Commits

Author Commit Message Labels Comments Date
Mike Bayer
- add a conftest to turn on logging config
Mike Bayer
add support for py.test
Mike Bayer
missing tag
Mike Bayer
setup for git
Mike Bayer
add gpg identity
Mike Bayer
add tags
Mike Bayer
Added tag rel_0_4_1 for changeset 3e6f0f86b4161b48898e656039154092426e5cc8
Mike Bayer
release date
Tags
rel_0_4_1
Mike Bayer
2013 copyright
Mike Bayer
fix the pullreq link
Mike Bayer
i always screw up this markup
Mike Bayer
- version bump - changelog
Ralph Bean
Update async/background unit test to reflect evolution of the pull request.
Ralph Bean
Rework background_runner as async_creator.
Ralph Bean
Added single standalone test for background_runner.
Ralph Bean
At zzzeek's suggestion, change threaded_creation into background_runner. More generic.
Ralph Bean
Docstring comment on threaded_creation.
Ralph Bean
Allow creation of new values by a background thread.
Mike Bayer
- move to in-doc changelog
Mike Bayer
Added tag rel_0_4_0 for changeset 3ec2934f49356612b72261b5e86c9365bbdbac65
Mike Bayer
take this out
Tags
rel_0_4_0
Mike Bayer
no, ad hoc
Mike Bayer
nameregistry is still relevant, though we wont use it in the examples here.
Mike Bayer
yield the value here...
Mike Bayer
changelog
Mike Bayer
0.3 branch
Branches
rel_0_3
Mike Bayer
- deprecate Dogpile, and rework all the docs - fix SyncReaderDogpile
Mike Bayer
still want to do docs and perhaps some deprecations, but this will be 0.4.0
Mike Bayer
keeping track of creationtime as create is called
Mike Bayer
- factor out the guts of Dogpile into a new object Lock, that simplifies the state model and is a per-use object only - instead of checking self.is_expired inside the "acquired createlock" region, call value_and_created_fn() to coordinate with the state of other processes; this doesn't occur in single-process Dogpile usage due to the whole NameRegistry thing, but we can't really rely upon that for cross process. This is the start of the fix for #1, however some dogpi…
  1. Prev
  2. Next