| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Change-Id: Ic4fa85ad08771ef0b421c102b088adce45f42238
|
|
|
|
| |
Change-Id: I1f5706f2d4aa85b30be5660f623ce4b4fe3101ca
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Change-Id: Iffcd0fbd7 has involuntarily triggered the
creation of a new HTTP Session for every invocation a Git-over-HTTP
request.
All came from the mistake of tracing the HTTP session instead
of the Gerrit session in the audit record.
The HTTP Servlet API specs say that any attempt to access
the current session of an incoming request would result
in the creation of a brand-new session.
The session involuntarily created also had an expiry time
equal to zero, which prevented the session housekeeper
to reclaim them later on, even though they were unused.
The consequence of creating an empty session for every
Git-over-HTTP request isn't immediately tangible, because
the session is empty and doesn't occupy a significant
amount of memory. However, longer-term, the in-memory
hashtable that records all the sessions, each one using
750 bytes on average, will be causing the overload
of the JVM heap and the crash of the process because of
lack of available memory.
Use the correct Gerrit session-id, retrieving
from the Provider<WebSession> the proper session, if active
and logged in, and make sure in tests that no HTTP sessions
are created as a result of a Git-over-http request.
Bug: Issue 13858
Change-Id: I8c086fed54b196c3f46fa88ac78c127784524d30
|
|\
| |
| |
| |
| |
| |
| |
| |
| | |
* stable-3.1:
Update git submodules
AccountResolver: limit number of ambiguous accounts in error message
Update highlightjs to 10.5.0
Change-Id: Ib49b3bb95a606db0d0418d9d2e467aadba330768
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This upgrades currently used version 10.0.2 from May 2020.
Build process is amended to use terser minification, as it provides
better result as closure-compiler: [1]
Terser (5.3.8):
$ ls -all build/highlight.min.js
-rw-r--r-- 1 davido users 977775 Jan 10 10:04 build/highlight.min.js
Closure-Compiler (v20210106):
$ ls -all build/highlight.min.js
-rw-r--r-- 1 davido users 994034 Jan 9 12:10 build/highlight.min.js
So, terser produces minified build that is 16259 bytes smaller.
Moreover, in latest gerrit releases terser is also used for gr-app.js
minification, so it appears to be consistent to also use terser for
highlight.js minification.
[1] https://github.com/highlightjs/highlight.js/issues/2956
Change-Id: I3a67dc05eb9c4fdac1bf578eacd777dfefb2a360
|
| |\ \
| | | |
| | | |
| | | | |
into stable-3.1
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Don't list more than 3 ambiguous accounts in error message of
UnresolvableAccountException in order to prevent flooding the log and
displaying a lot of sensitive account data in an error dialog.
Bug: Issue 13803
Change-Id: I95fe697988f3d900cce29ca93114a725b1c17ca6
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* Update plugins/replication from branch 'stable-3.1'
to 141f223240b15d3576f2d45525ed12b6ce7bc9bf
- Document that authGroup must have Access Database
Bug: Issue 13786
Change-Id: Iaf65252b25b9c40e5cfd1ac25d55fbf70536f83e
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* Update plugins/replication from branch 'stable-3.2'
to dedf422a7a0336096c5503d1f652f77873fbf11c
- Merge branch 'stable-3.1' into stable-3.2
* stable-3.1:
Document that authGroup must have Access Database
Change-Id: I09378f4288fd1335932bdf120bba8418fc8f51c7
- Document that authGroup must have Access Database
Bug: Issue 13786
Change-Id: Iaf65252b25b9c40e5cfd1ac25d55fbf70536f83e
|
|\| | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* stable-3.1:
Update jetty version to 9.4.35.v20201120
Change-Id: I97edf8fb6c5f97df93aff2ef62b710cd08c780ef
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Includes a fix for CVE-2020-27218 [1] that affects versions prior
to 9.4.35.v20201120.
As of: [2] new dependency was added to jetty-servlet: jetty-util-ajax.
[1] https://nvd.nist.gov/vuln/detail/CVE-2020-27218
[2] https://github.com/eclipse/jetty.project/issues/5539
Change-Id: I51549c9588f5748f36e145f8c4b347f97c08b4d8
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This is a backport of change 293242 to stable-3.2 branch and to
consume shadow-selection-polyfill module with npm.
Bug: Issue 11811
Change-Id: Iaf62137e48b7d119be3d8be96ad03a7f316a5e69
|
|\| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The change:
96ccc2388 Add shadow-selection-polyfill
is reverted, as it will be migrated so that shadow-selection-polyfill
dependency is consumed as npm module.
* stable-3.1:
Add shadow-selection-polyfill
Update git submodules
Revert "Upgrade soy to 2020-08-24"
Change-Id: I0a863e9de9f35d4a9a77302e2a05d9850a2b3538
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The selection on shadow-DOM is not implemented
in Safari and the document.getSelection() fallback
is not suitable for use.
The GoogleChromeLabs provided a library to solve
the problem: shadow-selection-polyfill.
What the library does is to manage the selection
change events and calculate and cache the results
for the shadow root elements, creating a custom
event '-shadow-selectionchange' once the selection
has been completed.
There is one gotcha that makes things slightly
more complicated: what the library provides is a Range
and not a Selection object. There are a couple of
adjustments needed in gr-diff and gr-diff-highlight
to use the range directly instead of getting it
from the selection.
NOTE: Even though the shadow-selection-polyfill could
also manage Chrome and Firefox, keep the existing
logic to avoid any possible regression, which is not
desireable on a stable branch.
On stable-3.1, because of the problems related to the inclusion
of the dependency via Bower, shadow.js has been removed
from its exports and used as pure JS file included in the
gr-diff.html.
This change needs would not be applied to stable-3.2 where,
thanks to the npm package mangement, it would be consumed
from the NPM registry directly.
Bug: Issue 11811
Change-Id: I41c4e94343010972c8a9f0f1ba3a059ca7af5292
|
| |\ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This reverts commit bcdbc45e85efe98ea820f5024c5ca566ba0576d5.
Han-Wen reported this soy version to show performance issues in Gerrit:
"Google is tracking this internally as b/176094888; hopefully we can get
it fixed in Soy proper.
The problem is that template rendering went from 9ms to 30ms. We render
quite a few templates for a single email, so adds ~120ms to each
operation that sends out an email.
It looks like this caused a significant slowdown on the benchmark under
contrib/benchmark-createchange.go (using reftable as refstorage)."
Hence reverting this update until it's fixed upstream.
See https://gerrit-review.googlesource.com/c/gerrit/+/280196
Change-Id: Id98a5d93e316268e086cf6aa2be514a586dd3eee
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* Update plugins/replication from branch 'stable-3.1'
to 22ca0b406a4efb9aebbbfcac8d2d986812423f01
- Split integration tests to different targets
Running all integration tests as part of one single 'replication_it'
target does not cope well with the addition of extra tests, because it
is bound to take longer and longer, eventually hitting any test timeout
threshold.
Splitting integration tests into different targets avoids timeout
failures and also provides additional benefits, such as:
- Better understanding of test failures
- More efficient utilization of bazel build outputs and remote caching,
effectively making tests execution faster.
Bug: Issue 13909
Change-Id: Ifc6cce9996d3a8a23ec2a66c377978205fb6680f
|
|\ \ \
| | | |
| | | |
| | | | |
stable-3.2
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This change solves the operator/operand aliasing not finding content in
gerrit.config (it would only be found in secure.config).
Change-Id: I10b96225526dc355e668357715f5e60a9d3d9751
|
|\ \ \ \
| |/ / /
|/| | | |
|
| |\| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* stable-3.1:
Add script for incremental reindexing during upgrade
Add query option allowing administrators to skip visibility filtering
Change-Id: Iade61ce701a6efcbdcc12bf2a0dda6a9af32419e
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
In order to shorten the downtime needed to reindex changes during a
Gerrit upgrade the following strategy can be used:
- index preparation
- create a full consistent backup
- note down the timestamp when the backup was created (backup-time)
- create a complete copy of the production system from the backup
- upgrade this copy to the new Gerrit version
- online reindex this copy
- upgrade of the production system
- make system unavailable so that users can't reach it anymore
e.g. by changing port numbers (downtime starts)
- take a full backup
- run
$ ./reindex.py -u gerrit-url -s backup-time
to write the list of changes which have been created or modified
since the backup for the index preparation was created to a file
"changes-to-reindex.list"
- upgrade the production system to the new gerrit version skipping
reindexing
- copy the bulk of the new index from the copy system to the
production system
- run
$ ./reindex.py -u gerrit-url
this reindexes all changes which have been created or modified after
the backup was taken reading these changes from the file
"changes-to-reindex.list"
- smoketest the system
- make the production system available to the users again
(downtime ends)
Change-Id: Ie736e0dc32180329ca6ed31bcb49eb6b96bf2b91
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
If an administrator wants to reindex changes which were created or
updated in a given period based on a query for that period the
query results are subject to visibility filtering. This can have the
effect that e.g. private changes are missed.
Add a query option "skip-visibility" to allow administrators
to skip visibility filtering.
Change-Id: I66c13659587b9459eb7cc585697c1655926ceac3
|
|/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* Update plugins/replication from branch 'stable-3.2'
to 99c822ebf3b72023f8b926444bbdf299fafb00df
- Merge branch 'stable-3.1' into stable-3.2
* stable-3.1:
Split integration tests to different targets
Don't check read permission when authgroup isn't set
Change-Id: Ic5c8f0468869476a01923b4d374f0188c271daf2
- Split integration tests to different targets
Running all integration tests as part of one single 'replication_it'
target does not cope well with the addition of extra tests, because it
is bound to take longer and longer, eventually hitting any test timeout
threshold.
Splitting integration tests into different targets avoids timeout
failures and also provides additional benefits, such as:
- Better understanding of test failures
- More efficient utilization of bazel build outputs and remote caching,
effectively making tests execution faster.
Bug: Issue 13909
Change-Id: Ifc6cce9996d3a8a23ec2a66c377978205fb6680f
- Don't check read permission when authgroup isn't set
It's unnecessary to check read permission when authGroup isn't set since
the then the user is a RemoteSiteUser that is-an InternalUser that has
read access to everything.
Change-Id: Ie6985250b0acb50c08fdcae75cc608222b1add35
|
|\| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* stable-3.1:
Fix 'is:submittable' query on multiple submit records
Update git submodules
NOTE: No-op merge, the only commit needed was already cherry-picked
onto stable-3.1 and the replication plugin should stay on its
own stable-3.2 branch point.
Change-Id: I18813f4cbb79ed900eeecef54a4cc676c6fba13b
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When a project has multiple submit rules, which produce
multiple submit records, the 'is:submittable' query stops
working as the documentation indicates it should. Rather
than returning changes that are ready to be submitted,
it returns any change where at least one submit record
is OK, despite the overall change not being submittable.
For example, with the code-owners plugin, which uses
a java-based submit rule, 'is:submittable' will return
changes that are passing owners checks, but might
have CodeReview:-2 or Verified:-1.
For projects with a single submit rule, the behavior
is exactly the same as before, since we simply check
that *any* of the submit records is OK, exactly as before,
AND that *none* of them are NOT_READY or RULE_ERROR.
Bug: Issue 13884
Change-Id: I4878ce13c6673852916d6891253d5e62b46f3db5
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Update plugins/replication from branch 'stable-3.1'
to 0995fe0445f507279c05cb5ee60a9413671be400
- Don't check read permission when authgroup isn't set
It's unnecessary to check read permission when authGroup isn't set since
the then the user is a RemoteSiteUser that is-an InternalUser that has
read access to everything.
Change-Id: Ie6985250b0acb50c08fdcae75cc608222b1add35
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
By default terser keeps all licensing comments as many licenses require
you to keep a copy of the license in redistributed source: [1]
--comments [filter] Preserve copyright comments in the output. By
default this works like Google Closure, keeping
JSDoc-style comments that contain "@license" or
"@preserve". You can optionally pass one of the
following arguments to this flag:
- "all" to keep all comments
- `false` to omit comments in the output
- a valid JS RegExp like `/foo/` or `/^!/` to
keep only matching comments.
Note that currently not *all* comments can be
kept when compression is on, because of dead
code removal or cascading statements into
sequences.
In minified gr-app.js version we don't need to preserve license
information and can remove it. This saves ca. 300KB in gr-app.js size.
[1] https://github.com/terser/terser
Bug: Issue 13870
Change-Id: I2736879d252994200ad2f52e414e224ebe56b65d
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When a project has multiple submit rules, which produce
multiple submit records, the 'is:submittable' query stops
working as the documentation indicates it should. Rather
than returning changes that are ready to be submitted,
it returns any change where at least one submit record
is OK, despite the overall change not being submittable.
For example, with the code-owners plugin, which uses
a java-based submit rule, 'is:submittable' will return
changes that are passing owners checks, but might
have CodeReview:-2 or Verified:-1.
For projects with a single submit rule, the behavior
is exactly the same as before, since we simply check
that *any* of the submit records is OK, exactly as before,
AND that *none* of them are NOT_READY or RULE_ERROR.
Bug: Issue 13884
Change-Id: I4878ce13c6673852916d6891253d5e62b46f3db5
|
| |
| |
| |
| | |
Change-Id: Ida8557013073be101b5349458af1380b471ef59a
|
| |
| |
| |
| | |
Change-Id: Ida9a38a340d6fc088b6350ed7cd42f9326b10374
|
|\|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* stable-3.1:
Set version to 3.1.12-SNAPSHOT
Set version to 3.1.11
Fix httpcore dependency needed by httpclient
Revert "Add support for Elasticsearch version 7.9.*"
Revert "Add support for Elasticsearch version 7.10.*"
Change-Id: I333a2434d66df525aa85687d9cc0134e7a398c21
|
| |
| |
| |
| | |
Change-Id: I4894bba96a34d5c5da8570d3d3a5bf04b77e70ee
|
| |
| |
| |
| | |
Change-Id: I6c87be906a961c2b938d8260390ed4255fcd5102
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
httpcore has removed annotations used by httpclient and therefore
it would fail to execute the annotation processor.
E.g. the high-availability plugin, which uses the httpclient, won't
compile anymore because is failing to find the missing annotation
removed from httpcore from v4.4.5 onwards.
The wrong httpcore dependency was introduced yesterday with the merge
of Change-Id: I1fc91bce.
Change-Id: I2d45001995806074534c11fc5230765e4c99e5a2
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Revert submission 291062
Reason for revert: requires an httpclient that Google cannot support
Reverted Changes:
I1fc91bce9:Align http dependencies with elasticsearch-rest-cl...
Ifb941eea6:Add support for Elasticsearch version 7.10.*
I7d76ece60:Add support for Elasticsearch version 7.9.*
Change-Id: Ib95a8c053fa7f046596458e7eb719a409847cb82
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Revert submission 291062
Reason for revert: requires an httpclient that Google cannot support
Reverted Changes:
I1fc91bce9:Align http dependencies with elasticsearch-rest-cl...
Ifb941eea6:Add support for Elasticsearch version 7.10.*
I7d76ece60:Add support for Elasticsearch version 7.9.*
Change-Id: I740e06058469a0ab12b14c7c616940bf429d0de7
|
|\|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* stable-3.1:
ForRef#check should permit internal users to read all refs
Add Jacek Centkowski to developers
Add support for Elasticsearch version 7.10.*
Add support for Elasticsearch version 7.9.*
Align http dependencies with elasticsearch-rest-client
Change-Id: Ia708f5a927da3f8ec8beef2470abfcf011843d5c
|
| |\
| | |
| | |
| | | |
stable-3.1
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
79d24d4 Make PermissionBackend#ForRef authoritative
Introduced a regression where InternalUsers where not taken into
consideration when checking READ permission.
Bug: Issue 13786
Change-Id: I3f18507f65044ac96321c1efecf1f2688f36859f
|
| |\ \
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* changes:
Add support for Elasticsearch version 7.10.*
Add support for Elasticsearch version 7.9.*
Align http dependencies with elasticsearch-rest-client
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Upgrade elasticsearch-rest-client to 7.10.1.
Add V7_10 to the version manager and use it for all the V7 tests.
Change-Id: Ifb941eea63b90e107ae37c6a08ab1815d6c0f22b
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Upgrade elasticsearch-rest-client to 7.9.3, which is the currently
latest 7.9 minor version available.
Add V7_9 to the version manager and use it for all the V7 tests.
Change-Id: I7d76ece60d3ed7400bcf28bcd45d0e2e02870ccd
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Address the upstream issue [1] which is caused by the lack of upgrades
done in this change. While waiting for Elastic to potentially provide
shaded dependencies [1,2] for these, upgrade them herein still for now.
Meaning, do not consider trying [2] this time around either. Favor the
upgrade approach which is still simpler and likelier to succeed today.
Base these new versions on [3]'s dependencies, even though the issue
([1]) didn't fail the build as such for version 7.8.1 of the client.
This issue's fix starts becoming necessary from client versions 7.9.x,
as shown by the follow-up upgrade change to 7.9.0 which requires this.
The most downstream branch of gerrit to include this change determines
from which Gerrit version the Elasticsearch versions 7.9.x can be
supported.
Set HTTPCOMP_VERS to the earlier 4.5.2, to comply with Google's
currently deployed internal version.
[1] https://github.com/elastic/elasticsearch/issues/61664
[2] https://www.elastic.co/guide/en/elasticsearch/client/java-rest/current/java-rest-low-usage-shading.html
[3] https://search.maven.org/artifact/org.elasticsearch.client/elasticsearch-rest-client/7.8.1/jar
Change-Id: I1fc91bce95d9e8094c19ad4044bd08b6285524af
|
| | | |
| | | |
| | | |
| | | | |
Change-Id: If92e2efd75b8e8230f6ffc47f550a8cc92edea3b
|
|\| | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* stable-3.1:
Upgrade testcontainers to 1.15.1
Change-Id: I38e8d0d4920fabbeb645c2e9dee8c7dd440f4b49
|
| |\| | |
|
| | |/
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
According to [1], "focuses on the stability and compatibility". Includes
bug fixes and an update of docker-java to 3.2.7, which gets reflected in
this change just as well.
This change is currently necessary to successfully execute all the
elastic core tests on macOS, with Docker Desktop, locally.
[1] https://github.com/testcontainers/testcontainers-java/releases/tag/1.15.1
Change-Id: I9a7c443c69e3d054a6c52451b3aa3cf57a7933d3
|
|\| |
| | |
| | |
| | |
| | |
| | |
| | | |
* stable-3.1:
SshDaemonIT: Perform some refactoring
Change-Id: I4431413b86a1d574ecc7d799ba3289069a5cea7a
|