DfsGarbageCollector will now enforce a maximum time to live (TTL) for
UNREACHABLE_GARBAGE packs. The default TTL is 1 day, which should be
enough time to avoid races with other processes that are inserting
data into the repository.
Change-Id: Id719e6e2a03cfc9a0c0aef8ed71d261dda14bd0c
Signed-off-by: Mike Williams <miwilliams@google.com>
1f86350 added initial support for include.path. Relative path and path
with tilde are not yet supported but config load was failing if one of
those 2 unsupported options was encountered. Another problem was that
config load was failing if the include.path file did not exist.
Change the behavior to be consistent with native git. Ignore unsupported
or nonexistent include.path.
Bug: 495505
Bug: 496732
Change-Id: I7285d0e7abb6389ba6983e9c46021bea4344af68
Signed-off-by: Hugo Arès <hugo.ares@ericsson.com>
Set all packs written by the DfsGarbageCollector to use the same
starting timestamp as lastModified. This makes it easier to see
which packs came from the same DfsGarbageCollector run, as they
share the same timestamp.
Change-Id: Id633573fbc3f0f360887b4745cacf33d6fc09320
Force reads to use a search ordering of:
INSERT / RECEIVE
COMPACT
GC (heads)
GC_REST (non-heads)
GC_TXN (refs/txn)
UNREACHABLE_GARBAGE
This has provided decent performance for object lookups. Starting
from an arbitrary reference may find the content in a newer pack
created by DfsObjectInserter or a ReceivePack server. Compaction of
recent packs also contains newer content, and then most interesting
data is in the "main" GC pack. As the GC pack is self-contained (has
no edges leading outside) readers typically do not need to go further.
Adding a new GC_REST PackSource allows the DfsGarbageCollector to
identify to the pack ordering code which pack is which, so the
non-heads are scanned second during reads. This removes a hack that
was unique to Google's implementation that enforced this behavior by
fixing up the lastModified timestamp.
Renumber the PackSource's categories to reflect this search ordering.
Change-Id: I19fdaab8a8d6687cbe8c88488e6daa0630bf189a
Treewalk has a member 'attr' which caches the attributes for the current
entry. We did not reset the cache always when moving to next entry. The
effect was that when there are no attributes for an entry 'a' but 'a'
was skipped by a Treewalk filter then Treewalk stopped looking for
attributes until TreeWalk.next() was called again.
Change-Id: Ied39b7fb5f56afe7a237da17801003d0abe6b1c7
Problem occurs when the checkout wants to create a file 'd/f' but
the workingtree contains a dirty file 'd'. In order to create d/f the
file 'd' would have to be deleted and since the file is dirty that
content would be lost. This should lead to a CheckoutConflictException
for d/f when failOnConflict was set to true.
This fix also changes jgit checkout semantics to be more like native
gits checkout semantics. If during a checkout jgit wants to delete a
folder but finds that the working tree contains a dirty file at this
path then JGit will now throw an exception instead of silently keeping
the dirty file. Like in this example:
git init
touch b
git add b
git commit -m addB
mkdir a
touch a/c
git add a/c
git commit -m addAC
rm -fr a
touch a
git checkout HEAD~
Change-Id: I9089123179e09dd565285d50b0caa308d290cccd
Signed-off-by: Rüdiger Herrmann <ruediger.herrmann@gmx.de>
Also-by: Rüdiger Herrmann <ruediger.herrmann@gmx.de>
The native wire protocol sends ref advertisements in the pkt-line
format, which requires encoding the ObjectId and ref name onto a byte
sequence. Busy servers show this is a very high source of garbage,
which pushes the garbage collector harder when there are many refs in
the repository (e.g. 70k, in a Gerrit managed repository).
Optimize the side band advertiser by retaining the CharsetEncoder,
minimizing the amount of temporary garbage built during encoding.
Change-Id: I406c654bf82c1eb94b38862da2425e98396134cb
Before this fix, ref directory removal did not work. That was because
the ref lock file was still in the leaf directory at deletion time.
Hence no deep ref directories were ever deleted, which negatively
impacted performance under large directory structure circumstances.
This fix removes the ref lock file before attempting to delete the ref
directory (which includes it). The other deep parent directories are
therefore now successfully deleted in turn, since leaf's content
(lock file) gets removed first.
So, given a structure such as refs/any/directory[/**], this fix now
deletes all empty directories up to -and including- 'directory'. The
'any' directory (e.g.) does not get deleted even if empty, as before.
The ref lock file is still also removed in the calling block's finally
clause, just in case, as before. Such double-unlock brought by this
fix is harmless (a no-op).
A new (private) RefDirectory#delete method is introduced to support
this #pack-specific case; other RefDirectory#delete callers remain
untouched.
Change-Id: I47ba1eeb9bcf0cb93d2ed105d84fea2dac756a5a
Signed-off-by: Marco Miller <marco.miller@ericsson.com>
Git servers supporting HTTP transport can send multiple WWW-Authenticate
challenges [1] for different authentication schemes the server supports.
If authentication fails now retry all authentication types proposed by
the server.
[1] https://tools.ietf.org/html/rfc2617#page-3
Bug: 492057
Change-Id: I01d438a5896f9b1008bd6b751ad9c7cbf780af1a
Signed-off-by: Christian Pontesegger <christian.pontesegger@web.de>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
setInput should always push at least 1 byte into the Inflater. If 0
bytes (or negative!) are being sent the DfsBlock is inconsistent with
the position passed in. This indicates a severe programming problem
in the caller, and may cause an infinite loop in DfsReader.
Today we saw a handful of live examples of this but don't know what
the cause is. Guard against this error condition and throw with a
more verbose failure, which may prevent an infinite loop. Callers
will eventually catch DataFormatException and rethrow with more detail
about the object that cannot be inflated, with the DFE in the chain.
Change-Id: I64ed2a471520e48283675c6210c6db8a60634635
When using a DfsInserter for high-throughput insertion of many
objects (analogous to git-fast-import), we don't necessarily want to
do a random object lookup for each. It'll be faster from the
inserter's perspective to insert the duplicate objects and let a later
GC handle the deduplication.
Change-Id: Ic97f5f01657b4525f157e6df66023f1f07fc1851
* master:
Fix javadoc errors and unused imports introduced by ddd0fe25
RepoCommand: record manifest shallow recommendation in .gitmodules
RepoCommand: record manifest groups as submodule labels
Remove the deprecated TestRepository.getClock() method
Replace use of deprecated method Repository.getRef()
[findBugs] Prevent potential NPE in
FileLfsRepository.getOutputStream()
Better report on client side if push failed due to too large object
[findBugs] Prevent potential NPE in CloneCommand.init()
RepoCommand: remove --record-remote-branches
RepoCommandTest: Improve assertion message for remote branch recording
Change-Id: I4fbce4f84925a933fcc9a48058ed6793f5821b97
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Git core learned about the submodule.<name>.shallow option in
.gitmodules files, which is a recommendation to clone a submodule
shallow. A repo manifest may record a clone depth recommendation as
an optional field, which contains more information than a binary
shallow/nonshallow recommendation, so any attempted conversion may be
lossy. In practice the clone depth recommendation is either '1' or doesn't
exist, which is the binary behavior we have in Git core.
Change-Id: I51aa9cb6d1d9660dae6ab6d21ad7bae9bc5325e6
Signed-off-by: Stefan Beller <sbeller@google.com>
Git core learned about attributes in pathspecs:
pathspec: allow querying for attributes
The pathspec mechanism is extended via the new
":(attr:eol=input)pattern/to/match" syntax to filter paths so that it
requires paths to not just match the given pattern but also have the
specified attrs attached for them to be chosen.
(177161a5f7, 2016-05-20)
We intend to use these pathspec attribute patterns for submodule
grouping, similar to the grouping in repo. So the RepoCommand which
translates repo manifest files into submodules should propagate this
information along. This requires writing information to the
.gitattributes file instead of the .gitmodules file. For now we just
overwrite any existing .gitattributes file and do not care about prior
attributes set. If this becomes an issue we need to figure out how to
correctly amend the grouping information to an existing .gitattributes
file.
Change-Id: I0f55b45786b6b8fc3d5be62d7f6aab9ac00ed60e
Signed-off-by: Stefan Beller <sbeller@google.com>
JGits PushCommand and BasePackPushConnection were throwing a generic
exception when the pushed pack-file was rejected by the server since it
contained too large objects. Teach JGit to better analyze the server's
response to detect this situation and throw a more specific exception.
Detect this situation 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.
Also see https://git.eclipse.org/r/#/c/46348/
Change-Id: I8d6d65e4585ebb3846f7207e7d1a2f82fa9cbd86
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
File.listFiles() returns null if the File is not a directory, improve
validation of directory and gitDir to fix this.
Change-Id: I763d08835faf96a0beb8e706992df0908526bd2c
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
* master:
JGit CLI: allow to call git init with specific directory
Redirect all Show output to outs
Support git config [include] section with absolute path(s)
Added filter for merge and non-merges commits.
[findBugs] Prevent potential NPE in FS_POSIX.readUmask()
[findBugs] Fix calculation of host header in SignerV4
Update Orbit repository to S20160518051658 for Neon RC2
Fix StashApply regarding handling of untracked files
GC should not pack objects only referenced by ORIG_HEAD,...
Make sure to overwrite files when "reset --hard" detects conflicts
Allow setting FileMode to executable when applying patches in
ApplyCommand
Fix config value get to return last instead of 1st just like git
Remove UTF-8 checking duplication in Config lib subclasses
Update Maven plugins
Fix type parameter in javadoc in TestRepository.delete(String ref)
TestRepository: Add delete() method
Make BaseReceivePack.setAtomic public
ReceivePack: Pass atomic setting from client to BatchRefUpdate
Change-Id: I5c9c5b7ccb23fb48b44b3da10b2c5d876d043d24
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
* stable-4.3:
Fix computation of id in WorkingTreeIterator with autocrlf and
smudging
Prepare 4.3.2-SNAPSHOT builds
JGit v4.3.1.201605051710-r
Scan loose ref before packed in case gc about to remove the loose
Fix possible NPEs when reporting transport errors
Fix calling of clean/smudge filters from Checkout,MergeCommands
Fix ApplyCommand when result of patch is an empty file
Change-Id: I829f06699f6670e519d04c927bdba4b82df29199
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
JGit failed to do checkouts when the index contained smudged entries and
autocrlf was on. In such cases the WorkingTreeIterator calculated the
SHA1 sometimes on content which was not correctly filtered. The SHA1 was
computed on content which two times went through a lf->crlf conversion.
We used to tell the treewalk whether it is a checkin or checkout
operation and always use the related filters when reading any content.
If on windows and autocrlf is true and we do a checkout operation then
we always used a lf->crlf conversion on any text content. That's not
correct. Even during a checkout we sometimes need the crlf->lf
conversion. E.g. when calculating the content-id for working-tree
content we need to use crlf->lf filtering although the overall operation
type is checkout.
Often this bug does not have effects because we seldom compute the
content-id of filesystem content during a checkout. But we do need to
know whether a file is dirty or not before we overwrite it during a
checkout. And if the index entries are smudged we don't trust the index
and compute filesystem-content-sha1's explicitly.
This caused EGit not to be able to switch branches anymore on Windows
when autocrlf was true. EGit denied the checkout because it thought
workingtree files are dirty because content-sha1 are computed on wrongly
filtered content.
Bug: 493360
Change-Id: I1072a57b4c529ba3aaa50b7b02d2b816bb64a9b8
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
As per [1], but limited to absolute paths indeed. No support yet for
tilde or $HOME expansion. Support for the --[no-]includes options
([1]) is not part of this commit scope either, but those options'
defaults are in effect as described in [1].
[1] https://git-scm.com/docs/git-config
Included path can be a config file that includes other path-s in turn.
An exception is thrown if too many recursions (circular includes)
happen because of ill-specified config files.
Change-Id: I700bd7b7e1625eb7de0180f220c707d8e7b0930b
Signed-off-by: Marco Miller <marco.miller@ericsson.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Added the option to retrieve either merge or non-merge commits in the
LogCommand.
Change-Id: Ie0e1c515a823f2392783f1a47d385c31230e8167
Signed-off-by: Alcemir Santos <alcemir.santos@gmail.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
BufferedReader.readLine() returns null if the end of the stream has been
reached.
Change-Id: I83102bbfb1316407247e0f29023077af9e8d9606
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
There was a bug regarding how JGit handled untracked files when applying
a stash. Problem was that untracked files are applied by doing a merge
of HEAD and untrackedFiles commit with a merge base of the stashed HEAD.
That's wrong because the untrackedFiles commit has no parent and
contains only the untracked files. Using stashed HEAD as merge base
leads to unneccessary conflicts on files not event included in the
untrackedFiles commit.
Imagine this graph directly before you want to apply a stash which was
based on 0. You want to apply the stash on current HEAD commit 5.
5 (HEAD,master)
/
0---+
\ \
1---3 (WIP on master)
/
2 (untracked files on master)
Imagine for a specific (tracked) file f
- commit 0 contains X
- HEAD contains Y
- commit 2 (the untracked files) does not contain file f
A merge of 2 and 5 with a merge base of 0 leads to a conflict. The 5
commit wants to modify the file and the 2 commit wants to delete the
file -> conflict.
If no merge base is set then the semantic is correct.
Thanks to Bow for finding this bug and providing the test case.
Bug: 485467
Change-Id: I453fa6ec337f81b2a52c4f51f23044faeec409e6
Also-by: Bow Ruggeri <bow@bow.net>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
There are references which are returned by
RefDatabase.getAdditionalRefs() which are allowed to point to
non-existing objects. These refs should not save objects from being
garbage collected. Examples for these references are ORIG_HEAD,
MERGE_HEAD, FETCH_HEAD and CHERRY_PICK_HEAD. Native git will not take
these references into account when doing a gc and therefore these
references may point to non-existing objects after a gc. Teach JGit's
GC to behave the same: ignore additional refs if they don't start with
"refs/". Examples for refs returned by getAdditionalRefs() which do
start with "refs/" are the bootstrap refs when using reftree's (see
commit 115f1ad3974d1162b33d1c8eff466019d1f249f3). See also
http://article.gmane.org/gmane.comp.version-control.git/294126.
Bug: 479697
Change-Id: I10e40589f13e72aacdd9f86f3b44696fd1cd068a
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
When performing a "reset --hard" a checkout is done. The pathes are
checked for potential checkout conflicts. But in the end for all
remaining conflicts these files are simply deleted from the working
tree. That's not the right strategy to handle checkout conflicts during
"reset --hard". Instead for every conflict we should simply checkout the
merge commit's content.
This is different from native gits behavior which reports errors when
during a "checkout --hard" a file shows up where a folder was expected.
"warning: unable to unlink d/c.txt: Not a directory"
Why it is like that in native git was asked in
http://permalink.gmane.org/gmane.comp.version-control.git/279482. Unless
it is explained why native git why this error is reported JGit should
overwrite the files.
Bug: 474842
Change-Id: I08e23822a577aaf22120c5137eb169b6bd08447b
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
git-apply allows modifying file modes in patched files using either
"new mode" or "new file mode" headers. This patch adds support for
setting files as executables and vice-versa.
Change-Id: I24848966b46f686f540a8efa8150b42e0d9c3ad1
Signed-off-by: Nadav Cohen <nadavcoh@gmail.com>
Before this fix, getting the value of 'key' below used to return
value1. This fix makes it so that value3 gets returned instead,
just like native git's get.
[section]
key = value1
key = value2
key = value3
Change-Id: Iccb24de9b63c3ad8646c909494ca3f8c9ed6e29c
Signed-off-by: Marco Miller <marco.miller@ericsson.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Change-Id: Ib9f0ae8207000a36c5bf1a92fcc2c32efc4c0984
Signed-off-by: Marco Miller <marco.miller@ericsson.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
* stable-4.3:
Scan loose ref before packed in case gc about to remove the loose
Fix possible NPEs when reporting transport errors
Fix calling of clean/smudge filters from Checkout,MergeCommands
Fix ApplyCommand when result of patch is an empty file
Change-Id: I0dc76b7a8c87ce8e0386a1b9e0fa92a3aa62abf7
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Before this change, jgit used to read packed-refs before scanning
loose refs. That was not a problem if gc didn't run concurrently. When
gc did run concurrently with such refs reading, that order sometimes
broke the latter. This lead to reading an older version of a ref's
tip, which meant "losing" the real tip or commit. The specific
read-Vs-gc concurrency scenario which broke reading that way follows:
1. let ref R be in packed-refs and R' be in loose
2. jgit starts reading packed-refs
3. gc also starts its business around that very time
4. jgit still has the time to read R from packed-refs
5. as gc is not done yet updating packed-refs with R'
6. jgit then starts scanning loose refs (or is about to)
7. gc quickly ends up being done moving loose R' to packed-refs
8. so gc (quickly) removes loose refs
9. -while jgit is scanning loose refs, now gone
10. so jgit assumes no loose to consider => packed-refs winning
11. so jgit wrongfully returns R (from 4.) as the tip, instead of R'.
This fix switches the order so loose refs are scanned (secured) before
taking the time to read packed-refs. This way, knowledge of the
likelier tip is guaranteed for ref reading to return the true tip
- despite concurrent gc. If there is no loose ref to scan, jgit reads
packed-refs and lands on R' (or S), which it then returns, as
expected. The gerrit issue [1] should be solved by this fix.
[1] https://code.google.com/p/gerrit/issues/detail?id=2302
Change-Id: Ibd120120a361a3a6ed565f3836afc1db706fbcdd
Signed-off-by: Marco Miller <marco.miller@ericsson.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
The special characters <> and '\n' interfere with parsing of
identities. C git strips these special characters, so we should too.
Rather than allocating extra strings by calling String#trim(), add a
few lines to our sanitization method to perform the same trimming as
described in String's Javadoc.
Change-Id: I96edcb93a2fc194ee354d60566d352299742a52f
This might be somewhat surprising behavior to users who might
naturally assume the following invariant:
ident.equals(parseIdent(ident.toExternalString()))
This invariant does not hold since whitespace is only trimmed during
serialization. We don't want to mess with the strings during
initialization, as this is called during the highly-optimized commit
parsing codepath.
Change-Id: I081a603f0ac0e33167462244779b0ff3ad51e80c
There was a bug in JGit which caused NPEs being thrown when Transport
errors should be reported. Avoid the NPE to let the original error show
up.
Change-Id: I9e1e2b0195bd61b7e531a09d0fc7bce109bd6515
We've found in Gerrit Code Review that it is common to pass around
both an ObjectReader (or more commonly a RevWalk wrapping one) and an
ObjectInserter. These code paths often assume that the ObjectReader
can read back any objects created by the ObjectInserter without
flushing. However, we previously had no way to enforce that constraint
programmatically, leading to hard-to-spot problems.
Provide a solution by exposing the ObjectInserter that created an
ObjectReader, when known. Callers can either continue passing both
objects and check:
reader.getCreatedFromInserter() == inserter
or they can just pass around ObjectReader and extract the inserter
when it's needed (checking that it's not null at usage time).
Change-Id: Ibbf5d1968b506f6b47030ab1b046ffccb47352ea
When CheckoutCommand or MergeCommand is called then not in all situation
the treewalks have been prepared to support clean/smudge filters. Fix
this
Bug: 491505
Change-Id: Iab5608049221c46d06812552ab97299e44d59e64
Repurpose RefDatabase#performsAtomicTransactions() slightly, to
indicate that the backend _supports_ atomic transactions, rather than
the current definition, which is that the backend always _uses_ atomic
transactions regardless of whether or not the caller actually wants
them. Allow BatchRefUpdate callers to turn off atomic transactions by
calling setAtomic(false). Defaulting to true means this is backwards
compatible.
Change-Id: I6df78d7df65ab147b4cce7764bd3101db985491c
Such hunks are identifiable by a zero value for "new start line". Prior
to the fix, JGit throws and ArrayIndexOutOfBoundsException on such
patches.
Change-Id: I4f3deb5e5f41a08af965fcc178d678c77270cddb
Signed-off-by: Jonathan Schneider <jkschneider@gmail.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Now if refs are unreadable when serving an upload pack the handler
will fail due to the actual underlying failure. Previously all wants
would be rejected as invalid because Repository.getAllRefs() returned
an empty map.
Testing this required a new subclass of InMemoryRepository so that
an IOException could be injected at the correct time.
Signed-off-by: Michael Edgar <adgar@google.com>
Change-Id: Iac708b1db9d0ccce08c4ef5ace599ea0b57afdc0
Change-Id: I5b3b7b0633354d5ccf0c6c320c0df9c93fdf8eeb
Signed-off-by: Ned Twigg <ned.twigg@diffplug.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
CommitCommand already provided a method to set the comment which should
be written into the reflog. The underlying RefUpdate class supported to
skip writing a reflog entry. But through the CommitCommand API it was
not possible to prevent writing a reflog entry. Fix this and allow
creating commits which don't occur in the reflog.
Change-Id: I193c53de71fb5958ea749c4bfa8360a51acc9b58
PackWriter.writeObject() can get into an infinite loop when corrupt
packs are present. When it finds a pack file with an object that can be
reused it calls DfsPackFile.copyAsIs(). If that method sees an invalid
CRC, it adds the object to the DfsPackFile's corrupt object list and
throws a CorruptObjectException, which it later catches as an
IOException and wraps in a
StoredObjectRepresentationNotAvailableException.
PackWriter.writeObjectImpl() catches that SORNAE and retries the
operation by calling DfsReader.selectObjectRepresentation(). But
currently that method returns the same object which was just seen to
be corrupt.
Change DfsPackFile.isCorrupt() from private to package private, and use
that method in DfsReader.selectObjectRepresentation() to filter out
corrupt objects.
The stack traces that show the problem are:
org.eclipse.jgit.errors.CorruptObjectException.<init>(CorruptObjectException.java:113)
org.eclipse.jgit.internal.storage.dfs.DfsPackFile.copyAsIs(DfsPackFile.java:624)
org.eclipse.jgit.internal.storage.dfs.DfsReader.copyObjectAsIs(DfsReader.java:491)
org.eclipse.jgit.internal.storage.pack.PackWriter.writeObjectImpl(PackWriter.java:1478)
org.eclipse.jgit.internal.storage.pack.PackWriter.writeObject(PackWriter.java:1455)
org.eclipse.jgit.internal.storage.dfs.DfsPackFile.getPackIndex(DfsPackFile.java:228)
org.eclipse.jgit.internal.storage.dfs.DfsReader.findAllFromPack(DfsReader.java:476)
org.eclipse.jgit.internal.storage.dfs.DfsReader.selectObjectRepresentation(DfsReader.java:455)
org.eclipse.jgit.internal.storage.pack.PackWriter.writeObjectImpl(PackWriter.java:1492)
org.eclipse.jgit.internal.storage.pack.PackWriter.writeObject(PackWriter.java:1455)
Change-Id: Iad7bbcaed1f11a6aa3b4f5af911a73a34c0fabfd
Signed-off-by: Terry Parker <tparker@google.com>
RepositoryCache has 2 methods to remove a repository from the cache but
they are never called when a repository is closed. Users of the cache
were expected to call one of those 2 methods but how could they have
called them at proper time without having visibility of the repository
usage count.
Ideally, I would have reworked the RepositoryCache to wrap any
repository it opens in a class that would be responsible to unregister
them from the cache when it's really closed, i.e. when usage counter
reaches 0. The problem preventing the wrapping solution is the
RepositoryCache.register method that allows to register an already
opened repository in the cache. Such repositories cannot be wrapped
because callers are still holding a reference on the unwrapped
repository.
Document that RepositoryCache.close method is removing the repository
from the cache as well as closing it and rework
RepositoryCache.unregister method to only remove the repository from the
cache. Use the latter to unregister repository when Repository.doClose
is getting executed.
Change-Id: Ia364816e4da8d7b6cfa72f10758ca31aa8a1f9db
Signed-off-by: Hugo Arès <hugo.ares@ericsson.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
When repositories are opened using the RepositoryCache, they are kept in
memory and when the repository usage counter reaches 0, the
Repository.close method is called which then calls close method on its
reference and object databases.
The problem is that RefDirectory.close method was a no-op and the
reference database was kept in memory. This problem is only happening
when opening a repository using the RepositoryCache because it never
evicts repositories, it's just calling the close method.
Change-Id: Iacb961de8e8b1f5b37824bf0d1a4caf4c6f1233f
Signed-off-by: Hugo Arès <hugo.ares@ericsson.com>
Repository has a usage counter that is initialized to 1 at
instantiation and this counter is decremented when Repository.close
method is called. There is also a Repository.incrementOpen method that
RepositoryCache uses to increment the usage count when it's returning a
repository that is already opened.
The problem was that RepositoryCache was incrementing the usage count
for repositories that it just opened or registered. The usage count was
2 when it should have been 1.
Incrementing usage count is now only be done for repository that are
served from the cache.
This bug is causing slow memory increase of our Gerrit server until the
server become slow. Even if the RepositoryCache is using SoftReference,
it seems that the JVM is not garbage collecting the repositories because
it's not yet on the edge of being out of memory.
To test this change, I replicated all repositories(11k) from Gerrit
master to one slave. The Gerrit master used memory after this test was
10GB without this change and 3.5GB with.
Change-Id: I86c7b36174e384f106b51fe92f306018fd1dbdf0
Signed-off-by: Hugo Arès <hugo.ares@ericsson.com>
* master:
Add config parameter gc.prunePackExpire for packfile expiration
In TestRepository, use a consistent clock
Change-Id: I7ac568e650fbd191e48a8f1a4068af72deb242e8
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
When checking out commits/branches JGit was triggering correctly
configured smudge filters. But when checking out paths (either from
index or from commits) JGit was not triggering smudge filters. Fix
CheckoutCommand to properly call filters.
Bug: 486560
Also-by: Pascal Krause <pascal.krausek@sap.com>
Change-Id: I5ff893054defe57ab12e201d901fe74e1376efea
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Implement the DIR_NO_GITLINKS setting with the same functionality
it provides in cGit.
Bug: 436200
Change-Id: I8304e42df2d7e8d7925f515805e075a92ff6ce28
Signed-off-by: Preben Ingvaldsen <preben@puppetlabs.com>
JGit's Garbage Collector is repacking relevant objects into new
packfiles and is afterwards deleting the now obsolete packfiles. But to
prevent problems caused by race conditions JGit was not deleting
packfiles when they are too young. The same mechanism as for loose
objects and the config parameter gc.pruneExpire was used.
But JGit was reusing the parameter gc.pruneExpire also for packfiles
which may cause a lot of filesystem consumption if gc.pruneExpire was
set to the default of 2 weeks. Only two weeks after packfile creation gc
was allowed to delete this packfile.
This change introduces a new config paramter gc.prunePackExpire with a
default of "1.hour". This parameter is used when packfiles are deleted.
Only packfiles older than the specified time can be deleted.
For loose objects the behaviour is not changed and only the old
parameter gc.pruneExpire is relevant.
Change-Id: I6209efb05678b15153bd22479dc13486907a44f8
This change replaced some tabs with spaces introduced in
change I8b3765713599e34f1411f9bbc7f575ec7c2384e0.
Change-Id: Ia5c23b38c9fbbb46f150e527347b61c64c8d9e87
Signed-off-by: Yuxuan 'fishy' Wang <fishywang@google.com>
With ignoreRemoteFailures set to true, we can ignore remote failures
(e.g. the branch of a project described in the manifest file does not
exist), skip that project and continue to the next one, instead of fail
the whole operation.
Change-Id: I8b3765713599e34f1411f9bbc7f575ec7c2384e0
Signed-off-by: Yuxuan 'fishy' Wang <fishywang@google.com>
This commit introduces a FileModeStrategy to
the FileTreeIterator class. This provides a way to
allow different modes of traversing a file tree;
for example, to control whether or not a nested
.git directory should be treated as a gitlink.
Bug: 436200
Change-Id: Ibf85defee28cdeec1e1463e596d0dcd03090dddd
Signed-off-by: Preben Ingvaldsen <preben@puppetlabs.com>
Allow access to the ObjectId of a DirCacheTree if known for low level
integration code.
Change-Id: I6f05b10c9ac781f5e8b38af4a19e653313c91fa8
Signed-off-by: Philipp Marx <smigfu@googlemail.com>
TreeWalk provides the new method getEolStreamType. This new method can
be used with EolStreamTypeUtil in order to create a wrapped InputStream
or OutputStream when reading / writing files. The implementation
implements support for the git configuration options core.crlf, core.eol
and the .gitattributes "text", "eol" and "binary"
CQ: 10896
Bug: 486563
Change-Id: Ie4f6367afc2a6aec1de56faf95120fff0339a358
Signed-off-by: Ivan Motsch <ivan.motsch@bsiag.com>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
The reason is that URIish(URL) and URIish(String) make different parsing
of path / rawPath with regard to drive letters. /C:/... for URL and
C:/... for String. This patch fixes the issue.
Change-Id: I8e2013fff30b7bb198ff733c038e21366667b8a0
Signed-off-by: Ivan Motsch <ivan.motsch@bsiag.com>
Capture the internal "want X not valid" state as a specific subclass
of PackProtocolException, allowing this to be more easily identified
in server stack traces and wrapper application code.
Change-Id: I4b1adb7497f396432da420b0f600ad25a261f912
If the client sends a SHA-1 that the server does not recognize echo
this back to the client with an explicit error message instead of
the generic "internal server error".
This was always the intent of the implementation but it was being
dropped on smart HTTP due to the UploadPackServlet catching the
PackProtocolException, discarding the buffered message UploadPack
meant to send, and sending along a generic message instead.
Change-Id: I8d96b064ec655aef64ac2ef3e01853625af32cd1
* stable-4.2:
Don't use deprecated LockFile constructor
Fix warnings about unchecked conversion of MergeResult
MockServletConfig: Fix warning about unchecked conversion of Enumeration
HugeFileTest: Make Git a class member and open in try-with-resource
Suppress "unchecked cast" warnings related to UploadPackFactory.DISABLED
DiffAlgorithms: Fix warnings about variable hiding
DirCacheBasicTest: Open ObjectInserter.Formatter in try-with-resource
DirCacheBuilderIteratorTest: Open TreeWalk in try-with-resource
DirCacheCGitCompatabilityTest: Open TreeWalk in try-with-resource
DirCacheCheckoutMaliciousPathTest: Open Git and RevWalk in t-w-r
DirCacheIteratorTest: Open TreeWalk instances in try-with-resource
ForPathTest: Open TreeWalk in try-with-resource
GitConstructionTest: Open Git instance in try-with-resource
IndexDiffTest: Open Git instances in try-with-resources
ManifestParserTest: Don't use deprecated StringBufferInputStream
InMemoryRepository: Remove unused RevWalk from batch method signature
IndexModificationTimesTest: Open Git instances in try-with-resource
InterIndexDiffFilterTest: Open TreeWalk in try-with-resource
LockFileTest: Open Git instance in try-with-resource
JGit v4.1.2.201602141800-r
MergeCommandTest: Use JUnit's assume to check preconditions
MergeCommandTest: Open Git instances in try-with-resource
Change-Id: Ie5dba6b9132a29e86958a04fa2b76465bcd2c6b5
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
The RevWalk given in the arguments is not used. According to the
comment at the top of the method, a new RevWalk is intentionally
used in the implementation.
Remove the unused argument.
Change-Id: Iec81a1341d5bf377801475845b96a465753096ef
Signed-off-by: David Pursehouse <david.pursehouse@sonymobile.com>
Attributes MacroExpander implements macros used in git attributes. This
is implemented inside the TreeWalk using a lazy created MacroExpander.
In addition, the macro expander caches the global and info attributes
node in order to provide fast merge of attributes.
Change-Id: I2e69c9fc84e9d7fb8df0a05817d688fc456d8f00
Signed-off-by: Ivan Motsch <ivan.motsch@bsiag.com>
* stable-4.2:
RepoProject: Fix warnings about variable hiding
RepoTest: Open Git in try-with-resources
RepositoryResolveTest: Open Git in try-with-resource
RepositoryTestCase: Open autocloseable types in try-with-resource
ResetCommandTest: Use Git member in testHardResetAfterSquashMerge
ResolveMergerTest: Open Git in try-with-resource
RevCommitListTest: Open Git and RevWalk in try-with-resource
RevCommitParseTest: Open ObjectInserter.Formatter in try-with-resource
RevObjectTest: Open RevWalk in try-with-resource
RevTagParseTest: Open ObjectInserter.Formatter in try-with-resource
RevertCommandTest: Open Git in try-with-resource
SquashMessageFormatterTest: Open git in try-with-resource
StatusCommandTest: Open Git in try-with-resource
SubmoduleAddTest: Open Git in try-with-resource
SymlinksTest: Open git and TreeWalk in try-with-resource
T0003_BasicTest: Open autocloseable types in try-with-resource
TextHashFunctions: Fix warnings about variable hiding
TreeFilterTest: Open TreeWalk in try-with-resource
TreeWalkJava7Test: Open TreeWalk in try-with-resource
Fix diff for added and removed submodule
Change-Id: If3ecc63f6dfac55474d3c1dd2f4105371f3d24fb
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
* stable-4.1:
Fix diff for added and removed submodule
Change-Id: I37dd71ed19b06e9bbcffe37370081ab875c6d8d4
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Method parameter names were hiding class members of the same
name.
Change-Id: I182f2715894ac4259b09a371cb4e0eb24f52518a
Signed-off-by: David Pursehouse <david.pursehouse@sonymobile.com>
The disableSslVerify method will be used in the follow up change.
Change-Id: Ie00b5e14244a9a036cbdef94768007f1c25aa8d3
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Implement LfsProtocolServlet handling the "Git LFS v1 Batch API"
protocol [1]. Add a simple file system based LFS content store and the
debug-lfs-store command to simplify testing.
Introduce a LargeFileRepository interface to enable additional storage
implementation while reusing the same protocol implementation.
At the client side we have to configure the lfs.url, specify that
we use the batch API and we don't use authentication:
[lfs]
url = http://host:port/lfs
batch = true
[lfs "http://host:port/lfs"]
access = none
the git-lfs client appends the "objects/batch" to the lfs.url.
Hard code an Authorization header in the FileLfsRepository.getAction
because then git-lfs client will skip asking for credentials. It will
just forward the Authorization header from the response to the
download/upload request.
The FileLfsServlet supports file content storage for "Large File
Storage" (LFS) server as defined by the Github LFS API [2].
- upload and download of large files is probably network bound hence use
an asynchronous servlet for good scalability
- simple object storage in file system with 2 level fan-out
- use LockFile to protect writing large objects against multiple
concurrent uploads of the same object
- to prevent corrupt uploads the uploaded file is rejected if its hash
doesn't match id given in URL
The debug-lfs-store command is used to run the LfsProtocolServlet and,
optionally, the FileLfsServlet which makes it easier to setup a
local test server.
[1]
https://github.com/github/git-lfs/blob/master/docs/api/http-v1-batch.md
[2] https://github.com/github/git-lfs/tree/master/docs/api
Bug: 472961
Change-Id: I7378da5575159d2195138d799704880c5c82d5f3
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Signed-off-by: Sasa Zivkov <sasa.zivkov@sap.com>
The Large File Storage extension specified by GitHub [1] uses SHA-256 to
compute the ID of large files stored by the extension. Hence implement a
SHA-256 abstraction similar to the SHA-1 abstraction used by JGit.
[1] https://git-lfs.github.com/
Bug: 470333
Change-Id: I3a95954543c8570d73929e55f4a884b55dbf1b7a
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Since If13f7b406, submodule difference are shown as a hunk. The issue
was that added and removed submodule were considered as Edit.REPLACE
instead of Edit.INSERT and Edit.DELETE in the DiffFormatter result.
Change-Id: I4330c2aa3f10e29d7d6b0b2e5286e59293a06239
Signed-off-by: Hugo Arès <hugo.ares@ericsson.com>
* stable-4.2:
DirCacheCheckoutTest: Open Git and TreeWalk in try-with-resource
CommitCommand: Remove declaration of unthrown exception
Branch: Fix variable hiding warning
ApplyCommandTest: Open Git in try-with-resource
PackFileTest: Open ObjectInserter.Formatter in try-with-resource
Change-Id: I8484b10fad5a4c35fcfaedc1cdf8ccf97471618e
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Repo manifest file allows revision attribute in <remote> tag. This change
teaches JGit to read that information.
Change-Id: I1c878a2505b9d09fa09fbd404a119b71f2fb8fdb
Signed-off-by: Yuxuan 'fishy' Wang <fishywang@google.com>
Monitoring progress of merges can be useful for users for large
repositories or complex merge processes that take some time.
This enables setting a monitor. Existing merge implementations in jgit
do not yet report progress if a monitor is set. This will be added in a
later change.
Change-Id: I17b978b3fc91750dd88649638b90a46820a0877c
Signed-off-by: Laurent Delaigue <laurent.delaigue@obeo.fr>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
When reading a bundle file, commit messages who's oneline format is
longer than 982 characters caused JGit to treat subsequent text in
the commit as a SHA, then throw a TransportException because it's
not a valid SHA.
Now the readLine method will read all the way to the end of the
line, not just the first 1024 characters of it.
Change-Id: If15b491aa9a1e4fd9b8bbed2dd9e6be47a64ccb7
Signed-off-by: Chris Gavin <chris@chrisgavin.me>
Monitoring progress of merges can be useful for users for large
repositories or complex merge processes that take some time.
This enables setting a monitor. Existing merge implementations in jgit
do not yet report progress if a monitor is set. This will be added in a
later change.
Change-Id: I17b978b3fc91750dd88649638b90a46820a0877c
Signed-off-by: Laurent Delaigue <laurent.delaigue@obeo.fr>
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
When reading a bundle file, commit messages who's oneline format is
longer than 982 characters caused JGit to treat subsequent text in
the commit as a SHA, then throw a TransportException because it's
not a valid SHA.
Now the readLine method will read all the way to the end of the
line, not just the first 1024 characters of it.
Change-Id: If15b491aa9a1e4fd9b8bbed2dd9e6be47a64ccb7
Signed-off-by: Chris Gavin <chris@chrisgavin.me>
* changes:
debug-rebuild-ref-tree: Copy HEAD into RefTree
debug-rebuild-ref-tree: Add --enable flag to turn the database on
RefTreeDatabase: Allow ORIG_HEAD, etc. on non-bare repositories
RefTreeDatabase: Expose bootstrap refs in getAdditionalRefs
* changes:
daemon: Add --ketch=LEADER flag
Ketch: Intercept push and route it through Ketch
Ketch: Simple map of running leader instances
Ketch: Basic replication system
This form of addTree() does not parse any objects and cannot throw the
declared checked exception. Callers are being forced to try-catch
CorruptObjectException that cannot occur when the iterator instance
has already been constructed.
Change-Id: Id338035302903bab81569d1576eab063eee0885a
ResetCommand was not returning the updated ref as a result of the call()
method. Since the ResetCommand is always updating the same ref (HEAD)
this should always be the HEAD ref.
Bug: 440750
Change-Id: I7974975c3ab05e68c208384e69cf0692ded6e8db
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
CommitCommand should allow to specify whether empty commits (commits
having the same tree as the sole predecessor commit) are allowed or not.
Similar to native git's "--allow-empty" flag.
The defaults differ between JGit and native git even after this change.
When not specifying paths then by default JGit allows to create empty
commits while native git does not. It would be API breaking to change
this now.
Bug: 460301
Change-Id: I88feb0c3ffb2c686b1d0594e669729b065cda4cb
Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Experimental flag to turn on the KetchLeader within this daemon JVM.
This is a manually elected leader process, set from the command line.
Remote followers for each repository are configured per-repository
using remote sections with ketch-type = FULL. For example:
Manually elected leader's $GIT_DIR/config:
[ketch]
name = A
[remote "A"]
ketch-type = FULL
[remote "B"]
url = git://127.0.0.1:9421/sample.git
ketch-type = FULL
[remote "C"]
url = git://127.0.0.1:9422/sample.git
ketch-type = FULL
Replica B and C daemons:
git daemon \
--export-all \
--enable=receive-pack \
--listen=127.0.0.1 --port=9421 \
--base-path=$HOME/ketch_test/follower_one \
$HOME/ketch_test/follower_one &
git daemon \
--export-all \
--enable=receive-pack \
--listen=127.0.0.1 --port=9422 \
--base-path=$HOME/ketch_test/follower_two \
$HOME/ketch_test/follower_two &
Change-Id: I165f85970a77e16b5263115290d685d8a00566f5
Git Ketch is a multi-master Git repository management system. Writes
are successful only if a majority of participant servers agree. Acked
writes are durable against server failures as a majority of the
participants store all required objects.
Git Ketch is modeled on the Raft Consensus Algorithm[1]. A ketch
sailing vessel is faster and more nimble than a raft. It can also
carry more source codes.
Git Ketch front-loads replication costs, which vaguely resembles a
ketch sailing vessel's distinguishing feature of the main mast on the
front of the ship.
[1] https://raft.github.io/
Change-Id: Ib378dab068961fc7de624cd96030266660b64fb4
This could have only happened during the getBytes call. Instead, use
Constants.encode, which is a non-throwing implementation.
This change is binary compatible with existing code compiled against
older versions of JGit, although it might break compilation of
previously compiling code due to dead catch blocks.
Change-Id: I191fec5cac718657407230de141440e86d0151fb
After creating a Transport instance callers should always call
its close() method. Use AutoCloseable to document this idiom
and allow use of try-with-resources.
Change-Id: I0c6ff3e39ebecdd7a028dbcae1856a818937b186
* changes:
Ignore API errors about Tree, TreeEntry, FileTreeEntry and friends being removed
Revert "Revert "Remove deprecated Tree, TreeEntry, FileTreeEntry and friends""
Mike Williams