Marco Miller [Mon, 29 Jul 2019 16:32:07 +0000 (12:32 -0400)]
Upgrade maven-site-plugin to 3.8.2
Change-Id: I79a3032e2ff7af14a41e6eb2618d2b1e22d1fde2 Signed-off-by: Marco Miller <marco.miller@ericsson.com> Signed-off-by: David Pursehouse <david.pursehouse@gmail.com>
* changes:
sshd: add missing javadoc in SshTestGitServer
sshd: shared reference in JGitClientSession must be volatile
sshd: correct the protocol version exchange
Jonathan Tan [Wed, 3 Apr 2019 17:52:48 +0000 (10:52 -0700)]
UploadPack: move writing of "packfile" header
In a subsequent patch, in some cases, PackWriter#writePack will be
responsible for both the "packfile-uris" and "packfile" sections,
meaning that (in these cases) it must write the "packfile" section
header itself.
In preparation for that patch, move the writing of the "packfile"
section header closer to the invocation of PackWriter#writePack when the
entire fetch response is configured to use the sideband. This means that
"packfile" is written *after* objects are counted (and progress messages
sent to the client in sideband 2) when the "sideband-all" feature is
used (whether "packfile-uris" is used or not), and written *before*
objects are counted otherwise.
Having code to write "packfile" in two places is unfortunate but
necessary. When "sideband-all" is not used, object counting has to
happen after "packfile" is written, because "packfile" activates the
sideband that allows counting progress to be transmitted. When
"packfile-uris" is used, object counting has to happen before "packfile"
is written, because object counting determines whether to send
"packfile-uris" or "packfile". When "sideband-all" is used but
"packfile-uris" is not used, either way works; this commit uses
"packfile-uris" behavior in this case.
Also make the naming of the sideband-activating methods in PacketLineOut
more consistent.
Change-Id: Ifbfd26cc26af10c41b77758168833702d6983df1 Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Thomas Wolf [Wed, 24 Jul 2019 18:36:51 +0000 (20:36 +0200)]
JGit pgm tests must quote paths on the command line
The JGit pgm tests parse the command line internally using Linux
semantics, treating '\' as an escape. File paths therefore must
be quoted, otherwise Windows paths are destroyed. See the
attachment[1] on bug 548598.
Jonathan Tan [Wed, 3 Apr 2019 00:00:12 +0000 (17:00 -0700)]
UploadPackTest: use Consumer to set up UploadPack
Use a Consumer instead of several nullable variables to further
configure UploadPack. This is in preparation for a test in a subsequent
patch needing further customization of the UploadPack object before
invoking it.
Change-Id: I074dff92c711a5ba74558bb4b06c42c115fb9b7f Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Jonathan Tan [Tue, 22 Jan 2019 21:17:12 +0000 (13:17 -0800)]
Support "sideband-all" in protocol v2 fetch
Allow the client to specify "sideband-all" in a fetch v2 request,
indicating that the whole response is to be multiplexed (with a sideband
indicator on every non-flush and non-delim pkt) instead of only the
packfile being multiplexed. This allows, for example, progress messages
to be sent at any point in the response.
This implements the "sideband-all" feature documented in
Documentation/technical/protocol-v2.txt in Git.
Change-Id: I3e7f21c88ff0982b1b7ebb09c9ad6c742c4483c8 Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Jonathan Tan [Tue, 22 Jan 2019 21:17:12 +0000 (13:17 -0800)]
UploadPackTest: refactor capability config test
UploadPackTest.java contains tests that check behavior when
"allowfilter" and "allowrefinwant" are not set, are set, and are not set
but the client insists on using them anyway. Because another capability
is to be included in a subsequent patch, refactor the common code in
these tests.
Remove setBoolean calls with "false", as they are no-ops.
Also take the opportunity to eliminate the overspecification of the
"fetch=" line returned by the capability advertisement.
Change-Id: I289bbd11c902a513cd8d53bc34767e61ebbd5f17 Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Thomas Wolf [Sat, 13 Jul 2019 08:57:11 +0000 (10:57 +0200)]
sshd: correct the protocol version exchange
RFC 4253 section 4.2 allows an ssh server to send additional lines
before its server identification string. Apache MINA sshd enforces
for these lines the constraints specified for the server identification
line, too: no NUL characters and not longer than 255 characters. That
is too strict. RFC 4253 doesn't mandate this, and it also doesn't
make sense given the rationale for these lines in RFC 4253: a TCP
wrapper may not be aware of SSH restrictions, and may not adhere to
these constraints.
Be more lenient when parsing the server's protocol version. Allow
NULs and longer lines in the preamble, and also handle line endings
more leniently. Only enforce the restrictions for the actual server
identification line.
Bug: 545939
Change-Id: I75955e9d8a8daef7c04fc0f39539c2ee93514e1c Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
Masaya Suzuki [Tue, 5 Mar 2019 03:01:39 +0000 (19:01 -0800)]
Do not ignore secondary failure
When another exception is thrown while handling another exception, that
exception can be attached to the original exception since Java 7
(Throwable#getSuppressed). Attach the secondary exception to the
original exception instead of throwing it away.
Masaya Suzuki [Fri, 14 Jun 2019 02:11:40 +0000 (19:11 -0700)]
Add assertThrows
This is a new assertion that will be introduced in JUnit 4.13. Unlike
ExpectedException rule, this makes it easy to test other aspects of the
thrown exception, such like ServiceMayNotContinueException's status
code. Introduce this as before making changes to UploadPackTest more.
Dave Borowitz [Mon, 3 Jun 2019 21:03:49 +0000 (14:03 -0700)]
RevWalk: Add a setFirstParent that mimics C git's --first-parent
RevWalk does not currently provide a --first-parent equivalent and the
feature has been requested.
Add a field to the RevWalk class to specify whether walks should
traverse first parents only. Modify Generator implementations to support
the feature.
Change-Id: I4a9a0d5767f82141dcf6d08659d7cb77c585fae4 Signed-off-by: Dave Borowitz <dborowitz@google.com> Signed-off-by: Alex Spradlin <alexaspradlin@google.com>
Jonathan Tan [Thu, 23 May 2019 20:28:53 +0000 (13:28 -0700)]
UploadPack: support lazy fetches if no bitmap
If a client clones with "--filter=blob:none", the checkout that "git
clone" automatically does causes the client to fetch all blobs at HEAD.
When fetching from a non-bitmapped repository, this will fail if an
object walk is ever needed, because JGit currently rejects such requests
- see the commit message of d3021788d2 ("Use bitmaps for non-commit
reachability checks", 2017-11-10) for more information.
Rejecting such requests in the absence of bitmaps is probably
overzealous: it is true that the server would prefer to have bitmaps in
this case, but there might be a small proportion of repos (for example,
very small repos or newly created ones) that do not have bitmaps, yet
the server would still like to have partial clones for them.
So, allow such requests, performing the object walk reachability check
if necessary. Limit this to servers with "uploadpack.allowFilter"
configured, so that servers wanting to support partial clone have this
functionality, and servers that do not support partial clone do not have
to pay the object walk reachability check cost.
Change-Id: I51964bafec68696a799625d627615b4f45ddbbbf Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Michael Keppler [Fri, 21 Jun 2019 15:39:20 +0000 (17:39 +0200)]
Fix bundle localization of Apache SSH bundle
The placeholders in manifest and plugin.properties did not match. To
avoid similar issues, all placeholders have been changed to
Bundle-Vendor and Bundle-Name now.
Bug:548503
Change-Id: Ibd4b9bc237b323e614506b97e5fbc99416365040 Signed-off-by: Michael Keppler <Michael.Keppler@gmx.de>
Matthias Sohn [Tue, 18 Jun 2019 00:20:13 +0000 (02:20 +0200)]
[Java 11] Update mockito to 2.23.0
This update is required for Java 11.
Update
- org.mockito to 2.23.0.v20190527-1420
- net.bytebuddy.byte-buddy to 1.9.0.v20181107-1410
- net.bytebuddy.byte-buddy-agent to 1.9.0.v20181106-1534
Matthias Sohn [Mon, 17 Jun 2019 23:48:10 +0000 (01:48 +0200)]
[Java 9] Fix GitDateFormatterTest
In JDK 9 and later, the default locale data uses data derived from the
Unicode Consortium's Common Locale Data Repository (CLDR). So there are
changes with respect to some locales.
For example the short date-time format is ‹{1}, {0}› in the CLDR locale,
as opposed to {1} {0} in the JRE locale data.
See: https://bugs.openjdk.java.net/browse/JDK-8206961
See: https://www.unicode.org/cldr/charts/29/by_type/date_&_time.gregorian.html#1141bf54834c4261
Change-Id: I7535821e8ecd8702a95db8732cbbf3a4a7385eca Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
David Ostrovsky [Mon, 17 Jun 2019 21:28:52 +0000 (23:28 +0200)]
[Java 11] Remove dependency on javax.xml.bind package
Replace javax.xml.bind.DatatypeConverter, that is not available any
more in Java 11 and later with Hex utility from non optional Bouncy
Castle library.
Bug: 540790
Change-Id: I9903c00ecc1a434e9795b8ba9267f02628fdc0e9 Signed-off-by: David Ostrovsky <david@ostrovsky.org>
Pavel Flaška [Fri, 24 Nov 2017 13:55:28 +0000 (14:55 +0100)]
Walk tag chains for --include-tag options
When cloning repository with --single-branch option, tag chains are not
packed and pack file is broken in some cases.
Typical test-case:
git tag -a test_tag <commit-id>
git tag -a test_prev_tag test_tag
git tag -d test_tag
git clone --single-branch <repository>
fatal: did not receive expected object <test_tag_id>
The reason for that is missing object for original test_tag reference,
which was deleted.
Problem description:
When pack-objects is given --include-tag, it peels each tag reference
down to a commit. If the commit is prepared to be packed, we we have to
include such tag too. The problem is when the tag points to through some
chain of other tag to commit. Then, the inner tags are not added leading
to broken pack.
Fix:
When going to commit, we have to check and add any of the tags on the
way (if they were not selected, which may happen with --single-branch
option).
Change-Id: I1682d4a2c52d674f90a1b021e0f6c3524c5ce5bc Signed-off-by: Pavel Flaška <Pavel.Flaska@gmail.com> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Matthias Sohn [Thu, 13 Jun 2019 00:44:31 +0000 (02:44 +0200)]
Merge branch 'stable-5.4'
* stable-5.4:
Prepare 5.4.1-SNAPSHOT builds
JGit v5.4.0.201906121030-r
JGit v5.4.0.201906120450-r
Increase severity of OperatorPrecedence to ERROR, and fix instances
PackWriter: Make internal class static
Prepare 5.3.3-SNAPSHOT builds
JGit v5.3.2.201906051522-r
Prepare 5.1.9-SNAPSHOT builds
JGit v5.1.8.201906050907-r
Test detecting modified packfiles
Enhance fsTick() to use filesystem timer resolution
Add debug trace to measure time needed to open pack index
Extend FileSnapshot for packfiles to also use checksum to detect changes
Wait opening new packfile until it can't be racy anymore
Avoid null PackConfig in GC
Add FileSnapshot test testing recognition of file size changes
Capture reason for result of FileSnapshot#isModified
Skip FileSnapshotTest#testSimulatePackfileReplacement on Windows
Tune max heap size for tests
Fix FileSnapshotTest.testNewFileNoWait() to match its javadoc
ObjectDirectory: fix closing of obsolete packs
Update API filters for methods added to fix bugs
Bazel: Increase severity of most error-prone checks to ERROR
Enable error-prone checks by default
Add bazel options to align with gerrit's
Include filekey file attribute when comparing FileSnapshots
Measure file timestamp resolution used in FileSnapshot
Fix FileSnapshot's consideration of file size
Fix API problem filters
Fix API problem filters
Matthias Sohn [Tue, 11 Jun 2019 21:09:19 +0000 (23:09 +0200)]
Merge branch 'master' into stable-5.4
* master:
Handle missing "ours" stage in WorkingTreeIterator.hasCrLfInIndex()
Config: Add helper method to check for empty value
ErrorProne: Increase severity of FutureReturnValueIgnored to ERROR
FS_Win32: Add missing parentheses on if-blocks
Upgrade spotbugs to 3.1.12
Abbreviated{Long}ObjectId: Make operator precedence explicit
GC: Update TODO comments
FS_POSIX: Fix reference comparison of Boolean.FALSE
Increase bazel timeout for long running tests
Use bazelisk to switch between used bazel version
Bump minimum Bazel version to 0.26.1
Bazel: Bump skylib library version to 0.8.0
Retry loading config when locked by another process
Make pull --rebase on an unborn branch do a checkout
Warn if configured cookie file is missing
Handle escaped CR-LF in git config files
DescribeCommand: use glob match instead of path match
Fix off-by-one error in RebaseTodoFile when reading a todo file
Consistently use "!isEmpty()" to detect non-empty list
TransportHttp: Check for non-empty list with "!isEmpty()" rather than
"size() > 0"
TransportHttp: Fix comparison of size with ">= 0"
NetscapeCookieFileTest: Split HttpCookiesMatcher to own class
Bazel: Add missing dependency on mockito for TransportHttpTest
Determine hard-linking and nlink support per FileStore
Support reading and writing cookies.
Repository: Add getIdentifier() method to avoid instanceof operator
Update to Orbit R20190602212107
PacketLineIn: Deprecate the END constant
PacketLineIn: Add an iterator over strings in the input stream
Replace most usages of PacketLineIn.END with PacketLineIn.end()
PacketLineIn: Deprecate DELIM constant
Replace trivial reference comparison of PacketLineIn.{DELIM,END}
PacketLineIn: Rename isDelim to isDelimiter
ProtocolV2ParserTest: Fix typo in comment
Upgrade Bouncy Castle to 1.61
Update to Orbit R20190531194818 and rollback update to Ant 1.10.6
cli: Add the --always option to describe
DescribeCommand: Support the "always" option
cli: Add the --tags option to describe
DescribeCommand: Consistenly omit the default value
Remove excess blank line in FileUtilsTest
PacketLineIn: Add helper methods to check for END and DELIM
UploadPackTest: Rename variable to avoid hiding class member
UploadPackTest: Add missing <> operator on instantiation of ArrayList
BitmapCalculator: javadoc fixes
RevWalkUtils: add progress callback to findBranchesReachableFrom
Upgrade maven-source-plugin to 3.1.0
Upgrade maven-jar-plugin to 3.1.2
Upgrade jacoco-maven-plugin to 0.8.4
BitmapCalculator and its test: add missing license header
RevWalk: new method createReachabilityChecker()
Change-Id: I4d76c7c0dbe6411c842f3468b709f7df51789c08 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Thomas Wolf [Tue, 11 Jun 2019 11:16:18 +0000 (13:16 +0200)]
Handle missing "ours" stage in WorkingTreeIterator.hasCrLfInIndex()
In a delete-modify conflict with the deletion as "ours" there may be
no stage 2 in the index. Add appropriate null checks. Add a new test
for this case, and verify that the file gets added with a single LF
after conflict resolution with core.autocrlf=true. This matches the
behavior of canonical git for this case.
Bug: 547724
Change-Id: I1bafdb83d9b78bf85294c78325e818e72fae53bc Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
The TODO comments say "in 5.0", but 5.0 was already released without
resolving them. Remove "in 5.0" on the assumption that the mentioned
improvements still need to be done at some point.
Change-Id: I3eb429803e2266de3fc490e1f3912991c08aa1ad Signed-off-by: David Pursehouse <david.pursehouse@gmail.com>
Matthias Sohn [Mon, 10 Jun 2019 21:35:20 +0000 (23:35 +0200)]
Increase bazel timeout for long running tests
EolRepositoryTest and GcCommitSelectionTest timed out frequently when
running unit tests using bazel with the default timeout "moderate"
(300s). Increase timeout of these tests to "long" (900s).
Change-Id: I43588cf950f55b50f868d9fe9c66d22bd428a54c Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Michael Keppler [Sun, 29 Jul 2018 05:53:11 +0000 (07:53 +0200)]
Retry loading config when locked by another process
When loading the config, a FileNotFoundException may occur if the file
exists but cannot be read (see [1]). This is the case on Windows with a
virus scanner checking the file. Therefore if the file exists and that
exception is thrown, retry multiple times, similar to how this was
already implemented for IOException.
Thomas Wolf [Sun, 3 Mar 2019 23:07:40 +0000 (00:07 +0100)]
Make pull --rebase on an unborn branch do a checkout
A merging pull on an unborn branch was already supported. But a
rebasing pull failed. If the user has pull.rebase = true in his
user config, the pull would try to rebase. Rebasing needs a parent
commit, though. Native git handles this case:
Thomas Wolf [Fri, 5 Apr 2019 19:33:14 +0000 (21:33 +0200)]
Handle escaped CR-LF in git config files
Canonical git treats CR-LF in config files as LF.[1][2] JGit does so,
too, except when escaped as a line continuation. Correct this and
treat the sequence \-CR-LF as a line continuation.
Bug: 545850
Change-Id: I51e7378a22c21b3baa3701163c423d04c900af5a Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Thomas Wolf [Sun, 21 Apr 2019 19:23:30 +0000 (21:23 +0200)]
Fix off-by-one error in RebaseTodoFile when reading a todo file
Commit messages of length 1 were not read. 'lineEnd' is the offset
of the last character in the line before the terminating LF or CR-LF,
and 'nextSpace' is actually the offset of the character _after_ the
next space. With a one-character commit message, nextSpace == lineEnd.
The code also assumes the commit message to be optional, but actually
failed in that case because it read beyond the line ending. Fix that,
too.
Add a test case for reading a todo file.
Bug: 546245
Change-Id: I368d63615930ea2398a6230e756442fd88870654 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
Thomas Wolf [Sat, 18 May 2019 21:48:54 +0000 (23:48 +0200)]
Determine hard-linking and nlink support per FileStore
It's quite possible that JGit can use the hard-linking mechanism
for atomic file creation on some volumes but not on others.
Ultimately it depends on the file systems on the mounted volumes.
Cache the information per FileStore instead of using a single
global flag. Also catch FileSystemException, it may be thrown
if the operating system reports a failure. The previously caught
AccessDeniedException is a sub-class of FileSystemException.
Bug: 547332
Change-Id: I1ef672b3468b0be79e71674344f16f28f9d11ba1 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
Matthias Sohn [Thu, 6 Jun 2019 09:34:56 +0000 (11:34 +0200)]
Merge branch 'stable-5.3' into stable-5.4
* stable-5.3:
Prepare 5.3.3-SNAPSHOT builds
JGit v5.3.2.201906051522-r
Prepare 5.1.9-SNAPSHOT builds
JGit v5.1.8.201906050907-r
Test detecting modified packfiles
Enhance fsTick() to use filesystem timer resolution
Add debug trace to measure time needed to open pack index
Extend FileSnapshot for packfiles to also use checksum to detect changes
Wait opening new packfile until it can't be racy anymore
Avoid null PackConfig in GC
Add FileSnapshot test testing recognition of file size changes
Capture reason for result of FileSnapshot#isModified
Skip FileSnapshotTest#testSimulatePackfileReplacement on Windows
Tune max heap size for tests
Fix FileSnapshotTest.testNewFileNoWait() to match its javadoc
ObjectDirectory: fix closing of obsolete packs
Update API filters for methods added to fix bugs
Bazel: Increase severity of most error-prone checks to ERROR
Enable error-prone checks by default
Add bazel options to align with gerrit's
Include filekey file attribute when comparing FileSnapshots
Measure file timestamp resolution used in FileSnapshot
Fix FileSnapshot's consideration of file size
Fix API problem filters
Fix API problem filters
Change-Id: Ie1c7e4752661aec9e1754660934921224e2408eb Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
David Ostrovsky [Thu, 16 May 2019 15:06:57 +0000 (17:06 +0200)]
Repository: Add getIdentifier() method to avoid instanceof operator
This change is needed to implement permission aware ref database in
Gerrit: [1], that is a pre-requisite to re-enable Git v2 protocol in
Gerrit: [2].
Background: Last year Git v2 protocol was enabled in Gerrit. The fact,
that JGit layer was not calling ref advertise filter for Git v2
protocol, introduced security vulnerability.
The lesson learned from this security incident: Gerrit should not rely
on ref advertise filter being called by JGit to implement crictical
security checks. Instead, the idea is to use the same approach as
currently used by Google's internal code on googlesource.com that
didn't suffer from this vulnerability: provide a custom repository to
JGit. The repository provides a RefDatabase that is permission-aware
and will only ever return refs that the user has access to.
However, due to hard coded instanceof operator usages in JGit code
base, some tests in Gerrit are failing with: [1] in place. This change
addresses this problem.
Change-Id: I67c0f53ca33b149442e7ee3e51910d19e3f348d5 Signed-off-by: David Ostrovsky <david@ostrovsky.org> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Matthias Sohn [Wed, 5 Jun 2019 19:18:21 +0000 (21:18 +0200)]
Merge branch 'stable-5.2' into stable-5.3
* stable-5.2:
Prepare 5.1.9-SNAPSHOT builds
JGit v5.1.8.201906050907-r
Test detecting modified packfiles
Enhance fsTick() to use filesystem timer resolution
Add debug trace to measure time needed to open pack index
Extend FileSnapshot for packfiles to also use checksum to detect changes
Wait opening new packfile until it can't be racy anymore
Avoid null PackConfig in GC
Add FileSnapshot test testing recognition of file size changes
Capture reason for result of FileSnapshot#isModified
Skip FileSnapshotTest#testSimulatePackfileReplacement on Windows
Tune max heap size for tests
Fix FileSnapshotTest.testNewFileNoWait() to match its javadoc
ObjectDirectory: fix closing of obsolete packs
Include filekey file attribute when comparing FileSnapshots
Measure file timestamp resolution used in FileSnapshot
Fix FileSnapshot's consideration of file size
Fix API problem filters
Fix API problem filters
Matthias Sohn [Wed, 5 Jun 2019 13:43:29 +0000 (15:43 +0200)]
Merge branch 'stable-5.1' into stable-5.2
* stable-5.1:
Prepare 5.1.9-SNAPSHOT builds
JGit v5.1.8.201906050907-r
Test detecting modified packfiles
Enhance fsTick() to use filesystem timer resolution
Add debug trace to measure time needed to open pack index
Extend FileSnapshot for packfiles to also use checksum to detect changes
Wait opening new packfile until it can't be racy anymore
Avoid null PackConfig in GC
Add FileSnapshot test testing recognition of file size changes
Capture reason for result of FileSnapshot#isModified
Skip FileSnapshotTest#testSimulatePackfileReplacement on Windows
Tune max heap size for tests
Fix FileSnapshotTest.testNewFileNoWait() to match its javadoc
ObjectDirectory: fix closing of obsolete packs
Include filekey file attribute when comparing FileSnapshots
Measure file timestamp resolution used in FileSnapshot
Fix FileSnapshot's consideration of file size
Fix API problem filters
Change-Id: I3ac77bfa03f7436de12ab86e1bba29afee5ccd01 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Test that JGit detects that packfiles have changed even if they are
repacked multiple times in one tick of the filesystem timer.
Test that this detection works also when repacking doesn't change the
length or the filekey of the packfile. In this case where a modified
file can't be detected by looking at file metadata JGit should still
detect too fast modification by racy git checks and trigger rescanning
the pack list and consequently rereading of packfile content.
Change-Id: I67682cfb807c58afc6de9375224ff7489d6618fb Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Enhance fsTick() to use filesystem timer resolution
RepositoryTestCase.fsTick() was was waiting 64, 128, 256, ... milliseconds
until it detected that the filesystem timer has ticked. Make use of
the filesystemtimer resolution information in FS to sleep a fraction
of the filesystem timer resolution. That raises probability to wake up
shortly after the filesystem timer has ticked.
Matthias Sohn [Sun, 26 May 2019 20:40:47 +0000 (22:40 +0200)]
Extend FileSnapshot for packfiles to also use checksum to detect changes
If the attributes of FileSnapshot don't detect modification of a
packfile read the packfile's checksum and compare it against the
checksum cached in the loaded packfile.
Since reading the checksum needs less IO than reloading the complete
packfile this may help to reduce the overhead to detect modficiation
when a gc completes while ObjectDirectory scans for packfiles in another
thread.
Bug: 546891
Change-Id: I9811b497eb11b8a85ae689081dc5d949ca8c4be5 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
Matthias Sohn [Wed, 8 May 2019 23:23:15 +0000 (01:23 +0200)]
Wait opening new packfile until it can't be racy anymore
If
- pack.waitPreventRacyPack = true (default is false)
- packfile size > pack.minSizePreventRacyPack (default is 100 MB)
wait after a new packfile was written and before it is opened until it
cannot be racy anymore.
If a new packfile is accessed while it's still racy at least the pack's
index will be reread by ObjectDirectory.scanPacksImpl(). Hence it may
save resources to wait one tick of the file system timer to avoid this
reloading. On filesystems with a coarse timestamp resolution it may be
beneficial to skip this wait for small packfiles.
Bug: 546891
Change-Id: I0e8bf3d7677a025edd2e397dd2c9134ba59b1a18 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
projects / jgit.git / log