| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Change-Id: Icc99ca45b13414d5f7af8ff2c9b994a57caa6e7b
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This version includes several bug fixes, but most are related to
the server rather than the client. This upgrade is only to make
sure we're using the latest version.
Refer to the release notes [1] for details.
Also update the test container to test against version 6.3.2.
[1] https://www.elastic.co/guide/en/elasticsearch/reference/current/release-notes-6.3.2.html
Change-Id: I80e65f4fe267cf487fddb759b3914e54e31164f3
|
|
|
|
|
|
|
|
|
|
|
|
| |
When change is retrieved from the Lucene index, the project
name is URLEncoded and needs to be decoded before checking
if the refs are stale or not.
Failing to decode the project name would result in an
infinite loop trying to check for project staleness.
Bug: Issue 9482
Change-Id: I422d3ecde1fa12d38ed508bed6982aef71c2330e
|
|
|
|
|
|
|
|
|
|
|
|
| |
encoding
This test is passing on both Lucene and Elasticsearch, but exposes
the problem in the staleness checker as reported in issue 9482. When
running the test, the RepositoryNotFoundException can be seen in the
logs.
Bug: Issue 9482
Change-Id: I3dfe66e0f2051a6743963c1660ba435b1f8694d3
|
|
|
|
|
|
|
|
| |
There are no specific fixes that we need in this version - in fact
most of the changes are related to tests - so this upgrade is only
to align to the latest version.
Change-Id: I6e01398a05859b411fdbcd7db388e3061f866f53
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This release contains several fixes since 4.7.1.201706071930-r.
The ones we're particularly interested in are:
- ResolveMerger: Fix encoding with string; use bytes
which fixes issue 9153, and:
- Delete all loose refs empty directories
- Use java.nio to delete path to get detailed errors
- GC: Remove empty references folders
- Do not ignore path deletion errors
which fix issues with cleanup of ref directories after running GC.
The release also includes:
- Honor trustFolderStats also when reading packed-refs
Users should note that for repositories with a high number of
references (for example in excess of 300K refs) and in a server with
high traffic, this solution may not scale well and should be tested
carefully.
Bug: Issue 9153
Change-Id: I4a229589dd1a3a14844538df0556fa1b62731029
|
|
|
|
|
| |
Change-Id: If7021f46e1df96d923371a91fba92e45a105126c
(cherry picked from commit faf05c292bb2e4d307fcac02ddfb4c9c938eefb2)
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit adds the missing permission check when delete a single
branch through the "Delete Branches" endpoint. The permission
check is forgotten because DeleteRef#deleteSingleRef assumes the
caller has done it before. But "Delete Branches" didn't do that.
It's created just for this stable branch. For "master", it's fixed
by I6a7f48e4a9. Cherry-pick is hard because these classes have been
refactored to other packages.
Change-Id: I6431ffe5d2625c54dab23516680c7a9a81019108
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Remove the postOK and getOK methods, and corresponding get/post
methods that take an 'expectedStatus' argument. They are only used
by the CacheOperationsIT tests, and can be easily replaced with
calls to RestResponse.assertOK().
Change-Id: I403bf2076b8a93efc63596d837ac25d971ca3fa9
Signed-off-by: David Pursehouse <dpursehouse@collab.net>
|
|/
|
|
|
|
|
|
| |
When update of project config fails there is no way
to know what is the cause.
Change-Id: I9bfd398c6e69865d332dd9849a0cb66a297473b4
Signed-off-by: Eryk Szymanski <eryksz@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Do so as already done in 2.15 and master versions.
Fix ElasticReindexIT so it passes also with this change. Contrary to
Lucene, configuring the Elasticsearch index type currently assumes
uninitialized indices -in Elasticsearch. Thus fix the test so it first
asserts that assumption, then reindexes or initializes indices offline.
The test then keeps testing as currently the case also for ReindexIT,
i.e., through their common parent test class methods.
Bug: Issue 8799
Change-Id: Ie52ea03a781c89fb3caf0f965f6fca1fb6c6a478
|
|
|
|
|
|
| |
Fix the corresponding IDEA warning.
Change-Id: I00ea3c69ffa7f813b4f7929c8ce0f88f4ffb42cc
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The new version provides better support for Java 9 and improves
performance. Multibindings artifact is now included in the guice
core artifact. The full release notes is here: [1].
[1] https://github.com/google/guice/wiki/Guice42
Change-Id: I208d53ce1cf9b99ed8f03ae34470e95c3b8f71ce
(cherry picked from commit 1fa7774d237d0ea864b9f03b6551fd1ebef0c619)
|
|/
|
|
| |
Change-Id: I7243f1ce175eabaaad54905e55c2c5e512ce2175
|
|
|
|
|
|
|
|
| |
Add "Delete Changes" permission to constants so that it is displayed
properly in GWT UI.
Change-Id: I8b71afa8ed1507b3ca6f8ae1a5c48575c2477281
Signed-off-by: Jacek Centkowski <jcentkowski@collab.net>
|
|
|
|
|
|
| |
ExpectedException#expect
Change-Id: I33f3a67134175d76ec7933b1b3ca54e22ea25904
|
|
|
|
|
| |
Change-Id: I5af2c5e9dfa4392baa824faf9c30f2c6b5e33c23
Signed-off-by: Edwin Kempin <ekempin@google.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Any additional statements after the statement that is expected to
throw will never be executed in a passing test. This can lead to
inappropriately passing tests where later incorrect assertions are
skipped by the thrown exception [1].
[1] http://errorprone.info/bugpattern/ExpectedExceptionChecker
Change-Id: I06a08ddd15e32decf5c18289ba228c2c78b31065
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Only define the password once; default one for the stock 'elastic' user.
The potential merge-up of this commit to master should consider the
other change I290c8af2. If the latter gets merged, then that merge-up
should also include amending master-only ElasticV5QueryProjectsTest.
Change-Id: I99b4e328be1f1b761edb0abe8e55e4115ce3365a
|
|/
|
|
|
|
|
| |
As that (static) 'container' member was not read elsewhere than in this
static method. Thus fix the corresponding IDEA warning.
Change-Id: Ie38e14ed13354aa91e313e141e1959a6752f24cf
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The fail function takes an already formatted message,
it does not provide for formatting arguments, the
way Python's logger.log does, for example.
I did not find any other improperly made call to fail
in the rest of the bazlets project.
C.f. https://docs.bazel.build/versions/master/skylark/lib/globals.html#fail
Change-Id: If187b4d1f97ef331ead8a34f499c97cd8d5728f9
|
|
|
|
| |
Change-Id: I73ac5e27a8bf9b96efb4551752c7536660fed7c9
|
|
|
|
| |
Change-Id: I4010235c44fecd27f0286eb17e0e132a4bd4f220
|
|
|
|
|
| |
Bug: Issue 9146
Change-Id: I11d8f4da87a0e27e9117e2c0087cbaff5d185338
|
|
|
|
|
| |
Bug: Issue 9147
Change-Id: Id33c105351b47c38a1f87258cde83d4c8fb4e8de
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When using Elasticsearch, doing a query that involved the characters "."
and "_", from full text fields, did not include results with keywords as
a substring. This behavior was different from Lucene, where these two
characters are mapped to the space character (" ") so that the query
returns keywords separated by them.
This change adds character mappings for Elasticsearch in order to ensure
that the full-text queries return same results as when using Lucene. At
index creation time, this change creates a new elasticsearch setting
where an analyzer with character mappings is configured. This analyzer
is then added to the elasticsearch mappings to be used by the full-text
field queries.
Because the elasticsearch mappings and settings can only be configured
at index creation time, one should take the following steps to apply
this change:
1. delete the index (changes, accounts, groups).
2. initialize an Elasticsearch site with this change.
3. reindex the documents (changes, accounts, groups).
4. start the site.
This change applies to all the currently supported Elasticsearch
versions.
Bug: Issue 9146
Bug: Issue 9147
Change-Id: I6da7a98d35d912b5bee7cc510d02db4433f25538
|
|
|
|
|
|
| |
As the latter has been refactored through the one following it straight.
Change-Id: Iabeced6beae5e937d9c667b75981819de9511f86
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As per Bazel test documentation [1] that is: run all the Elasticsearch
docker tests with the 'exclusive' tag set in. Do so to minimize the
slower execution speed factor, thus running them tests in a more natural
fashion. Doing so should then minimize flakiness or only leave true
failures to be fixed. -Actual test execution speeds depending on the
running host resources, of course.
[1] https://docs.bazel.build/versions/master/test-encyclopedia.html
Bug: Issue 9407
Change-Id: Ic9b57b84d06921abab91de57ecd4b86bcfb4272a
|
|
|
|
|
|
|
|
| |
Make this test rule generic enough, so that other potential new test
classes get executed with no change required. Exclude test classes that
are already picked up by sibling test rules.
Change-Id: I2814885b364d8b50017b728ea6fc6301516c45cd
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add execution of the V5 and V6 tests to the already included V2 default.
Include flaky tag support for the V6 tests though, as their accounts and
groups subset currently fails (cf. Issue 9252). Adding such a tag here
allows to exclude the V6 tests in a CI context. The V5 tests still pass
(just like V2's) thus require no such flaky tag for passing CI coverage.
Before this change, it was not possible to benefit from V5/V6 elastic
tests feedback. Thus the previous risk of elastic support regressions.
Once a fix for Issue 9252 becomes available, one will be able to get
such feedback for V6, on top of V5 added herein and default V2. Until
then, excluding the hereby flaky V6 tests allows for V5 and V2 checks:
bazel test //... --test_tag_filters=elastic,-flaky (e.g.).
Change-Id: I62fef6dca20a73065df6b3dd4b21deb1ac7653fd
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
$ bazel test -t- //gerrit-elasticsearch/...
INFO: Build options have changed, discarding analysis cache.
INFO: Analysed 3 targets (2 packages loaded).
INFO: Found 3 test targets...
INFO: Elapsed time: 126.687s, Critical Path: 124.80s
INFO: 9 processes: 2 remote cache hit, 4 darwin-sandbox, 3 worker.
INFO: Build completed successfully, 24 total actions
//gerrit-elasticsearch:elasticsearch_query_accounts_test PASSED in 65.0s
//gerrit-elasticsearch:elasticsearch_query_changes_test PASSED in 122.5s
//gerrit-elasticsearch:elasticsearch_query_groups_test PASSED in 53.5s
Executed 3 out of 3 tests: 3 tests pass.
INFO: Build completed successfully, 24 total actions
Change-Id: I1bc530879ae69d74d39750a173d5458b2fb72bd9
|
|
|
|
| |
Change-Id: I7556af36c770d1af8822c3d5bd36abc043f19092
|
|
|
|
|
|
|
|
| |
Buildifier is now also used for formatting .bzl files.
This change was created by running buildifier 0.12 over our source tree.
Change-Id: I9f15112d4fe23e5cec0700cfe47f1ca649f61d2a
|
|\
| |
| |
| |
| |
| |
| | |
* stable-2.13:
Update Bower to 1.8.2
Change-Id: I690bc00105d4987bec0aa14bc7936a85256b3f3b
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
See https://twitter.com/bower/status/918073147789889536
bower moved their mirror to in house so they can put it behind a cdn.
This also resulted in a domain change.
Change-Id: Ie26d464687fe10a46c1fd83b9445c5fcf3cbbca9
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This dependency is not used in gerrit core. The only reason it's
included in WORKSPACE is because core replication plugin depends on it
and there is no way in Bazel to consume external dependency in core
plugin without defining this dependency in WORKSPACE file.
It turns out, that gitblit-plugin also depends on commons-io, but on the
newer version. Sync the versions to the same version used by gitblit.
Change-Id: Iae3545df95b23439d3c7fd385bfeae482d6d60ac
(cherry picked from commit 6488535f737cdce5feeca2550ce402d803d4d5a9)
|
| |
| |
| |
| | |
Change-Id: I9d8e186e5836e83ea8bb3d4917b349ff75e4374b
|
| |
| |
| |
| | |
Change-Id: I2d85bbade228d89caabad334742641ecccbe4af7
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If a change is pushed without a Change-Id footer, but with a potential
Change-Id line somewhere else in the body of the commit message, the
error message says:
Hint: A potential Change-IdChange-Id was found, but it was not
in the footer (last paragraph) of the commit message.
i.e. with repeated "Change-Id" in the first line.
Change-Id: I035e20b72b0b48b090a9183310f3460f97755d00
|
| |
| |
| |
| |
| |
| | |
Change-Id: I5e1a4550701a9eaad2bbfdb9e7183a21dfda7cec
Signed-off-by: David Pursehouse <dpursehouse@collab.net>
Helped-by: Patrick Hiesel <hiesel@google.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When change is pushed for review typical response looks like:
remote: New Changes:
remote: http://[server]/gerrit/[change-id] [change subject]
It is not sufficient for sites that serve Gerrit as a service that is
integrated under unified UI (with unified Code Browser). In such
environment links to Gerrit are no longer valid and it should be
possible to alter them with desired format.
This patch introduces ChangeReportFormatter that can be replaced with
instance loaded from plugin (DynamicItem pattern) with default
implementation that preserves current behavior.
Change-Id: Ia2de44c121b6a73fcf93e1aaf11690697adb9abf
Signed-off-by: Jacek Centkowski <jcentkowski@collab.net>
Signed-off-by: David Pursehouse <dpursehouse@collab.net>
|
| |
| |
| |
| |
| |
| | |
Also update the test container to use version 6.3.1.
Change-Id: I78d24ba8ba1a8612276a78ab3549b37cd53708d2
|
| |
| |
| |
| | |
Change-Id: I4f7b4da57c29f95218c2e75beb256182d7f69f54
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Changes can only be deleted by Administrators, or by the change owner
when they are granted the "Delete Own Changes" permission either on the
"Change Owners" virtual group or an internal group they are a member of.
This means that it is not possible to allow users to delete other users'
changes other than by granting the "Administrate Server" permission,
which also gives them all the other administrative capabilities. This is
also inconsistent with how the "Delete Drafts" permission works.
An example of where it would be better to have more flexible permissions
is allowing changes to be deleted by Project Owners on a per-project
basis, or by a group of users who are trusted (but not enough to grant
them Administrator capabilities).
Add a new permission "Delete Changes" which allows these use cases, thus
a change can now also be deleted by users who are:
- Member of a group that is explicitly granted "Delete Changes" on the
change's destination branch.
- Member of a group that is given the "Owner" permission on the project,
when the virtual group "Project Owners" is granted "Delete Changes" on
the change's destination branch.
Bug: Issue 9354
Change-Id: I9d5d779eb0a6ce18faca02b9fc90904ec5da91d9
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Instead of cloning the repo as "user" and then using git push to create
the change, use the change API to create the change. This avoids the
unnecessary cloning of the repository, and makes it more explicitly
clear which user is supposed to be the owner of the created change.
Change-Id: Id4a3ff77a6128d4500eeaba37d01d69b8253b165
|
| |
| |
| |
| | |
Change-Id: Ib23c6bdc18212aba8b8e35ab56c25e6dc3fd8429
|
| |
| |
| |
| | |
Change-Id: I3100fac2c3e00d50a52d78c0eed58c376c102019
|