Download git version 2.16.2

I'm Santosh and I'm here to post some cool article for you. If you have any query and suggestion please comment in comment section. Notice: It seems you have Javascript disabled in your Browser. In order to submit a comment to this post, please write this code along with your comment: 8afd47f49e86f7d1cdc. Thank you!

Create Git Repository. You may also like. Santosh Prasad Hi! View all posts. Keep it up!! Allow to filter list by project state in Projects. Issue : Allow per-project on-line notedb migration. Change : Fix buggy Index-Interactive Executor. Fix index. Change Change : Disable auto flushing during offline Lucene indexing. Change : Use thread count specified on offline reindex command line.

Honor the --threads command line option when provided instead of batchThreads property from gerrit. Change : Avoid creating loose objects in Schema Change : Avoid creating loose objects in schema Change : Fix schemas to use the appropriate ObjectInserter.

Change : Update schemas , to create a desired initial commit. Change : Create initial commit in schema only when necessary. Change Change Change Change Change : Avoid creating loose objects in schemas , , , , , and Change : Parallelize inserts into accountPatchReviewDb in schema Reduce the migration time on large sites by parallelizing inserts to accountPatchReviewDb. Change : Make schema migraitons more efficient by listing repositories only once.

Change Change : Parallelize schemas ,, which perform updates on each repository. Issue : Fix display the GC action on project commands for Gerrit in multi-site configuration.

Change : Add --migrate-draft-to flag on init doc. Change : Handle null account-ids when sending e-mails. Change : Make shuffling project slices optional in notedb migrator. Change : Make GC on repositories optional in notedb migrator. Change : Fix NPE while accessing draft refs. Change : EqualsLabelPredicate: fix bug that prevents matching owner votes. Issue : Inline edit allows the modification of Change-Id across patch-sets.

Unauthenticated users could exploit this problem in a Denial of Service attack, causing the server to go out-of-memory. Fixes a misconception that leads to data being accessible through Gerrit APIs that should be locked down. The former was always capable of providing correct results for all refs. This component was unaware of Git tags and notedb-related refs. That was problematic, because it bypassed documented behavior where a user should only have access to a change if he can see the destination ref.

The same goes for other database references. When the target repository is All-Users , the optimization turns into a security issue because it allows seeing all personal information associated with all accounts, i.

This fix now blocks Gitiles or any other part of Gerrit to abuse this power when the target repository is All-Users , where nobody can be authorized to skip the ACLs evaluation anyway. Introduce a new option --force-state-change-with-skip in the NoteDb migrator to force state change of the migration if projects are skipped.

Introduce a new option --verbose in the NoteDb migrator to display more information about the migration progress.

Fix contention on single database connection which all migration threads were competing for. Instead use one database connection per thread. Use multiple threads per project for improving the performance when there are a lot of changes per project. The impact is negligible on projects with a limited number of changes. Issue : Enable the ability to modify the commit message as part of change in-line edit. Introduce the documentation of Gerrit logs format and the meaning of each field.

Clarify how to resolve common Bazel issues with Eclipse project generation. Document X-Forwarded-For header in the Gerrit reverse-proxy section. Issue : BadMessageException: Response header too large.

Bump Jetty version to 9. Response header overflow leads to buffer corruptions Jetty server always allocates maximum response header size. Issue Issue To avoid idling threads during reindex operation, split the big projects into smaller parts and let the thread pool index these parts.

Issue : Fix plugin builds on OSX. Issue : Fix issue with auto registering ssh commands. Command interface. Correct interface name is org. Issue : Avoid closing System. By using try-with-resource wrappers around System. Due to the closed System. Issue : Honor project watches also for changes created via cherry-pick. When a change was cherry-picked and a project watch expression from a user matched the newly created change a notification email because of the project watch was often not sent.

Change the default for the CherryPickInput. Switching between changes in a change series randomly fails. Also, when adding a file to a change using in-line edit fails constantly with an HTTP status The new configuration option notedb.

This helps to speed up online migration at the expense of imposing a higher load on the Gerrit server running at the same time. Using that optional property, scenario steps can take some more or less time prior to expecting proper completion.

The way to set that property locally then depends on the target runtime environment, or SUT latency. The property may be used for either core or plugin scenarios.

FlushProjectsCache related scenarios added to core and the high-availability plugin. Each scenario can now either weight like any other by default, or override that default with a greater weight value, compared to siblings that are lighter on execution times. Beside core, make the high-availability and multi-site plugin scenarios reuse this. Add the corresponding core support for an optional http request body to every such GerritSimulation. Either automate or allow environment properties to feed the related input test data.

Beside core, make the high-availability and multi-site plugin scenarios test changes that way. The latter currently has Issue as a known limitation. Issue : Fix SQL connection timeout during migration to schema and A full GC is run at the beginning of the migration, after every k accounts, and at the end of the migration.

Additionally, bitmap index generation is disabled during the migration. Issue : Fix visibleto predicate for groups. If a group name was used in the visibleto predicate, changes were not returned even if the group had permission to see them.

Issue : Fix change query visibility for internal user. Issue : Block the removal of the Realm primary external ids. It was possible to remove the email address associated with the primary id of a Realm, effectively resulting in the account being removed. The response was returned when the auth backend could not be reached, instead of the response Some security tools will flag the absence of the autocomplete flag as a possible security issue, as it could allow for passwords to be saved and automatically filled in on shared computers.

In some cases the highlighting was shown on an adjacent line rather than on the line that was actually changed. The documentation is fixed, and --ownedby is added as an alias so that it also works for anyone still referring to the old documentation. Fix various misuses of the Flogger API causing exception details to be omitted in the log output.

Issue : Fix emitting debug logs while running tests. Issue : Fix parent of previous patch sets not being available. Issue : Fix missing replication Id in replication logs. The migration to use Flogger in 2. This is fixed by reverting back to slf4j logging for the replication log. Issue : Fix replication start --wait to track in-flight collisions and to not fail.

If there was already a pending push not an in-flight push to the same endpoint, the start for the push would be dropped when adding the push to the Destination. This meant that a replication start --wait command would never complete when one of its pushes was pending since its state would never receive the completion notification for that push.

Issue : Make persistent task keys stable. Calculation of the cache disks stat metrics for persistent caches is disabled by default because it is a very expensive operation that can block a thread for several minutes on larger installations.

Installations that still want to use this metric must explicitly enable it with the cache. Make the older version of the commit-msg hook from gerrit 2.

Fix handling of ambiguous name in visibleto search predicate. If the visibleto search predicate was used with a display name that resolved to more than one account, the first account in the result was arbitrarily used. Now, this case will result in an error.

Issue : Avoid multiple notifications for existing reviewers. Use 1. See PR Issue : Upgrade Guice to 4. Issue : Upgrade gwtjsonrpc to 1. It was originally introduced to allow classes to explicitly set their log level at level other than DEBUG. All the plugins API have been amended to avoid the use of concrete class EnumSet in their method signatures. Source-level compatibility is preserved because Set is a super-class of EnumSet.

In order to enable detailed performance and problem analysis, upload-pack metrics are now included in the sshd log. Add new configuration setting core. When set to true, this option enables using strong references to reference packfile pages cached in the WindowCache.

If this option is set Java gc can no longer flush the WindowCache to free memory if the used heap comes close to the maximum heap size. On the other hand this provides more predictable performance. This new setting works alongside the existing sendemail. When log. Note that the error log was already formatted in JSON when this option was enabled. Issue : Add support for going to a specific line number in the inline editor.

NamedFluentLogger is an extension to Flogger that allows to create a logger with a given name rather than the name of the enclosing class. Issue : Fix the access-path for AbstractGitCommand subclasses. The access path for the Receive. This allowed project owners to force-update a ref using git-over-SSH without having en explicit permission for that.

Issue : Cancel deprecation of change identifiers that was planned since 2. Since then the identifiers have still not been removed and there is no clear plan to do so. Due to this, and since continuing to support all of the types does not add any technical burden, ESC decided that the deprecation should be cancelled.

The change. Issue : Fix internal server errors when setting project access permission with bad regex. Issue : Fix time attribute of file entries in plugin artifacts. When a change Id passed as argument was resolved to multiple changes, the first returned change was arbitrarily used. Now an error will be raised and logged. Issue : Allow direct upgrade from v2.

Document dependency from account deactivator to autoUpdateAccountActiveStatus and log a warning if the account deactivation task is configured but will not be scheduled.

Modify draft comments generation so that, when removed, they to point to an empty parent, so that All-Users can avoid size explosion. Issue : Fix broken links in Elasticsearch configuration documentation. Issue : Open commentlinks to changes on the same server in the same tab.

Issue : Fix issues with caching edited commit message. Add method to push changes directly to given replica; intended to be used by other plugins that extend the replication plugin. When a ref-updated event is received, persist the event in the directory defined by the replication.

When the updated ref is replicated delete the persisted event. While we could conclude from the stack trace what this thread was doing, we missed some important information like repository name, user name, etc. The HTTP threads now have descriptive names. Deletion of draft comment refs was broken until 2.

The new site program DeleteZombieCommentsRefs can be used to remove them. Bugfixes in release 2. New in release 2. Improved how the Group By Columns collection editor works you can now select multiple columns in one go. For example, when you copy a measure, a new LineageTag is automatically generated. This also solves issue Added right-click options to certain properties.

Similar for KPIs. There is now also a right-click option for clearning the collection of annotations and extended properties. Bugfixes: Fixed an issue where many properties did not correctly indicate if they are read-only Fixed issue pbix file name should now be used as database name whenever the pbix metadata is saved as a. Assets 2 Source code zip Source code tar. Hoping this fixes issue Fixed unhandled exception when changes are made in a connected instance of Power BI Desktop see New in release 2.

This should fix issue More info. Saving a Power BI Desktop model as a folder structure now uses the file name as the database name in the serialised files, instead of the guid generated by PBI Desktop.