Merge lp://staging/~chipaca/snappy/lock-ness into lp://staging/~snappy-dev/snappy/snappy-moved-to-github
Proposed by
John Lenton
Status: | Needs review |
---|---|
Proposed branch: | lp://staging/~chipaca/snappy/lock-ness |
Merge into: | lp://staging/~snappy-dev/snappy/snappy-moved-to-github |
Diff against target: |
988 lines (+597/-51) 8 files modified
cmd/snappy/common.go (+4/-5) daemon/api.go (+53/-4) daemon/api_test.go (+24/-40) daemon/daemon.go (+4/-0) daemon/daemon_test.go (+23/-2) daemon/mmutex/mmutex.go (+213/-0) daemon/mmutex/mmutex_test.go (+274/-0) dirs/dirs.go (+2/-0) |
To merge this branch: | bzr merge lp://staging/~chipaca/snappy/lock-ness |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Gustavo Niemeyer | Needs Information | ||
Review via email:
|
Commit message
REST API now uses hierarchical package-grained locking.
Description of the change
Moo. TeX. What's not to like.
To post a comment you must log in.
Unmerged revisions
- 772. By John Lenton
-
Introducing the MMutex. It is not a cow in affordable clothing.
- 771. By John Lenton
-
Committing in search for a name
Thanks for this branch! It looks good (but I need to read through it again as its quite complex at first). I got a bit of a meta-question. It seems like all the locking is done in the rest-api which is fine as its the only concurrent user of the API right now. But it also means that if someone builds on top of the snappy api he/she will have to rebuild the same locking. Is it a huge amount of work to put the locks inside snappy/* itself or is there another downside that I have not considered?