Performance testing has indicated the per-process ChunkCache isn't
very effective for the DHT storage implementation. If a server is
using the DHT storage backend, it is most likely part of a larger
cluster where requests are distributed in a round-robin fashion
between the member servers.
In such a scenario there is insufficient data locality between
requests to get a good hit ratio on the per-process ChunkCache. A low
hit ratio means the cache is actually hurting performance by eating up
memory that could otherwise be used for transient request data, and
increasing pressure on the GC when it needs to find free space.
Remove all of the ChunkCache code. Installations that want to cache
(to reduce database usage) should wrap their Database with a
CacheDatabase and use a network based CacheServer.
I left the ChunkCache in the original DHT storage commit because I
wanted to document in the history of the project that its probably
worth *not* having, but leave open a door for someone to revert this
change if they find otherwise at a later date.
Change-Id: I364d0725c46c5a19f7443642a40c89ba4d3fdd29
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Chris Aniszczyk <caniszczyk@gmail.com>
Adds a class which can be used to calculates a SHA1 of the diff
associated with a patch, similar to git patch-id.
In this version whitespace is not ignored.
Change-Id: I421d15ea905e23df543082786786841cbe3ef10d
Signed-off-by: Stefan Lay <stefan.lay@sap.com>
Signed-off-by: Chris Aniszczyk <caniszczyk@gmail.com>
Since this change may affect performance and memory consumption on every
access to a loose ref I explicitly made it a RFC to collect opinions.
Previously RefDirectory.scanRef() was not detecting an update of a
loose ref when the update didn't changed the modification time of
the backing file. RefDirectory cached loose refs and the way to detect
outdated cache entries was to compare lastmodification timestamp on the
file representing the ref. If two updates to the same ref happen faster
than the filesystem-timer granularity (for linux this is 2 seconds)
there is the possiblity that we don't detect the update.
Because of this bug EGit's PushOperationTest only works with 2 second
sleeps inside.
This change let RefDirectory use FileSnapshot to detect such situations.
FileSnapshot helps to remember when a file was last read from disk and
therefore enables to decide when to load a file from disk although
modification time has not changed.
Change-Id: I03b9a137af097ec69c4c5e2eaa512d2bdd7fe080
Signed-off-by: Christian Halstrick <christian.halstrick@sap.com>
Signed-off-by: Chris Aniszczyk <caniszczyk@gmail.com>
A checkout conflict during rebase setup should leave the repository
in SAFE state which means ensuring that the rebase temporary files
need to be removed.
Bug: 346813
Change-Id: If8b758fde73ed5a452a99a195a844825a03bae1a
Signed-off-by: Chris Aniszczyk <caniszczyk@gmail.com>
Change Ia2ab4f8dc95020f2914ff01c2bf3b1bc62a9d45d added merge
support for when OURS or THEIRS was simultaneously deleted
and modified. That changeset however did not add create an
entry in the conflicts table so clients would see a CONFLICTING
result but getConflicts() would return null.
This change creates a MergeResult for the conflicting file.
Bug: 345684
Change-Id: I52acb81c1729b49c9fb3e7a477c6448d8e55c317
Signed-off-by: Chris Aniszczyk <caniszczyk@gmail.com>
Change Ib9898fe0f982fa08e41f1dca9452c43de715fdb6 added support for
the 'cherry-pick' fast forward case where the upstream commit history
does not include any merge commits. This change adds support for the
case where merge commits exist and the local branch has no changes.
Bug: 344779
Change-Id: If203ce5aa1b4e5d4d7982deb621b710e71f4ee10
Signed-off-by: Chris Aniszczyk <caniszczyk@gmail.com>
Previously when merging two contents with a non-empty base and one of
the contents was empty (size == 0) and the other was modified there
was a potentially expensive calculation until we finally always come
to the same result -> the complete non-deleted content should collide
with the empty content.
This proposal adds an optimization to detect empty input content and
to produce the appropriate result immediatly.
Change-Id: Ie6a837260c19d808f0e99173f570ff96dd22acd3
Signed-off-by: Christian Halstrick <christian.halstrick@sap.com>
For the following patch on the linux 2.6.32 tag:
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -685,6 +685,7 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sc
static void check_spread(struct cfs_rq *cfs_rq, struct sched_entity *se)
{
+#if 0
#ifdef CONFIG_SCHED_DEBUG
s64 d = se->vruntime - cfs_rq->min_vruntime;
@@ -694,6 +695,7 @@ static void check_spread(struct cfs_rq *cfs_rq, struct
sched
if (d > 3*sysctl_sched_latency)
schedstat_inc(cfs_rq, nr_spread_over);
#endif
+#endif
}
static void
JGit produced an incorrect diff, attempting to add a new "}" instead
of the new "#endif" at the end of the hunk. This was caused by a prior
fix for bug 328895 where we wanted to "slide" a diff down in the file
when adding a new method/function and want to show the closing curly
brace as being added after the new method, rather than added onto the
end of the prior function or method just before the insertion point.
Bug: 345956
Change-Id: I32b9e24f1e2980258b1b39dd1807919ab1c5f9b2
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
The problem occurred when the first text ends in the middle
of the last line of the other text and the first text has no
end of line.
Bug: 344975
Change-Id: I1f0dd9f8062f2148a7c1341c9122202e082ad19d
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
The reflog message fix was done in
I4f1c3cd6b2cf543be213f061afb94223062dde51
Change-Id: I44817ccf4bf226ed3e4ce6fb2d923e88788221dd
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
This makes the message look the same as in C Git (the "."):
This reverts commit <sha1>.
Change-Id: I4c254c122277b127e7b039c0d1c7f7a0d691530d
Signed-off-by: Robin Stocker <robin@nibor.org>
Since d1718a the method getHumanishName was broken on windows since
the URIish is not normalized anymore. For a path like
"C:\gitRepositories\egit" the whole path was returned instead of
"egit".
Bug: 343519
Change-Id: I95056009072b99d32f288966302d0f8188b47836
Signed-off-by: Stefan Lay <stefan.lay@sap.com>
Before this change any files in the conflicting set would
also be listed in the the other IndexDiff Sets which is
confusing. With this change a conflicting file will not
be included in any of the other sets.
Change-Id: Ife9f2652685220bcfddc1f9820423acdcd5acfdc
Signed-off-by: Chris Aniszczyk <caniszczyk@gmail.com>
Submodules are supposed to be handled by separate operations, so
we should ignore them on checkout, just like C Git does.
This fix does not add submodule support. We just try harder
to ignore them.
Bug: 343566
Change-Id: I2c5ae1024ea7bb57adf27072da6acc9643018eda
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
jgit.storage.dht is a storage provider implementation for JGit that
permits storing the Git repository in a distributed hashtable, NoSQL
system, or other database. The actual underlying storage system is
undefined, and can be plugged in by implementing 7 small interfaces:
* Database
* RepositoryIndexTable
* RepositoryTable
* RefTable
* ChunkTable
* ObjectIndexTable
* WriteBuffer
The storage provider interface tries to assume very little about the
underlying storage system, and requires only three key features:
* key -> value lookup (a hashtable is suitable)
* atomic updates on single rows
* asynchronous operations (Java's ExecutorService is easy to use)
Most NoSQL database products offer all 3 of these features in their
clients, and so does any decent network based cache system like the
open source memcache product. Relying only on key equality for data
retrevial makes it simple for the storage engine to distribute across
multiple machines. Traditional SQL systems could also be used with a
JDBC based spi implementation.
Before submitting this change I have implemented six storage systems
for the spi layer:
* Apache HBase[1]
* Apache Cassandra[2]
* Google Bigtable[3]
* an in-memory implementation for unit testing
* a JDBC implementation for SQL
* a generic cache provider that can ride on top of memcache
All six systems came in with an spi layer around 1000 lines of code to
implement the above 7 interfaces. This is a huge reduction in size
compared to prior attempts to implement a new JGit storage layer. As
this package shows, a complete JGit storage implementation is more
than 17,000 lines of fairly complex code.
A simple cache is provided in storage.dht.spi.cache. Implementers can
use CacheDatabase to wrap any other type of Database and perform fast
reads against a network based cache service, such as the open source
memcached[4]. An implementation of CacheService must be provided to
glue this spi onto the network cache.
[1] https://github.com/spearce/jgit_hbase
[2] https://github.com/spearce/jgit_cassandra
[3] http://labs.google.com/papers/bigtable.html
[4] http://memcached.org/
Change-Id: I0aa4072781f5ccc019ca421c036adff2c40c4295
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Merging into a non-master branch would result in the following message:
Merge branch 'a' into HEAD
Now the merge message is correct:
Merge branch 'a' into b
Change-Id: I488f97190e4c1711c23a7a3cbd64f8b13a87bbac
Signed-off-by: Robin Stocker <robin@nibor.org>
The filter did not correctly match smart HTTP client requests,
so it always fell back on HTTP status codes for errors. This
usually causes a smart client to retry a dumb request, which
is not what the server wants.
Change-Id: I42592378dc42fbe308ef30a2923786c690f668a9
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Shawn O. Pearce
Chris Aniszczyk
Stefan Lay
Abhishek Bhatnagar
Christian Halstrick
Kevin Sawicki
Ketan Padegaonkar
Bernard Leach
Christoph Brill
Piotr Janik
Robin Rosenberg
Robin Stocker
Christian Halstrick
Matthias Sohn