Fredrik Medley [Fri, 29 May 2015 18:04:50 +0000 (20:04 +0200)]
UploadPack: Use reachable-sha1-in-want configuration
C git 2.5 supports setting the equivalent of
RequestPolicy.REACHABLE_COMMIT with uploadpack.allowreachablesha1inwant.
Parse this into TransportConfig and use it from UploadPack. An explicitly
set RequestPolicy overrides the config, and the policy may still be
upgraded on a unidirectional connection to avoid races.
Change-Id: Id39771a6e42d8082099acde11249306828a053c0 Signed-off-by: Fredrik Medley <fredrik.medley@gmail.com>
Jonathan Nieder [Fri, 5 Jun 2015 23:08:55 +0000 (16:08 -0700)]
Allow lookup of multiple exact refs in one shot
exactRef(ref1, ref2, ref3) requests multiple specific refs in a single
lookup, which may be faster in some backends than looking them up one by
one.
firstExactRef generalizes getRef by finding the first existing ref from
the list of refs named. Its main purpose is for the default
implementation of getRef (finding the first existing ref in a search
path). Hopefully it can be useful for other operations that look for
refs in a search path (e.g., git log --notes=<name>), too.
Change-Id: I5c6fcf1d3920f6968b8b97f3d4c3a267258c4b86 Signed-off-by: Jonathan Nieder <jrn@google.com>
Jonathan Nieder [Fri, 5 Jun 2015 21:14:55 +0000 (14:14 -0700)]
Introduce exactRef to read a ref whose exact name is known
Unlike getRef(name), the new exactRef method does not walk the search
path. This should produce a less confusing result than getRef when the
exact ref name is known: it will not try to resolve refs/foo/bar to
refs/heads/refs/foo/bar even when refs/foo/bar does not exist.
It can be faster than both getRefs(ALL).get(name) and getRef(name)
because it only needs to examine a single ref.
A follow-up change will introduce a findRef synonym to getRef and
deprecate getRef to make the choice a caller is making more obvious
(exactRef or findRef, with the same semantics as getRefs(ALL).get and
getRefs(ALL).findRef).
Change-Id: If1bd09bcfc9919e7976a4d77f13184ea58dcda52 Signed-off-by: Jonathan Nieder <jrn@google.com>
David Pletcher [Thu, 4 Jun 2015 18:48:32 +0000 (11:48 -0700)]
Restore checkObjectCollisions flag
I am developing an offline pack verification feature based on
PackParser. The birthday collision check is a prohibitive obstacle
to performance at scale because it interacts with the repository
to perform collision checks. This CL restores the checkObjectCollisions
flag that was removed in 9638e0aa87614a6fb4f109bbeac0cde3462b9769,
while changing the flag getter and setter to protected from public as a
precaution against misuse.
Change-Id: I363cd0c9de57c5e8659cdfe2d51b17823f4fe793 Signed-off-by: David Pletcher <dpletcher@google.com>
Make JGit make the same distinction. This is in line with the current
Javadoc of getString, which claims to return "a String value from the
config, null if not found". It is more reasonable to interpret "x ="
in the above example as "found" rather than "missing".
We need to maintain the special handling of a key name with no "="
resolving to a boolean true, but "=" with an empty string is still not
a valid boolean.
Return -1 from PackWriter.Statistics.getBitmapIndexMises() when no
bitmap indices were found, to differentiate it from the case where
the bitmap indices contained all of the want/have commits.
Change-Id: I78d4600b462c19f62b347217a0b2c19eaaf3a14b Signed-off-by: Terry Parker <tparker@google.com>
Matthias Sohn [Mon, 1 Jun 2015 22:13:07 +0000 (00:13 +0200)]
Merge branch 'master' into stable-4.0
* master:
Add more reports to maven site
Update project meta-data in pom.xml used for site generation
Update dependencies to the versions used in target platform
Update 4.5 target platform to use final Mars Orbit repository
Compare API changes in clirr report against 3.7.0
Fix CommitCommand.setOnly()
Add "src" folder to source folders of org.eclipse.jgit.test
Improve exception thrown when pull can't find advertised ref
Silence unchecked conversion warning in TransportSftp
Silence deprecation warning in WindowCacheConfig
Silence deprecation warning in DirCacheCheckout
Fix unchecked conversion warning in MergeFormatter
Fix hidden field warnings in bundle org.eclipse.jgit
Close WindowCursor using try-with-resources in UnpackedObject
Fix WindowCursor memory leak.
archive: Drop unnecessary empty comments and 'final' qualifiers on locals
Close 'out' consistently in ArchiveCommand.call
Fix that exceptions in ReceivePack cause Invalid Channel 101 exceptions
Better report too large pack-files from PushCommand
FS: Extract GobblerThread into a private static class
Add bitmap index misses to PackWriter.Statistics
Enable public access to SimilarityIndex scoring function
Add getters to RepoProject.
Silence unused object warning in MyersDiff
Silence resource leak warnings where caller is responsible to close
Silence false potential null pointer access warnings
Fix potential null pointer access in IndexDiffFilter
Add tests for ObjectFilter
Expose Sets helper to tests outside org.eclipse.jgit.api
Change-Id: I34b2bb45f51ed6f52a6bb1215de654ebb2ffde10 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Matthias Sohn [Fri, 29 May 2015 12:39:33 +0000 (14:39 +0200)]
Update project meta-data in pom.xml used for site generation
- Update list of committers
- Update scm information
- Expose jgit hudson for Maven site
- Name our project's organization
- Scm URL should point to a web-enabled view on the repository
- Extract tycho-extras version used during signing to a property
Change-Id: If32bed323ec283b5e83d28ffe8775220f7c4cecd Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
When CommitCommand.setOnly(path) is used a temporary index has to be
created containing only modifications for the specified pathes. The
process to fill this temporary index was broken because filling the
index was stopped when we reached the first untracked file. Instead the
process to fill the temporary index should continue until we processed
all pathes.
Bug: 451465
Change-Id: I22af50d70bd3b81e7c056358724956122b0d158d Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Matthias Sohn [Wed, 15 Apr 2015 06:40:05 +0000 (08:40 +0200)]
Improve exception thrown when pull can't find advertised ref
- throw an API exception instead of an internal exception to allow
applications to handle this problem
- improve error message to give hints about possible root causes
Bug: 464660
Change-Id: Ib7d18bb2eeeac0fc218daea375b290ea5034bda1 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Shawn Pearce [Thu, 28 May 2015 19:39:07 +0000 (15:39 -0400)]
Merge changes I7b6d7be4,I63a74651,I39c2ea6b
* changes:
Silence unused object warning in MyersDiff
Silence resource leak warnings where caller is responsible to close
Fix potential null pointer access in IndexDiffFilter
Jonathan Nieder [Wed, 27 May 2015 23:25:49 +0000 (16:25 -0700)]
archive: Drop unnecessary empty comments and 'final' qualifiers on locals
Early JGit code used comments to inform the Eclipse formatter about
where to break lines and used final in the hope of making code faster.
The ArchiveCommand command implementation imitated that style.
Nowadays the project relies less on the Eclipse formatter and relies
more on Java having sane performance with local variables that are not
explicitly marked 'final'. Removing the unnecessary empty comments and
'final' qualifiers makes this code more readable and more consistent
with recent JGit code.
Change-Id: I7a181432eda7e18bd32cf110d89c0efbe490c4f1 Signed-off-by: Jonathan Nieder <jrn@google.com>
Jonathan Nieder [Wed, 27 May 2015 23:25:32 +0000 (16:25 -0700)]
Close 'out' consistently in ArchiveCommand.call
Whether the output stream specified with setOutputStream() is closed by
ArchiveCommand.call() is murky and inconsistent:
- on success, it is closed
- if an exception is encountered when writing the archive, it is closed
- if an exception is encountered when calling createArchiveStream to
open the archive, we forget to close it
Close the output stream consistently to avoid leaks.
Now that the inner try-with-resources doesn't have its own finally
block, this allows us to merge the two try blocks.
It would be even better to never close the output stream. That will
involve more API changes to avoid silently breaking callers, so it is
deferred to a later change.
Change-Id: I0185bdaa60ecee4a541eab5d8ff6c9c4dbe40bf1 Signed-off-by: Jonathan Nieder <jrn@google.com>
Fix that exceptions in ReceivePack cause Invalid Channel 101 exceptions
When during a PushOperation the server hits an exception different from
UnpackException the JGit server behaved wrong. That kind of exceptions
are handled so late that the connection is already released and the
information whether to talk sideband to the client is lost. In detail:
ReceivePack.receive() will call release() and that will reset the
capabilities. But later on the stack in ReceivePackServlet.doPost() it
is tried to send a response to client now with reset capabilities (no
sideband!).
Change-Id: I0a609acc6152ab43b47a93d712deb65bb1105f75 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Better report too large pack-files from PushCommand
JGits PushCommand and BasePackPushConnection were throwing generic
TransportExceptions when the pushed pack-file was rejected by the server
since it was too big. Let JGit better interprete the server's response
to detect this situation and throw a more specific exception.
This detection works by parsing the status line sent by the server. This
change only recognizes the response sent by a JGit based server. All
other servers which report such problems in a different way still lead
to a generic TransportExceptions.
Change-Id: Ic075764ea152939ce72c446252464620dd54edea Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Dave Borowitz [Wed, 27 May 2015 16:40:32 +0000 (09:40 -0700)]
FS: Extract GobblerThread into a private static class
The primary goal is to improve exception readability. Since this is a
standalone thread, just logging the stack trace of the caught
exception is not very useful:
java.io.IOException: Stream closed
at java.io.BufferedInputStream.getBufIfOpen(BufferedInputStream.java:162)
at java.io.BufferedInputStream.read(BufferedInputStream.java:258)
at org.eclipse.jgit.util.FS$2.run(FS.java:451)
Providing a named class eliminates the "FS$2", and including the
command name provides a little more context in the error message.
A future improvement might include the stack trace that created the
GobblerThread as well.
Terry Parker [Tue, 12 May 2015 00:37:02 +0000 (17:37 -0700)]
Add bitmap index misses to PackWriter.Statistics
RevWalks to find commits that are not in bitmap indices are expensive.
Track the count of commits that are enumerated via RevWalks as "bitmap
index misses" in the PackWriter.Statistics class.
Change-Id: Ie0135a0a0aeba2dfb6df78839d545006629f16cb Signed-off-by: Terry Parker <tparker@google.com>
David Pletcher [Mon, 25 May 2015 21:49:58 +0000 (14:49 -0700)]
Enable public access to SimilarityIndex scoring function
The SimilarityIndex class implements the useful capability of scoring
the similarity between two files. That capability is required for a
feature that's being developed in another package, to detect files
derived from a set of potential sources.
This CL adds a public factory method to create a SimilarityIndex from
an ObjectLoader. It grants public access to the SimilarityIndex class,
the score method, an inner exception class and a special marker
instance of that exception class.
Change-Id: I3f72670da643be3bb8e261c5af5e9664bcd0401b Signed-off-by: David Pletcher <dpletcher@google.com>
Matthias Sohn [Tue, 26 May 2015 09:39:19 +0000 (11:39 +0200)]
Merge branch 'master' into stable-4.0
* master:
Silence non-externalized string warnings in org.eclipse.jgit
Externalize translatable texts in org.eclipse.jgit
Don't invalidate pack file on InterruptedIOException
Update Mars target platforms to use Mars RC2 orbit
Update build to use eclipse-jarsigner-plugin 1.1.2
Guard agains null ReflogReader if named ref does not exist
FS: Allow to manually set the path to the Git system config file
FS: Fix a minor typo in runInShell() docs
FS: Improve javadoc of some recently introduced methods
Cleanup code and Eclipse compile errors in new gitrepo API
Refactor to expose ManifestParser.
FS: Remove the gitprefix logic
SystemReader: Use discoverGitSystemConfig() in openSystemConfig()
FS: Add a method to discover the system-wide config file
FS: Extend readPipe() to optionally take additional environment
FS: Document readpipe()'s encoding parameter
Split discoverGitPrefix() code out into discoverGitExe()
Equalize discoverGitPrefix() implementations between POSIX and Win32
Move resolveGrandparentFile() to the base class for wider use
Replace deprecated release() methods by close()
Use AutoClosable to close resources in bundle org.eclipse.jgit
ReceivePack: support quiet capability
Fix ObjectReader resources leak
Change-Id: I0cd9f7ad57f26f0a0cbf412845d00ba1efbea346 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Saša Živkov [Wed, 20 May 2015 14:57:24 +0000 (16:57 +0200)]
Don't invalidate pack file on InterruptedIOException
If the thread reading a pack file is interrupted don't invalidate that
pack file.
This could happen when Gerrit invoked JGit for computing a diff in one
thread and waited for the call to finish from another thread, with a
timeout. When the timeout was reached the "diff" thread was interrupted.
If it happened to be in an IO operation, reading a pack file, an
InterruptedIOException was thrown and the pack file was marked as
invalid and removed from the pack list.
Invalidating the pack in that case could cause the project disappearing in
Gerrit as discussed in [1] and [2].
Change-Id: I2eb1f98370936b5be541d96d70c3973cbfc39238 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com> Signed-off-by: Sasa Zivkov <sasa.zivkov@sap.com>
FS: Allow to manually set the path to the Git system config file
Now that d7a4473 removed the gitprefix property, we did not have a way to
specify the path to the Git system config file in case
discoverGitSystemConfig() fails. Fix that by introducing a member variable
that caches the result of discoverGitSystemConfig() as well as a setter
method to overwrite the content of that variable.
Change-Id: Icd965bffbe2f11b18c9505ee2ddd2afad5b64d70 Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com>
FS: Improve javadoc of some recently introduced methods
Change-Id: I31e788ee20ac3e8439559d9060d39e9792f6dc7d Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Andrey Loskutov [Sun, 24 May 2015 07:31:17 +0000 (09:31 +0200)]
Cleanup code and Eclipse compile errors in new gitrepo API
Commit d3348e introduced few errors in Eclipse.
This commit cleans up the new API:
- fixes API error in RepoCommand after moving IncludedFileReader type
- fixes unused imports in RepoCommand & RepoCommandTest
- fix javadoc errors in ManifestParser & RepoProject
- makes three (implicitly final) fields in ManifestParser final.
The only purpose of the gitprefix logic was to determine the path to the
system-wide config file. This is now done by discoverGitSystemConfig()
independent of the gitprefix, so get rid of this unused code.
Change-Id: Iaa88df9bd066dc1ed4067c18618af809e49876b3 Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com>
FS: Add a method to discover the system-wide config file
Change-Id: I969e26a5ab5f8ca3ab29024f405c1e34afdba493 Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Jonathan Nieder [Tue, 19 May 2015 01:28:31 +0000 (18:28 -0700)]
Add tests for ObjectFilter
Test that
- the default ObjectFilter is ALL
- ObjectFilter affects nextObject() and not next()
- omitting a tree implies omitting its subtrees
- a blob or tree reached by another path is still returned
- ObjectFilter can be mixed with RevFilter
Change-Id: I144a53fe677070fff8c3ddf8cba07a848773bc1b Signed-off-by: Jonathan Nieder <jrn@google.com>
FS: Extend readPipe() to optionally take additional environment
Change-Id: I4db7763826e4ada92074317d4d1c9a32299f3af8 Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Split discoverGitPrefix() code out into discoverGitExe()
Change-Id: I700540eec06efb24eeb09bfcb40420820c32d156 Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Move resolveGrandparentFile() to the base class for wider use
Change-Id: I67ec732ea2e5345a6946783f0c5ef60c07ce254e Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Shawn Pearce [Wed, 20 May 2015 07:34:00 +0000 (00:34 -0700)]
ReceivePack: support quiet capability
git-core has supported this for a long time; allowing clients to
avoid progress messages from the server if they are dumping to a
pipe instead of a tty.
Avoid the two progress monitors going on side-band and expose
isQuiet() method to allow hooks to also reduce their output if
this is sensible for them.
Matthias Sohn [Tue, 19 May 2015 14:07:17 +0000 (16:07 +0200)]
Merge branch 'master' into stable-4.0
* master:
Fix warnings in ObjectFilter
Fix typo in reflog message written by RebaseCommand.tryFastForward()
Correct @since tags for ObjectFilter
Fix typo in ObjectWalk#getObjectFilter javadoc
Allow ObjectWalk to be filtered by an arbitrary predicate
Remove SoftReference from dfs.DeltaBaseCache
Fix memory leak in dfs.DeltaBaseCase
Update to Jetty 9.2.10
Update javax.servlet to 3.1
Use ANY_DIFF filter in ResolveMerger only for bare repositories
FS_POSIX: Rework umask detection to make it settable
Expose disposeBody() on RevCommit and RevTag
ObjectReader: remove the walkAdvice API
RevWalk: Discard uninteresting commits unless RevSort.BOUNDARY
ObjectWalk: make setRetainBody(false) the default
Do not concatenate strings as arguments to StringBuilder.append()
IndexDiffFilter: Simplify a boolean expression
GroupHead: Remove a redundant call to String.format()
FS_Win32: Avoid an IOException on Windows if bash is not in PATH
Skip logging stack trace on corrupt objects
Add repository name to failures in HTTP server log
Fix possible AIOOB in DirCacheTree.contains()
Delete deprecated PackWriter.preparePack() methods
Delete deprecated class IgnoreRule
Delete deprecated checkoutEntry() methods in DirCacheCheckout
Fix IllegalArgumentException in AmazonS3
Change-Id: Ica3d4f0675c81684fbe48fcf0053f2a949bc5c9b Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Hugo Arès [Fri, 15 May 2015 19:15:28 +0000 (15:15 -0400)]
Fix ObjectReader resources leak
In 77030a5e, AutoClosable was implemented on classes that use release().
This caused a resource leak because the ObjectReader.close method was
not calling the now deprecated release method, which is the method that
sub classes implements to release resources.
Change-Id: I247651ec8fd7ca9941d256ca46d14cc43cc35c6e Signed-off-by: Hugo Arès <hugo.ares@ericsson.com>
Jonathan Nieder [Fri, 15 May 2015 18:37:57 +0000 (11:37 -0700)]
Correct @since tags for ObjectFilter
Although the stable-4.0 branch already exists, 4.0 development is
still happening on master until IP logs are sent for review, which
will happen at the end of May.
Change-Id: I863ba85c6303f8ef2eb13bca5e2d30e5d3c58b29 Signed-off-by: Jonathan Nieder <jrn@google.com>
Jonathan Nieder [Thu, 14 May 2015 23:24:15 +0000 (16:24 -0700)]
Allow ObjectWalk to be filtered by an arbitrary predicate
This will make it possible to declare a collection of objects as
ineligible for the walk en masse, for example if they are known to be
uninteresting via a bitmap.
Change-Id: I637008b25bf9fb57df60ebb2133a70214930546a Signed-off-by: Jonathan Nieder <jrn@google.com>
Shawn Pearce [Sun, 10 May 2015 02:53:53 +0000 (19:53 -0700)]
Remove SoftReference from dfs.DeltaBaseCache
The Java GC doesn't always clear these before running out of memory
and failing allocations. In practice OpenJDK 7 is leaving these live,
removing any advantage of the SoftReference to attempt to shed memory
when the GC is unable to continue allocating.
Instead follow the pattern of the DfsBlockCache and use hard refs
to the object data. Require applications to configure the cache
size more accurately given expected memory usage.
Shawn Pearce [Sun, 10 May 2015 01:11:20 +0000 (18:11 -0700)]
Fix memory leak in dfs.DeltaBaseCase
The LRU chain management code was broken leading to situations where
the chain was incomplete. This prevented the cache from removing
items when it exceeded its memory target, causing a leak.
One case was repeated hit on the head of the chain. moveToHead(e)
was invoked linking the head back to itself in a cycle orphaning
the rest of the table.
Add some unit tests to cover this and a few other paths.
projects / jgit.git / log