| *** NM has joined #swift3 | 00:49 | |
| *** NM has quit IRC | 01:21 | |
| *** [-mugatu-] has quit IRC | 05:41 | |
| kota_ | oh, new swift3 patch for ProxyFS!! | 06:03 |
|---|---|---|
| kota_ | timburke, notmyname, tdasilva: I updated my thought on 4 top patches at https://etherpad.openstack.org/p/s3api-middleware. And also I reviewed them. I think one patch is close to land and I like to get it into the release. Others seem not be rushed and they don't block the release. | 10:51 |
| kota_ | timburke, notmyname, tdasilva: if you have any opinions and ideas, please let me know. Thanks. | 10:51 |
| kota_ | oops, this is swift3 channel. ah... it's still correct channnel :P | 10:52 |
| *** NM has joined #swift3 | 12:35 | |
| *** remix_tj has joined #swift3 | 13:08 | |
| *** remix_tj has left #swift3 | 13:09 | |
| *** Guest36 has joined #swift3 | 14:13 | |
| *** [-mugatu-] has joined #swift3 | 15:09 | |
| timburke | thanks kota_! | 16:28 |
| *** NM has quit IRC | 16:32 | |
| *** NM has joined #swift3 | 16:36 | |
| *** NM has quit IRC | 16:41 | |
| *** Guest98091 has joined #swift3 | 16:44 | |
| *** Guest98091 has quit IRC | 16:58 | |
| *** Guest36 has quit IRC | 17:14 | |
| *** Loiuek has joined #swift3 | 17:35 | |
| *** Loiuek has quit IRC | 17:39 | |
| *** Loiuek has joined #swift3 | 17:47 | |
| *** Loiuek has quit IRC | 18:10 | |
| *** NM has joined #swift3 | 18:15 | |
| *** chsc has joined #swift3 | 18:34 | |
| *** chsc has quit IRC | 18:34 | |
| *** chsc has joined #swift3 | 18:34 | |
| openstackgerrit | Samuel Merritt proposed openstack/swift3 master: Make multipart uploads compatible with ProxyFS https://review.openstack.org/508633 | 20:57 |
| timburke | torgomatic: i'm still torn. i could actually see legitimate reasons for swift3 to want to guess content types *differently* from stock swift | 21:00 |
| torgomatic | timburke: If we want to do something different, that's fine by me. I just wanted to avoid reimplementing any more bits of the proxy than I had to. | 21:01 |
| timburke | (speaking of -- do we have any guarantees about what mimetypes will return? doesn't that depend on the contents of /etc/mime.types or something?) | 21:01 |
| timburke | yeah, that's fair too... honestly, i haven't played around with AWS's content type detection much, so i don't actually know if we *need* to do anything different... | 21:02 |
| torgomatic | timburke: mimetypes.py has a list of known files it'll try, and it loads the database from there; /etc/mime.types is one of the places it'll look (the first one, FWIW) | 21:03 |
| timburke | and it *would* suck to have to play catch-up in dropping some code to use python-magic or something... | 21:04 |
| torgomatic | yup, that's my thought too | 21:04 |
| torgomatic | if swift3 were in the Swift repo, this'd be really easy; pull the type-guessing out to a utility function, call it from both places, and be done with it | 21:07 |
| timburke | makes me think we should move some portion of https://github.com/openstack/swift/blob/2.15.1/swift/proxy/controllers/obj.py#L447-L456 to utils and stop repeating it in https://github.com/openstack/swift/blob/2.15.1/swift/common/middleware/slo.py#L1056-L1058 | 21:07 |
| timburke | yeah | 21:07 |
| timburke | *shrug* i guess we can just go with the extra header -- i'm just hesitant to have to support it indefinitely | 21:10 |
| *** DSLegends has joined #swift3 | 21:13 | |
| *** NM has quit IRC | 21:30 | |
| *** chsc has quit IRC | 23:31 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!