstack-keystone | 03:23 | |
| stevemar | jogo, oh noes, i have no idea stable/havana uses latest keystoneclient code! | 03:42 |
| stevemar | had* | 03:42 |
| *** dstanek_afk is now known as dstanek | 03:47 | |
| *** stevemar has quit IRC | 03:48 | |
| *** derek_c has quit IRC | 03:49 | |
| *** wchrisj has joined #openstack-keystone | 04:29 | |
| *** wchrisj has quit IRC | 05:02 | |
| *** zhiyan` has left #openstack-keystone | 05:43 | |
| *** zhiyan` has joined #openstack-keystone | 05:43 | |
| *** dstanek has quit IRC | 05:43 | |
| *** stevemar has joined #openstack-keystone | 05:47 | |
| *** gyee has quit IRC | 05:48 | |
| *** derek_c has joined #openstack-keystone | 05:56 | |
| *** jraim has quit IRC | 06:20 | |
| *** jraim has joined #openstack-keystone | 06:40 | |
| *** stevemar has quit IRC | 07:15 | |
| *** zhiyan` is now known as zhiyan_ | 07:30 | |
| *** saju_m has joined #openstack-keystone | 07:44 | |
| jogo | ayoung morganfainberg: one more patch to revert https://review.openstack.org/80766 | 08:09 |
| morganfainberg | jogo, another? | 08:09 |
| morganfainberg | jogo, let me take a look | 08:09 |
| jogo | second part of the BP | 08:09 |
| jogo | both patches landed around the same time | 08:10 |
| morganfainberg | jogo ahh yeah | 08:10 |
| jogo | http://logs.openstack.org/35/80435/1/check/check-tempest-dsvm-full/d061a95/logs/horizon_error.txt.gz | 08:10 |
| jogo | stack trace to back it up | 08:10 |
| morganfainberg | hmm. ayoung is out for the night | 08:11 |
| morganfainberg | i'll see who we can round up to get that in | 08:12 |
| jogo | speaking of which I should be out for the night as well | 08:12 |
| jogo | morganfainberg: thanks! | 08:12 |
| * morganfainberg goes back to playing some diablo3 | 08:12 | |
| morganfainberg | jamielennox|away, bknudson, ayoung, stevemar, dolphm, https://review.openstack.org/#/c/80766/ could use a seoncd +2/A | 08:12 |
| *** derek_c has quit IRC | 08:51 | |
| *** henrynash has joined #openstack-keystone | 08:52 | |
| morganfainberg | henrynash, ping | 08:59 |
| morganfainberg | henrynash, could use a second +2 on https://review.openstack.org/#/c/80766/ if you're around. | 09:00 |
| henrynash | morganfainberq: hi….looking | 09:00 |
| morganfainberg | henrynash, trying to help infra unbreak stable/havana gate | 09:00 |
| morganfainberg | henrynash, thanks! :) | 09:00 |
| henrynash | morganfainberg: looks like it has failed pep8 | 09:03 |
| morganfainberg | doh | 09:03 |
| henrynash | morganfainberg: but other than that, fine | 09:03 |
| morganfainberg | i'll fix the pep8 if i can | 09:06 |
| *** leseb has joined #openstack-keystone | 09:11 | |
| henrynash | morganfainberg: as an aside, when you have a moment perhaps you could look at: https://review.openstack.org/#/c/79897/3 | 09:32 |
| *** leseb has quit IRC | 09:35 | |
| *** leseb has joined #openstack-keystone | 09:43 | |
| morganfainberg | henrynash, sure i'll take a gander at it | 09:45 |
| morganfainberg | then i need to go to sleep it's 245am | 09:45 |
| henrynash | morganfainberq: understand!!!!! | 09:45 |
| morganfainberg | henrynash, +2 on that, a couple of comments, but no reasont o block the change | 09:53 |
| henrynash | morganfainberq: thx | 09:53 |
| *** saju_m has quit IRC | 09:54 | |
| henrynash | morganfainberg: as an aside, do you prefer to see the internal rules to be all clustered at the top…or placed (as I did) ahead of the block of api calls that use them? | 09:55 |
| morganfainberg | at the top | 09:55 |
| morganfainberg | personally | 09:55 |
| morganfainberg | but honestly, i don't care :) | 09:55 |
| henrynash | :-) | 09:55 |
| morganfainberg | i think it's easy to see all the rules if they're all in one place | 09:55 |
| morganfainberg | but, meh | 09:55 |
| morganfainberg | if it works | 09:56 |
| morganfainberg | henrynash, also if you have any input on https://review.openstack.org/#/c/80409/ (e.g. is this a terrible idea) that would be great | 09:59 |
| morganfainberg | henrynash, i don't expect that to go in until RC is cut though | 09:59 |
| henrynash | morganfainberg: ok, will look | 09:59 |
| morganfainberg | just changing how to run live tests | 09:59 |
| morganfainberg | jogo, +2/+A on that keystoneclient fix | 10:01 |
| morganfainberg | really? domains aren't immutable now? | 10:02 |
| morganfainberg | henrynash, does anyone _actually_ move things between domains? I really hope not | 10:03 |
| henrynash | morganfainberq: I hope not as well…..but unfortunately we kind of support it | 10:03 |
| morganfainberg | hmm. i wonder if this is a ML topic we can ask operators if they do this (explaining why it's a bad idea) and see about getting it inverted | 10:04 |
| morganfainberg | henrynash, can we add (even if it doesn't change functionality) a config option this late in the cycle? | 10:05 |
| morganfainberg | henrynash, also it looks like you have a new string change. | 10:05 |
| henrynash | morganfainberq: we could indeed…i was trying to obviously not change existing functionally for now…and then maybe for Juno we change the option to disabled by default | 10:05 |
| morganfainberg | henrynash, that one i am fairly certain we can't w/o doing an exceptionr equest | 10:06 |
| henrynash | morganfainberq: oh, you mean a string change? | 10:06 |
| morganfainberg | yeah we're past String Freeze | 10:06 |
| morganfainberg | i'm checking the string now. | 10:07 |
| henrynash | morganfainberq: oops, damn | 10:07 |
| morganfainberg | yep | 10:07 |
| morganfainberg | henrynash, line https://review.openstack.org/#/c/80769/2/keystone/common/controller.py 528 | 10:07 |
| morganfainberg | you're adding a new string in for translation | 10:08 |
| morganfainberg | henrynash, we could try and get an exception, but my gut feeling is this should land J1 | 10:08 |
| morganfainberg | henrynash, unless there is a real "it | 10:09 |
| morganfainberg | 's broken" reason to get this into I | 10:09 |
| morganfainberg | whichcase, we should request the exception :) | 10:09 |
| henrynash | morganfainberq: so I can only come up with half of the security hole….i.e. using policy.v3cloudsample, it means you can move a user into a domain you don't have a roles on | 10:10 |
| morganfainberg | oh ick :( | 10:10 |
| morganfainberg | that is worth getting an exception in i think | 10:10 |
| henrynash | morganfainberq: that's why I thought it was worth pushing it…although in theory you can't then do anything with that moved user (since you don't have a role on that domain), it feels like the a preparation step for a hole somewhere! | 10:11 |
| morganfainberg | yeah | 10:12 |
| morganfainberg | lets hit dolph up on Monday and confirm | 10:12 |
| morganfainberg | it seems benign, but... yanno | 10:12 |
| henrynash | morgainfainberq: my thoughts exactly…. | 10:12 |
| morganfainberg | actually odd thought... | 10:12 |
| morganfainberg | by moving the user to a domain they dont' control is there any way that user would gain extra rols on the new domain | 10:13 |
| henrynash | morganfainberq: (btw, i fixed the commit comment in https://review.openstack.org/#/c/79897) | 10:13 |
| morganfainberg | e.g. escalate to admin? | 10:13 |
| morganfainberg | hold on | 10:13 |
| *** saju_m has joined #openstack-keystone | 10:19 | |
| *** leseb has quit IRC | 10:27 | |
| *** morganfainberg is now known as morganfainberg_Z | 10:34 | |
| *** saju_m has quit IRC | 10:53 | |
| *** saju_m has joined #openstack-keystone | 10:59 | |
| *** leseb has joined #openstack-keystone | 11:24 | |
| *** leseb has quit IRC | 11:29 | |
| *** david-lyle has quit IRC | 12:47 | |
| *** bvandenh has joined #openstack-keystone | 13:04 | |
| *** dstanek has joined #openstack-keystone | 13:32 | |
| *** wchrisj has joined #openstack-keystone | 13:44 | |
| *** saju_m has quit IRC | 14:02 | |
| *** henrynash has quit IRC | 14:09 | |
| *** henrynash has joined #openstack-keystone | 14:50 | |
| *** openstackgerrit has joined #openstack-keystone | 14:54 | |
| *** openstackstatus has joined #openstack-keystone | 14:55 | |
| *** henrynash has quit IRC | 14:59 | |
| *** david-lyle has joined #openstack-keystone | 15:12 | |
| *** flaper87|afk is now known as flaper87 | 15:34 | |
| *** dstanek has quit IRC | 15:47 | |
| *** wchrisj has quit IRC | 16:19 | |
| *** wchrisj has joined #openstack-keystone | 17:00 | |
| *** wchrisj has quit IRC | 17:14 | |
| *** henrynash has joined #openstack-keystone | 17:20 | |
| *** wchrisj has joined #openstack-keystone | 17:26 | |
| *** wchrisj has quit IRC | 17:32 | |
| *** henrynash has quit IRC | 17:40 | |
| *** leseb has joined #openstack-keystone | 19:56 | |
| *** wchrisj has joined #openstack-keystone | 20:08 | |
| *** wchrisj has joined #openstack-keystone | 20:10 | |
| *** leseb has quit IRC | 20:14 | |
| *** wchrisj has quit IRC | 20:43 | |
| *** thedodd has joined #openstack-keystone | 21:04 | |
| *** leseb has joined #openstack-keystone | 21:25 | |
| *** tellesnobrega has quit IRC | 21:26 | |
| *** leseb has quit IRC | 21:30 | |
| *** stevemar has joined #openstack-keystone | 21:41 | |
| *** dstanek has joined #openstack-keystone | 21:41 | |
| *** tellesnobrega has joined #openstack-keystone | 21:45 | |
| *** stevemar has quit IRC | 21:50 | |
| *** leseb has joined #openstack-keystone | 21:59 | |
| *** morganfainberg_Z is now known as morganfainberg | 22:13 | |
| *** thedodd has quit IRC | 22:20 | |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/keystone: Don't automatically enable revocation events. https://review.openstack.org/80441 | 22:44 |
| openstackgerrit | Morgan Fainberg proposed a change to openstack/keystone: Don't automatically enable revocation events. https://review.openstack.org/80441 | 22:47 |
| morganfainberg | ayoung, proposed a fix to address the comments i had about your patch. | 22:49 |
| ayoung | morganfainberg, that "catch" is what happens when you actyally disable the extension | 22:54 |
| morganfainberg | ayoung, yeah i know | 22:54 |
| morganfainberg | ayoung, ++ | 22:54 |
| ayoung | morganfainberg, why didn't you change dict() to {}? | 22:55 |
| ayoung | I thought {} meant something else...or am I getting confused | 22:56 |
| morganfainberg | {} is a literal dict | 22:56 |
| ayoung | there was one python collection where initializing with the empty was not the same as initializing with a value | 22:56 |
| morganfainberg | in py27 {key: value} is a dict, {key, key, key} is a set | 22:56 |
| ayoung | is it []? or () ? | 22:56 |
| morganfainberg | () | 22:56 |
| ayoung | ah, right | 22:56 |
| morganfainberg | if you do (blah) that is just blah | 22:56 |
| ayoung | so I thought {} was an empty set, not an empty dict | 22:57 |
| morganfainberg | if you do (blah,) | 22:57 |
| morganfainberg | that is the tuple | 22:57 |
| morganfainberg | yeah, imo never use the set literal | 22:57 |
| morganfainberg | sets are used far less frequently, so set() is better | 22:57 |
| ayoung | which is why I did dict() | 22:57 |
| ayoung | no question | 22:57 |
| morganfainberg | and you can't initilaize an empty set (afaict) with a literal | 22:57 |
| morganfainberg | ayoung, in short, i forgot | 22:57 |
| ayoung | and it should be done once at startup | 22:57 |
| morganfainberg | it was such a nit, i didn't care too much | 22:57 |
| ayoung | go for it if it is important to you | 22:58 |
| morganfainberg | nah | 22:58 |
| morganfainberg | not worth the bits on the wire :P | 22:58 |
| morganfainberg | we've already over-talked it | 22:58 |
| morganfainberg | it was more of a future looking thing. I'd never -1 for that unless someone made a dict() 100s of times | 22:58 |
| morganfainberg | so, i wanted to run an interesting thought by you. | 22:59 |
| ayoung | fire away | 23:00 |
| morganfainberg | when it comes to a design, does it make sense to ensure everything that goes back to the driver (there are exceptions but not the majority) are always args vs kwargs? | 23:00 |
| ayoung | I kindof hate kwargs | 23:00 |
| morganfainberg | any optional argument stuff happens at the business logic layer | 23:00 |
| morganfainberg | and that changes interactions | 23:01 |
| ayoung | I think we abuse them | 23:01 |
| morganfainberg | rather than the driver being smart | 23:01 |
| morganfainberg | and knowing all the variations | 23:01 |
| ayoung | ++ | 23:01 |
| morganfainberg | and i mean from a pythonic standpoint less so from a keystone or openstack (we'd inherit it) | 23:01 |
| ayoung | anything that could be done in the manager should be done in the manager | 23:01 |
| ayoung | instead of the driver | 23:01 |
| morganfainberg | it's been bothering me how much kwargs is abused across python | 23:02 |
| morganfainberg | i just don't know if i'm the one taking crazy pills ;) | 23:02 |
| ayoung | I was thinking about that. Inversion of control and dependency injection in particular says "I need something that meets this contract: blah" but without type safety....how dso you specify the contract? | 23:02 |
| ayoung | I mean, all you get is the names of the parameters to a constructor | 23:03 |
| morganfainberg | ayoung, yeah. | 23:03 |
| morganfainberg | ayoung, welcome to duck-type hell | 23:03 |
| ayoung | and, with the .requires appraoch we don't even really get that | 23:03 |
| ayoung | if it smells like brimstone, and burns like brimstone.... | 23:03 |
| morganfainberg | the way we handle this is we make everything a type-defined object | 23:03 |
| morganfainberg | something protobuf like | 23:03 |
| morganfainberg | if you're doing user things, you have a user object | 23:03 |
| morganfainberg | with a defined schema | 23:04 |
| ayoung | so then how do you specify that a function only accepts an object of that type? | 23:04 |
| morganfainberg | i think that becomes "at the entry point" e.g. public interface we validate | 23:04 |
| morganfainberg | probably w/ a decorator | 23:04 |
| ayoung | protobuf? smells like IDL to me | 23:04 |
| morganfainberg | @validate_arg(type) | 23:05 |
| ayoung | yeah, I'm guessing thatis the best we are going to get | 23:05 |
| morganfainberg | protobuf is probably not the right solution, but it could be a solution | 23:05 |
| ayoung | but...I don;t know that I really like even that. It only works for code we write ourselves, and ideally DI is done with anyone's code | 23:05 |
| ayoung | OK...back to the family | 23:05 |
| ayoung | thanks for looking at that patch | 23:06 |
| morganfainberg | yeah. | 23:06 |
| morganfainberg | have a good weekend | 23:06 |
| ayoung | we can noodge people about it on Monday | 23:06 |
| *** ayoung is now known as ayoung-zzz | 23:06 | |
| *** flaper87 is now known as flaper87|afk | 23:38 | |
| *** stevemar has joined #openstack-keystone | 23:46 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!