2b6a4e8d9a
[entropy.client.cache] improve _repositories_hash() robustness against collisions
2014-01-30 14:56:37 +01:00
f6591e1ade
[entropy.client.cache] drop _selective_repositories_hash()
2014-01-30 14:56:37 +01:00
f5eb6c1803
[entropy.client.cache] drop _all_repositories_hash, duplicate of _repositories_hash()
2014-01-30 14:56:37 +01:00
3c3a376a3b
[entropy.client] rename _filter_available_repositories() to filter_repositories()
2014-01-30 14:56:37 +01:00
a83935d33c
[entropy.client.client] move clear_cache() to Client
2014-01-30 14:56:37 +01:00
34a03796cb
Release Entropy 269
2014-01-28 14:39:33 +01:00
1841234eb4
Release Entropy 268
2014-01-28 14:17:45 +01:00
4988c54502
Release Entropy 267
2014-01-28 11:18:50 +01:00
49b210e6b0
Release Entropy 266
2014-01-24 14:04:11 +01:00
96b3b7115a
[entropy.locks] call const_setup_directory only, it's enough
...
Also deal with possible directory creation errors. We can safely
ignore them, since we would fail right after anyway.
This will allow the code to deal with the case where the directory
exists but we don't have enough perms to mkdir it.
2014-01-16 07:22:26 +01:00
5423663751
[entropy.const] fix default makedirs permissions, const_setup_file does it anyway
2014-01-16 07:21:57 +01:00
0b881e238f
Release Entropy 265
2014-01-14 09:10:34 +01:00
e76775701e
[entropy.client.package] multifetch: define failed_downloads before use
2014-01-05 15:51:29 +01:00
64b0e8baa9
[entropy.client] Client: rewrite lazy loading support, improve reliability
2014-01-05 14:53:05 +01:00
301aa281d5
[entropy.client] avoid races in Entropy Client @property methods
2014-01-05 13:59:14 +01:00
881f8831cf
[entropy.client] set _real_enabled_repos atomically
...
The race was triggered by the startup code path in Rigo.
2014-01-05 13:55:31 +01:00
8c4737ef7a
[entropy.core] SystemSettingsPlugin: make parse() return plugin metadata
...
This is needed in order to avoid infinite recursion through
__maybe_lazy_load()
2014-01-05 13:32:36 +01:00
64fb91dcae
[entropy.core] SystemSettings: enable lazy loading of external plugins as well
2014-01-05 13:00:31 +01:00
92a5f2fcc1
Release Entropy 264
2014-01-04 11:16:46 +01:00
3ebba8b14f
[entropy.db.sql] fix typo in SQL query, s/tag/versiontag/
2014-01-04 11:15:27 +01:00
c48869bb31
[entropy.core] SystemSettings: drop mtime based in-RAM cache
...
It does more harm than good and we have lazy loading now
2014-01-04 11:01:27 +01:00
c98c39db06
[entropy.core] SystemSettings: drop mtime based cache validation
...
This is no longer necessary and error prone on filesystems with
coarse mtime granularity (like ext3). Individual methods are able
to validate their own caches.
2014-01-04 10:31:09 +01:00
5da276bfb6
[entropy.core] SystemSettings: __remove_repo_cache() is no longer necessary
2014-01-04 10:10:49 +01:00
9af5689a13
[entropy.core] SystemSettings: simplify __getitem__
2014-01-03 20:57:10 +01:00
ab6d1f9fc4
[entropy.core] SystemSettings: lazy load item only if it hasn't been loaded yet
2014-01-03 20:55:36 +01:00
2053989b27
[entropy.core] SystemSettings: add lazy loading of metadata
2014-01-03 20:25:52 +01:00
7877126107
Release Entropy 263
2014-01-03 16:12:21 +01:00
b0bd765b50
[entropy.db.sqlite] do not acquire the repo lock if it's not strictly needed
...
Try to avoid to acquire the repository lock in exclusive mode if
it's not strictly needed. We were acquiring the lock in exclusive mode
even when there was nothing to update due to schema changes.
2014-01-03 16:10:33 +01:00
f1de6accef
Release Entropy 262
2014-01-03 15:50:06 +01:00
74f8bb63f5
[entropy.spm] PortagePlugin: update portage.doebuild() call, remove deprecated root argument
2014-01-03 15:45:39 +01:00
d03466ea10
[entropy.db.cache] disable in-RAM cache if total memory is less than 4G
2014-01-03 14:54:04 +01:00
ee48b812ed
[entropy.tools] add total_memory() function
2014-01-03 14:53:42 +01:00
338b3f6cc5
[entropy.db] expose and use the currently set cache policy
2014-01-02 23:00:00 +01:00
b6ed5f9948
[entropy.db.cache] introduce EntropyRepositoryCachePolicies
...
It is now possible to define a policy for the in-RAM cache through
EntropyRepository constructor or EntropyRepositoryCachePolicies.
2014-01-02 22:58:54 +01:00
2d22414086
Release Entropy 261
2014-01-01 11:03:52 +01:00
e16ad0a828
[entropy.db.sqlite] catch LockAcquireError in _maybeDatabaseSchemaUpdates
...
User may not have the privileges to acquire the lock, but we should
not fail here.
2014-01-01 10:58:54 +01:00
999f0509dd
[entropy.db] raise LockAcquireError in case of lock acquire error failures
2014-01-01 10:58:17 +01:00
f0d437a090
Release Entropy 260
2013-12-31 18:01:25 +01:00
2f713f73d5
[entropy.db.sqlite] raise timeout to 300 seconds
...
This will better fit slow disks.
2013-12-31 18:00:35 +01:00
1598f5e501
[entropy.*] use get_repository() to return a GenericRepository as well
...
All the repository classes must go through get_repository() to allow
further monkey patching and other external, API consumer related features.
2013-12-31 17:50:31 +01:00
4edaf614f4
[entropy.db.sql] in global context, the current thread is always MainThread
2013-12-31 17:38:57 +01:00
72aa0a07d1
[entropy.*] turn get_repository() into a classmethod
2013-12-31 17:30:29 +01:00
956b31d05e
[entropy.db.sql] add overridable isMainThread method
...
a MainThread may be a "liquid" concept. A dedicated worker thread
may be considered a "MainThread" from the point of view of EntropySQLRepository.
_start_cleanup_monitor checks whether the calling thread is the MainThread
in order to determine if a cleanup thread should be started joining
the current thread. The problem with this approach is that if our
worker thread is a long lasting thread, we will end up with hundreds
of cleanup monitor threads waiting for the worker thread to terminate.
This will never happen though.
We can now reimplement isMainThread in order to tell the internal
function not to start the cleanup monitor thread also for arbitrary
threads (like the worker thread in RigoDaemon).
2013-12-31 17:24:24 +01:00
88ea4c7427
[entropy.client.services] fix typo in HTTP arguments for get_documents()
...
s/revision/rev/
2013-12-31 17:04:36 +01:00
af6fa0d32f
[entropy.spm] PortagePlugin: add clear() to API, see documentation
2013-12-31 16:52:52 +01:00
0a9fbaf0df
Release Entropy 259
2013-12-31 16:02:38 +01:00
cf9fc1b947
Release Entropy 258
2013-12-31 15:17:41 +01:00
0346f74c1a
[entropy.client.package] catch InterruptError raised by fetch_abort functions
2013-12-31 12:28:29 +01:00
c5adbc0b20
Release Entropy 257
2013-12-27 19:21:09 +01:00
b088df88bd
[entropy.cache] EntropyCacher: drop CACHE_IDS (\o/) and clear_cache()
2013-12-27 19:15:20 +01:00
Fabio Erculiani