| *** ngupta has joined #openstack-keystone | 00:00 | |
| *** dims has quit IRC | 00:11 | |
| *** shoutm has joined #openstack-keystone | 00:14 | |
| *** shoutm has quit IRC | 00:14 | |
| *** ngupta has quit IRC | 00:23 | |
| *** ngupta has joined #openstack-keystone | 00:23 | |
| *** ngupta has quit IRC | 00:28 | |
| *** su_zhang has joined #openstack-keystone | 00:33 | |
| *** dims has joined #openstack-keystone | 00:39 | |
| ayoung | something just broke. Running tox against master: keystone.tests.unit.test_token_provider.TestPKIProviderWithStdlib.test_get_token_id_error_handling fails | 00:39 |
|---|---|---|
| stevemar | ayoung: yep | 00:40 |
| ayoung | http://fpaste.org/314145/45359601/ | 00:40 |
| stevemar | ayoung: i opened a bug about it this morning https://bugs.launchpad.net/keystone/+bug/1537401 | 00:40 |
| openstack | Launchpad bug 1537401 in OpenStack Identity (keystone) "keystone gate failing test_get_token_id_error_handling" [Undecided,New] | 00:40 |
| stevemar | eventlet broke it | 00:40 |
| stevemar | ayoung: subscribe to keystone bugs :O | 00:41 |
| ayoung | stevemar, I see that...giving it a test now | 00:41 |
| stevemar | ayoung: dstanek wrote "Eventlet 0.18.0 was released to PyPI today and seems to be what broke it. When I constrained by environment to installing 0.17.4 things were fine" | 00:41 |
| ayoung | stevemar, I have them going to a subfolder...I was unable to keep up with the flow of email from OpenStack and something had to give | 00:41 |
| stevemar | ah | 00:41 |
| *** topol_ is now known as topol | 00:43 | |
| *** ChanServ sets mode: +v topol | 00:43 | |
| stevemar | ayoung: fwiw, looks like henry and notmorgan pushed my skip through: https://review.openstack.org/#/c/271723/ | 00:44 |
| stevemar | logging off for a bit | 00:45 |
| *** browne has quit IRC | 01:06 | |
| *** EinstCrazy has quit IRC | 01:14 | |
| *** EinstCrazy has joined #openstack-keystone | 01:15 | |
| *** EinstCrazy has quit IRC | 01:18 | |
| ayoung | you can tell whenever I've been working in another language. I got back to using " for all strings | 01:23 |
| *** shoutm has joined #openstack-keystone | 01:24 | |
| ayoung | stevemar, yeah. I saw that, but I think we'll need to use dstanek 's approach of constraining eventlet to the earlier version or we will break people using PKI. | 01:29 |
| * dstanek glares at eventlet | 01:29 | |
| openstackgerrit | Merged openstack/keystone: skip test_get_token_id_error_handling to get gate passing https://review.openstack.org/271723 | 01:38 |
| ayoung | dstanek, guess what? We are getting eventlet out of Tripleo as we speak. One less place we'll have to deal with, and that is a big one | 01:39 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone: Updating sample configuration file https://review.openstack.org/269479 | 01:40 |
| openstackgerrit | ayoung proposed openstack/keystone: Implied Roles API https://review.openstack.org/242614 | 01:40 |
| *** EinstCrazy has joined #openstack-keystone | 01:53 | |
| *** agireud has quit IRC | 01:55 | |
| *** agireud has joined #openstack-keystone | 01:57 | |
| notmorgan | ayoung: basically unbreak gate until requirements fix can land | 02:06 |
| ayoung | notmorgan, yep...works just fine | 02:07 |
| ayoung | and I am not super worried about PKI people chasing top of tree | 02:07 |
| notmorgan | hah | 02:08 |
| notmorgan | true enough | 02:08 |
| notmorgan | had a limequat today. Holy crap tart. | 02:08 |
| notmorgan | painfully so.. like sour candy... but even better | 02:09 |
| *** EinstCrazy has quit IRC | 02:14 | |
| *** lhcheng has joined #openstack-keystone | 02:42 | |
| *** ChanServ sets mode: +v lhcheng | 02:42 | |
| *** EinstCrazy has joined #openstack-keystone | 02:45 | |
| *** phalmos has joined #openstack-keystone | 02:47 | |
| *** markvoelker has joined #openstack-keystone | 03:00 | |
| *** EinstCrazy has quit IRC | 03:01 | |
| *** EinstCrazy has joined #openstack-keystone | 03:05 | |
| *** markvoelker has quit IRC | 03:07 | |
| *** phalmos has quit IRC | 03:12 | |
| *** lhcheng has quit IRC | 03:12 | |
| *** phalmos has joined #openstack-keystone | 03:13 | |
| *** phalmos has quit IRC | 03:16 | |
| *** EinstCrazy has quit IRC | 03:29 | |
| *** EinstCrazy has joined #openstack-keystone | 03:32 | |
| *** EinstCrazy has quit IRC | 03:38 | |
| *** EinstCrazy has joined #openstack-keystone | 03:43 | |
| *** EinstCrazy has quit IRC | 03:47 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone: Updated from global requirements https://review.openstack.org/271642 | 03:58 |
| *** lhcheng has joined #openstack-keystone | 04:01 | |
| *** ChanServ sets mode: +v lhcheng | 04:01 | |
| *** markvoelker has joined #openstack-keystone | 04:03 | |
| *** markvoelker has quit IRC | 04:07 | |
| openstackgerrit | Merged openstack/keystone: Remove keystone/common/cache/_memcache_pool.py https://review.openstack.org/271485 | 04:27 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone: Updating sample configuration file https://review.openstack.org/269479 | 04:29 |
| *** dims has quit IRC | 04:42 | |
| *** alex_xu has quit IRC | 04:54 | |
| *** alex_xu has joined #openstack-keystone | 04:55 | |
| *** fawadkhaliq has joined #openstack-keystone | 04:58 | |
| *** markvoelker has joined #openstack-keystone | 05:04 | |
| *** alex_xu has quit IRC | 05:04 | |
| *** alex_xu has joined #openstack-keystone | 05:05 | |
| *** markvoelker has quit IRC | 05:09 | |
| openstackgerrit | Merged openstack/keystone: Updated from global requirements https://review.openstack.org/271642 | 06:08 |
| *** fawadkhaliq has quit IRC | 06:28 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone: Updating sample configuration file https://review.openstack.org/269479 | 06:37 |
| *** davechen has joined #openstack-keystone | 06:44 | |
| *** lhcheng has quit IRC | 06:54 | |
| *** markvoelker has joined #openstack-keystone | 07:05 | |
| *** shoutm has quit IRC | 07:09 | |
| *** henrynash has joined #openstack-keystone | 07:38 | |
| *** ChanServ sets mode: +v henrynash | 07:38 | |
| *** markvoelker has quit IRC | 07:39 | |
| *** ajayaa has joined #openstack-keystone | 07:41 | |
| *** rcernin has joined #openstack-keystone | 08:23 | |
| *** su_zhang has quit IRC | 08:33 | |
| *** rcernin_ has joined #openstack-keystone | 08:36 | |
| *** rcernin_ has quit IRC | 08:36 | |
| *** lhcheng has joined #openstack-keystone | 08:43 | |
| *** ChanServ sets mode: +v lhcheng | 08:43 | |
| *** lhcheng has quit IRC | 08:48 | |
| *** hogepodge has quit IRC | 09:32 | |
| *** davechen has left #openstack-keystone | 10:52 | |
| *** markvoelker has joined #openstack-keystone | 11:36 | |
| *** markvoelker has quit IRC | 11:41 | |
| *** fawadkhaliq has joined #openstack-keystone | 12:03 | |
| *** fawadkhaliq has quit IRC | 12:25 | |
| *** alex_xu has quit IRC | 12:42 | |
| *** alex_xu has joined #openstack-keystone | 12:43 | |
| *** EinstCrazy has joined #openstack-keystone | 13:01 | |
| *** dims has joined #openstack-keystone | 13:11 | |
| *** brad[] has quit IRC | 13:18 | |
| *** markvoelker has joined #openstack-keystone | 13:37 | |
| *** markvoelker has quit IRC | 13:41 | |
| *** lhcheng has joined #openstack-keystone | 14:09 | |
| *** ChanServ sets mode: +v lhcheng | 14:09 | |
| *** lhcheng has quit IRC | 14:14 | |
| *** ayoung has quit IRC | 14:18 | |
| *** jasondotstar_ has joined #openstack-keystone | 14:22 | |
| *** jasondotstar has quit IRC | 14:23 | |
| *** jasondotstar_ is now known as jasondotstar | 14:23 | |
| *** rdo has quit IRC | 14:25 | |
| *** rdo has joined #openstack-keystone | 14:27 | |
| openstackgerrit | Brant Knudson proposed openstack/keystone: keystonemiddleware is not a requirement https://review.openstack.org/271480 | 14:31 |
| *** rcernin has quit IRC | 14:40 | |
| *** phalmos has joined #openstack-keystone | 14:47 | |
| *** shoutm has joined #openstack-keystone | 14:49 | |
| *** shoutm has quit IRC | 14:51 | |
| *** rcernin has joined #openstack-keystone | 15:11 | |
| *** huats_ has quit IRC | 15:18 | |
| *** phalmos has quit IRC | 15:24 | |
| *** EinstCrazy has quit IRC | 15:27 | |
| *** ayoung has joined #openstack-keystone | 15:30 | |
| *** ChanServ sets mode: +v ayoung | 15:30 | |
| *** markvoelker has joined #openstack-keystone | 15:37 | |
| *** markvoelker has quit IRC | 15:42 | |
| *** su_zhang has joined #openstack-keystone | 15:44 | |
| *** jsavak has joined #openstack-keystone | 15:52 | |
| *** lhcheng has joined #openstack-keystone | 15:55 | |
| *** ChanServ sets mode: +v lhcheng | 15:55 | |
| *** fawadkhaliq has joined #openstack-keystone | 16:14 | |
| *** jsavak has quit IRC | 16:32 | |
| *** jsavak has joined #openstack-keystone | 16:33 | |
| *** EinstCrazy has joined #openstack-keystone | 16:33 | |
| *** jsavak has quit IRC | 16:37 | |
| *** fawadkhaliq has quit IRC | 16:40 | |
| *** fawadkhaliq has joined #openstack-keystone | 16:40 | |
| *** EinstCrazy has quit IRC | 16:41 | |
| *** lhcheng has quit IRC | 16:43 | |
| *** ayoung has quit IRC | 16:43 | |
| *** dims has quit IRC | 17:05 | |
| *** fawadkhaliq has quit IRC | 17:29 | |
| openstackgerrit | henry-nash proposed openstack/keystone: Allow project domain_id to be nullable at the manager level https://review.openstack.org/264533 | 17:38 |
| *** markvoelker has joined #openstack-keystone | 17:38 | |
| *** markvoelker has quit IRC | 17:43 | |
| openstackgerrit | henry-nash proposed openstack/keystone: Verify project unique constraints for projects acting as domains https://review.openstack.org/158372 | 17:43 |
| *** EinstCrazy has joined #openstack-keystone | 17:56 | |
| *** EinstCrazy has quit IRC | 18:02 | |
| *** huats_ has joined #openstack-keystone | 18:17 | |
| *** huats_ has quit IRC | 18:17 | |
| *** huats_ has joined #openstack-keystone | 18:17 | |
| *** phalmos has joined #openstack-keystone | 18:44 | |
| *** phalmos has quit IRC | 18:59 | |
| *** ayoung has joined #openstack-keystone | 19:01 | |
| *** ChanServ sets mode: +v ayoung | 19:01 | |
| ayoung | not sure if the others went through: jamielennox|away gyee, please look at implie roles api when you have a chance: https://review.openstack.org/#/c/242614/ | 19:06 |
| *** lhcheng has joined #openstack-keystone | 19:09 | |
| *** ChanServ sets mode: +v lhcheng | 19:09 | |
| *** phalmos has joined #openstack-keystone | 19:14 | |
| *** ajayaa has quit IRC | 19:18 | |
| openstackgerrit | henry-nash proposed openstack/keystone: Add tests for role management with v3policy file https://review.openstack.org/261846 | 19:22 |
| openstackgerrit | henry-nash proposed openstack/keystone: Add CRUD support for domain specific roles https://review.openstack.org/261870 | 19:23 |
| openstackgerrit | henry-nash proposed openstack/keystone: Modify rules in the v3 policy sample for domain specifc roles https://review.openstack.org/262078 | 19:25 |
| openstackgerrit | henry-nash proposed openstack/keystone: Modify implied roles to honor domain specific roles https://review.openstack.org/263064 | 19:26 |
| openstackgerrit | henry-nash proposed openstack/keystone: Modify rules for domain specific role assignments https://review.openstack.org/263549 | 19:27 |
| *** su_zhang has quit IRC | 19:31 | |
| *** rcernin has quit IRC | 19:37 | |
| *** markvoelker has joined #openstack-keystone | 19:39 | |
| *** markvoelker has quit IRC | 19:43 | |
| *** lhcheng has quit IRC | 19:44 | |
| openstackgerrit | henry-nash proposed openstack/keystone: Projects acting as domains https://review.openstack.org/231289 | 19:50 |
| *** henrynash has quit IRC | 19:55 | |
| *** henrynash has joined #openstack-keystone | 19:59 | |
| *** ChanServ sets mode: +v henrynash | 19:59 | |
| *** dims has joined #openstack-keystone | 20:00 | |
| *** dims has quit IRC | 20:05 | |
| *** su_zhang has joined #openstack-keystone | 20:05 | |
| *** henrynash has quit IRC | 20:06 | |
| *** brad[] has joined #openstack-keystone | 20:09 | |
| *** henrynash has joined #openstack-keystone | 20:09 | |
| *** ChanServ sets mode: +v henrynash | 20:09 | |
| *** dims has joined #openstack-keystone | 20:22 | |
| *** simondodsley has joined #openstack-keystone | 20:23 | |
| *** EinstCrazy has joined #openstack-keystone | 20:38 | |
| *** markvoelker has joined #openstack-keystone | 20:40 | |
| *** EinstCrazy has quit IRC | 20:42 | |
| *** markvoelker has quit IRC | 20:44 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/keystone: Updated from global requirements https://review.openstack.org/271831 | 20:49 |
| *** dslev has joined #openstack-keystone | 20:50 | |
| *** dslev_ has joined #openstack-keystone | 20:50 | |
| *** dslev_ has quit IRC | 20:50 | |
| openstackgerrit | henry-nash proposed openstack/keystone: Projects acting as domains https://review.openstack.org/231289 | 21:02 |
| *** su_zhang has quit IRC | 21:11 | |
| *** dims has quit IRC | 21:17 | |
| *** phalmos has quit IRC | 21:28 | |
| *** dims has joined #openstack-keystone | 21:38 | |
| *** rcernin has joined #openstack-keystone | 22:01 | |
| *** amit213 has quit IRC | 22:09 | |
| *** amit213 has joined #openstack-keystone | 22:10 | |
| *** mattt has left #openstack-keystone | 22:12 | |
| *** dims has quit IRC | 22:13 | |
| *** jamielennox|away is now known as jamielennox | 22:27 | |
| *** dims has joined #openstack-keystone | 22:30 | |
| openstackgerrit | Jamie Lennox proposed openstack/keystoneauth: Use positional library instead of our own copy https://review.openstack.org/267300 | 22:32 |
| jamielennox | notmorgan: i had ^ from a while ago as well | 22:34 |
| notmorgan | jamielennox: heh. | 22:36 |
| notmorgan | sure. i was aiming for "use it" then play cleanup | 22:37 |
| notmorgan | just so we didn't make anything grumpy, but for KSA it's not as importnat | 22:37 |
| notmorgan | abandoned mine | 22:38 |
| *** lhcheng has joined #openstack-keystone | 22:38 | |
| *** ChanServ sets mode: +v lhcheng | 22:38 | |
| *** EinstCrazy has joined #openstack-keystone | 22:39 | |
| *** markvoelker has joined #openstack-keystone | 22:40 | |
| *** markvoelker has quit IRC | 22:45 | |
| *** LukeH has joined #openstack-keystone | 22:47 | |
| *** LukeH has quit IRC | 22:48 | |
| *** dims has quit IRC | 22:50 | |
| jamielennox | notmorgan: yea, for ksc we need to keep it for a while because of auth_token, but ksa i think we just replace it | 22:55 |
| openstackgerrit | Jamie Lennox proposed openstack/python-keystoneclient: Replace TestResponse with requests_mock https://review.openstack.org/258741 | 22:57 |
| notmorgan | right | 22:57 |
| notmorgan | we should also propose the fix to ATM for future / next release | 22:57 |
| notmorgan | since positional is now "in" | 22:57 |
| notmorgan | also, yay we can use it in Keystone | 22:57 |
| notmorgan | i want to start making all the cachable interfaces use positional to prevent kwargs | 22:57 |
| notmorgan | actually i have an idea for kwarg keygen first | 22:58 |
| notmorgan | now that i think about it | 22:58 |
| jamielennox | notmorgan: is there an oslo deprecations thing or is it just debtcollector? | 23:01 |
| notmorgan | debtcollector | 23:01 |
| notmorgan | oooor the versionutils thing | 23:01 |
| notmorgan | but debtcollector is the "right" choice iirc | 23:01 |
| *** rcernin has quit IRC | 23:01 | |
| openstackgerrit | Steve Martinelli proposed openstack/keystone: Revert "skip test_get_token_id_error_handling to get gate passing" https://review.openstack.org/271851 | 23:03 |
| notmorgan | stevemar: i assume this means the requirements pin landed? | 23:03 |
| jamielennox | notmorgan: so the versionutils thing was mentioned in one of my reviews, but i don't know it | 23:04 |
| notmorgan | we use versionutils all over | 23:04 |
| jamielennox | i wasn't sure which was supposed to be used | 23:04 |
| notmorgan | it just logs | 23:04 |
| jamielennox | client? | 23:04 |
| notmorgan | debtcollector uses warning | 23:04 |
| notmorgan | i'd use debtcollector if you can make it work | 23:04 |
| notmorgan | but versionutils is fine imo otherwise | 23:04 |
| stevemar | notmorgan: no, new eventlet version | 23:10 |
| notmorgan | still will want a pin to blacklist the broken version before we undo that | 23:12 |
| notmorgan | fwiw | 23:12 |
| notmorgan | s/pin/exclude | 23:12 |
| *** EinstCrazy has quit IRC | 23:13 | |
| stevemar | notmorgan: sure, i was gonna wait and see if jenkins passes | 23:13 |
| stevemar | notmorgan: we also need tor release keystonemiddleware with the caching stuff brought back in | 23:13 |
| stevemar | forgot about that, lemme do that | 23:13 |
| openstackgerrit | henry-nash proposed openstack/keystone: Allow project domain_id to be nullable at the manager level https://review.openstack.org/264533 | 23:16 |
| notmorgan | jamielennox: so, cache_key generator. i think this is doable [check my logic]: we already have the argspec to the FN, and we can determine default values by doing a reverse ordering of the argspec and lookin at defaults, so deepcopy(argspec.args) then .reverse() it. we then store it, when we get a *args **kwargs passed in we look for the args and kwargs and populate the defaults in our | 23:16 |
| notmorgan | key-generator-list. aphabatize the list and we have a consistent key | 23:16 |
| notmorgan | jamielennox: i'm thinking of proposing this to dogpile upstream. so let me draw up pseudo code. | 23:16 |
| notmorgan | stevemar: ^ if you want to weigh in too | 23:16 |
| notmorgan | stevemar: also i think sdague proposed a blacklist to requirements | 23:16 |
| notmorgan | since it was also affecting nova | 23:16 |
| jamielennox | notmorgan: your going to have to backtrack a little and tell me what you're trying to solve | 23:16 |
| notmorgan | jamielennox: dogpile does not handle kwarg caching | 23:17 |
| * jamielennox has never used @MEMOIZE | 23:18 | |
| notmorgan | jamielennox: because the key_fn_generator can't determine the values since it is a decorator, so *args, **kwargs miss defaults | 23:18 |
| notmorgan | also kwargs are not-order specific | 23:18 |
| notmorgan | so.. they just punted on handling kwargs | 23:18 |
| notmorgan | https://bitbucket.org/zzzeek/dogpile.cache/src/c6913eb143b24b4a886124ff0da5c935ea34e3ac/dogpile/cache/util.py?at=master&fileviewer=file-view-default#util.py-76:85 | 23:18 |
| jamielennox | ok, then yes you can absolutely use inspect to get the default values of a function, sort the dict and use that as part of your key | 23:19 |
| notmorgan | they already use inspec | 23:19 |
| notmorgan | t | 23:19 |
| jamielennox | that works so long as you have defaults, it gets risky when taking **kwargs | 23:19 |
| notmorgan | but they still punt because cache keys need to be absolutely consistent | 23:19 |
| notmorgan | right. now. since we know the underlying function will explode if it gets unknown values | 23:20 |
| notmorgan | we can just populate a list for non-passed kwargs, and do an alphabatize of the kwargs [or even all args] | 23:20 |
| notmorgan | to generate the key | 23:20 |
| notmorgan | that way we don't care about the ordering | 23:20 |
| notmorgan | user can do func(a=1,c=2,b=3) and we'll still end up with func|a|b|c | 23:21 |
| stevemar | notmorgan: https://review.openstack.org/#/c/271853/ | 23:21 |
| notmorgan | the only issue is we break compat with currently populated cache-keys | 23:21 |
| notmorgan | i guess we could just do the alpha work if **kwargs is populated | 23:21 |
| notmorgan | so.. we only use the new code path that way | 23:22 |
| notmorgan | so, func(a, c, b=12) would be func|a|c|b=12 | 23:22 |
| jamielennox | notmorgan: so i would expect a user to specify which args are important | 23:22 |
| notmorgan | but func(a, c=10, b=12) would be func|a|b=12|c=10 | 23:22 |
| jamielennox | (no real knowledge of dogpile) | 23:23 |
| notmorgan | ugh, we always need to alpha though. | 23:23 |
| jamielennox | so i would expect so say | 23:23 |
| jamielennox | @MEMOIZE(keys=['a', 'b', 'c']) | 23:23 |
| notmorgan | otherwise we have different cache keys between func(a, b, c) and func(a=a, b=b, c=c) | 23:23 |
| jamielennox | so that you can say that those are the args that i care about | 23:23 |
| jamielennox | ie, d=4 is not important to the cache | 23:23 |
| notmorgan | jamielennox: sure, that is in our code. i'm trying to avoid carrying that | 23:23 |
| notmorgan | a generic way would be nice. | 23:23 |
| notmorgan | and this is pretty deep in the dogpile code fwiw | 23:24 |
| jamielennox | you could to keys=dogpile.ALL or something | 23:24 |
| *** dslev has quit IRC | 23:24 | |
| jamielennox | but i'm not sure i want too much magic in this | 23:24 |
| notmorgan | i could write a simple keygen for oslo.cache that does this but i know this has been a general complaint in dogpile.cache | 23:24 |
| notmorgan | "why can't i handle kwargs" | 23:24 |
| notmorgan | i guess i could just make it a new function_key_generator in utils and let people opt into it | 23:24 |
| jamielennox | with keys=[] you should be able to | 23:24 |
| notmorgan | then openstack can default | 23:24 |
| jamielennox | keys doesn't care if it's arg or kwargs | 23:25 |
| jamielennox | shouldn't - again i'm coming from theory | 23:25 |
| notmorgan | sure, but in 99% of the cases we never want to not capture all args | 23:25 |
| jamielennox | sure so you can default to an ALL | 23:25 |
| notmorgan | becuase otherwise we have a cachekey that is duplicated between (a=1, b=2, c=3, d=10) and (a=1, b=2, c=3, d=12) | 23:26 |
| notmorgan | i'm thinking we don't really have a solid case where that is normal in openstack | 23:26 |
| jamielennox | again - not always a problem | 23:26 |
| jamielennox | log=False | 23:26 |
| jamielennox | umm | 23:26 |
| jamielennox | debug=XX | 23:26 |
| jamielennox | anything that doesn't actually impact what is calculated | 23:26 |
| notmorgan | that is pretty far down and often not part of the memoization | 23:27 |
| notmorgan | maybe i'd rather do an @memoize(ignore_keys=[list]) | 23:27 |
| jamielennox | again, i haven't used it like that, in keystone i trust you did a good job, in auth_token i want/have to use the object directly | 23:27 |
| jamielennox | notmorgan: either way works, you can do both if you put some sanity checking in | 23:28 |
| notmorgan | you're not going to @memoize in auth_token | 23:28 |
| *** rcernin has joined #openstack-keystone | 23:28 | |
| notmorgan | in almost any case | 23:28 |
| jamielennox | i can't | 23:28 |
| notmorgan | well there are a couple cases we could | 23:28 |
| notmorgan | but they are all pure calculation stuff that isn't worth it | 23:28 |
| jamielennox | fetching tokens from keystone | 23:28 |
| notmorgan | i mean, we *could* @memoize the whole fetch. and rehydrate the object as though it came from keystone | 23:29 |
| notmorgan | but ... that is something i'd rather move towards down the line if we want rather than start there | 23:29 |
| notmorgan | since invalidation is much much harder with @memoize | 23:29 |
| jamielennox | notmorgan: that's what memcache is doing in auth_token now | 23:29 |
| notmorgan | right you're using the memcache-like interface | 23:30 |
| notmorgan | the right answer initially | 23:30 |
| jamielennox | but i can't use memoize because i don't have a full CONF object to construct the memoize globally | 23:30 |
| notmorgan | jamielennox: oh i've solved that before | 23:30 |
| notmorgan | we can lazy config by passing a region into our decorator | 23:30 |
| notmorgan | it's a little more expensive and much harder to follow | 23:31 |
| jamielennox | notmorgan: i don't particularly want to use memoize, just saying i can't | 23:31 |
| notmorgan | but i don't see a huge win | 23:31 |
| notmorgan | basically we do a lazy region.cache_on_arguments(func)(thing) | 23:31 |
| notmorgan | it's ugly | 23:31 |
| jamielennox | anyway, i'd default to keys=ALL | 23:31 |
| jamielennox | if you do ignore_keys then inspect and do all but those | 23:31 |
| notmorgan | right if ignore_keys=[] then we can just use keys=all | 23:32 |
| jamielennox | and some sanity checking that keys and ignore_keys aren't both passed | 23:32 |
| notmorgan | like i said i think i have a solution here for generic kwarg handling if so, i'm going to push to make this the default in openstack (oslo.cache) | 23:32 |
| *** gildub has joined #openstack-keystone | 23:33 | |
| notmorgan | the @memoize stuff is very much part of keystone's code base still | 23:33 |
| notmorgan | it does some behind the scenes offloading of config etc | 23:33 |
| *** hogepodge has joined #openstack-keystone | 23:33 | |
| notmorgan | oslo.cache is a much rawer interface but still not as raw as dogpile.cache directly | 23:33 |
| *** hogepodge has quit IRC | 23:38 | |
| *** markvoelker has joined #openstack-keystone | 23:41 | |
| *** henrynash has quit IRC | 23:43 | |
| *** markvoelker has quit IRC | 23:46 | |
| openstackgerrit | Merged openstack/keystonemiddleware: Group common PKI validation code - Tests https://review.openstack.org/268556 | 23:54 |
| *** dims has joined #openstack-keystone | 23:54 | |
| openstackgerrit | Merged openstack/keystonemiddleware: Group common PKI validation code - Refactor https://review.openstack.org/268557 | 23:54 |
| openstackgerrit | Merged openstack/keystone: Updated from global requirements https://review.openstack.org/271831 | 23:58 |
| jamielennox | notmorgan: interesting https://review.openstack.org/#/c/271357/ | 23:58 |
| notmorgan | i don't agree with that in the slightest | 23:59 |
| notmorgan | but that is a result of an issue that i am fixing here: https://review.openstack.org/#/c/271536/ | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!