| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
The name is actually `Gerrit-Has-Labels`, not `Gerrit-HasLabels`. The
email footers do not follow a consistent naming schema, some of them use
dashes and others use camelcase. This label, contrary to the current
documentation, uses dashes.
Release-Notes: skip
Google-Bug-Id: b/291061494
Change-Id: Iae5e49cfa00c95800838599375d231801c6cd0c8
(cherry picked from commit 9da5f5d99b1c8e8e60f572abaa11a99bb46a6fa7)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The RunnableScheduledFuture task inside a Task has a few layers that
don't help provide a meaningful thread name whereas the Task itself
often has a toString() that does, so use the Task itself.
Here's an example for a SendEmail thread. It's a similar result for
Index-Interactive and other threads.
Before (line breaks added by me):
SendEmail-1[
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask@57310449[
Not completed,
task = java.util.concurrent.Executors$RunnableAdapter@3a7a2870[
Wrapped task = com.google.gerrit.server.logging.LoggingContextAwareRunnable@9252ca5
]
]
]
After:
SendEmail-1[send-email comments]
Change-Id: Iea78b6412b2259a968eed0a16744638f87796d91
Release-Notes: Improved WorkQueue task thread names
(cherry picked from commit d4c1e0a6ed7eb2b8fffa078a7ccec0c0edb0e668)
|
|
|
|
|
| |
Release-Notes: skip
Change-Id: I8f607aa6d7512bc6933662ed022b92f3cd5993c7
|
|
|
|
|
| |
Release-Notes: skip
Change-Id: Iad3a38e786767a5ae1b0543f25eb40a424b899c2
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* stable-3.7:
Set version to 3.7.10-SNAPSHOT
Set version to 3.7.9
Update Getting Started to Gerrit 3.7.8
Release-Notes: skip
Change-Id: Ice602304fe9c3703a735b6f1adc4e4923868304e
|
| |
| |
| |
| |
| | |
Release-Notes: skip
Change-Id: Ic99c0d13a99f9577f0762e3c8bd41ca2b798f744
|
| |
| |
| |
| |
| | |
Release-Notes: skip
Change-Id: Id6feef84ad0578f2a09c0b97591d8be768293507
|
| |\ |
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: Id9ec22f0688adf290b27f32ffb71d63df3f6a763
Release-Notes: skip
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: Ie4bc10a1e760202296ff98007586eb9ea358918f
Release-Notes: skip
|
|\ \ \ |
|
| |\| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* stable-3.7:
Mention who owns the change in comment emails
Release-Notes: skip
Change-Id: I29535eb95eb23576567f06337c3657a25856418e
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
When receiving the email, knowing the change owner lets the receiver
knows it is one of their change and might refine reviewing priority for
some people.
We had that customization for a few years now, and I felt like it can
be upstreamed (ref: https://phabricator.wikimedia.org/T43608 ).
Release-Notes: Emails notifications for comments now mentions who owns the change.
Change-Id: If676e4f1b75bdbe2d39dcfe1fbe841520e81068f
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
There is generally a slight misalignment of icon and text as can be
seen in this screenshot: https://imgur.com/a/tDUHGk0
It was also noted in the referenced bug that increasing the icon height
is impossible, because then the alignment gets totally messed up.
3 problems contribute to this misalignment:
1. `.titleText` element being a span. This generally makes alignment of
icons next to text very tricky. Let's use a flex element instead.
2. `vertical-align: text-bottom` for the icon is a very unusual choice
for the vertical-align property and does not really make a lot of
sense. Let's just use the `align-items: center` style for flex
elements, see above.
3. `.bigTitle` has a custom font size, but not a custom line-height.
That is generally problematic as can be seen here: The element will
get a height of 20px as an inline element, but contains text with a
24.5px font. No surprise that we see misalignment. Let's make sure
that the line-height and thus the height of the text element has a
proper value. 1.2*font-size is what browsers would normally use.
This was checked to behave well on gerrit-review, android-review and
chromium-review.
Release-Notes: skip
Bug: Issue 337076005
Change-Id: Ia09bbe92158f7b4b1d6b88e599e0635fa9b15d5e
(cherry picked from commit 06cfcb6ef54d78d7e68dfc0f17a15d30d6aae414)
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The variable was spelled PACTCH_SET_ID by accident.
Change-Id: I96696ffaed79281611173e158ad5bec7a5e475e1
Release-Notes: skip
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Not sure why this wasn't caught by gjf earlier.
Change-Id: I9b1306f51e6218364e8ae9f43274e12b520d4797
Release-Notes: skip
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This was supposed to be using the length of the originalSubject when it
was different from the subject, but a likely copy/paste error had it
using the length of the subject again.
Change-Id: I4dd18dfb9662e76b8e9fc85c29fdc022b6d58af3
Release-Notes: skip
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This annotation was not applied consistently to fields and methods, with
some fields having it and their respective getters missing it and vice
versa. Also fix the one instance where it was applied but formatted
differently than in the rest of the code base.
Change-Id: Ie4436fe64642affad7d775c59b10b6bacc0c4acd
Release-Notes: skip
Forward-Compatible: checked
|
|/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The Change.getServerid() is flagged as @Nullable and therefore
could lead to a NPE when performing a permission check upon
Gerrit clone.
Bug: Issue 338826950
Release-Notes: Fix NPE during git clone
Change-Id: I7d9978ceb8397c9dde5a41fa92af93bce54bf1d9
|
|\| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* stable-3.7:
Index Change: Remove misleading documentation
Document introduction of `virtual_id_number` to `ChangeInfo`
Release-Notes: skip
Change-Id: I33df959e79da89bba2b0ee7215c17de7cbfbb8c6
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
It has not been possible for change owners to index their own changes
for a while, but the documentation still claims that they can.
Bug: Issue 337070245
Release-Notes: Skip
Change-Id: I67a973da01b57c5cd779314132c7e586c92f6c9b
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Note that the underscore (`_`) has been removed from the variable name
to avoid resemblance to the prefix `_` used in `_number`, which
(according to the documentation) denotes deprecation.
Follow-Up-To: I85c946ef
Bug: Issue 325309573
Release-Notes: Add `virtual_id_number` to `ChangeInfo`
Forward-Compatible: checked
Change-Id: Iddf5dee45f80f863ec8dbc113e15e9499bfee757
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* Update plugins/codemirror-editor from branch 'stable-3.8'
to 6f916580c3f26ecfd9b2a10d96cb1c0530103bc6
- Fix xml mode
Change-Id: Ib419a2fadea2eb321559269b8e23ed14c0c20f3b
(cherry picked from commit 06719abee6c38d17008599f074050db5153bffa3)
- Add missing javascript language support
Whoops!
Change-Id: I56a493db26eb785769cc0cb3058a566f5fc4c9f6
(cherry picked from commit 469f617ed3102be445615396f3047e7ba498d261)
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
assignee is no longer a member of Change, but serverId is. Also,
revertOf and cherryPickOf can be null, so only count their weight when
non-null.
Change-Id: I7343670ccee83ef244a0fa6301980f2913c158e7
Release-Notes: skip
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This was missed in 1fba739a5f317e75383500e06d94ce12091e986d.
Change-Id: Ie84657e016bc69aa43b8602f6d9bd5a843c4f93c
Release-Notes: Removed missed reference to assignee config in docs
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
By using `flex: 0 0 auto` it makes sure that the text doesn't get cut
off. It's a little less ugly then it wrapping below.
This is a better fix than [0]. Correct fix is still a mobile optimised
header but that is a seperate change.
[0] I9da124be1ecd911cb16803fe709fb23a0860cd9e
Release-Notes: gr-main-header: fix header to wrap the better
Change-Id: Id7ebb8def48f5f58e31b3e2e297128fbbc8d0443
(cherry picked from commit 14b349c60cb3a90307670c1069e3e82136b785e4)
|
| | |
| | |
| | |
| | |
| | |
| | | |
Release-Notes: Add baseUrl to INITIAL_DATA in IndexHtmlUtil
Change-Id: I6a74f06eb1bb04365784722704043cac1b79541d
(cherry picked from commit b2b9d67de54218a3a45ea8bef3a12b95c602ed77)
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We include the base url in urlWithParams so we have to do the same
inside gr-rest-api by calling getBaseUrl().
Also clear cache in savePreferences().
Release-Notes: Fix clearing cache in gr-rest-api
Bug: 335613081
Change-Id: Ic0fb6af3316c8cd77bcaa64b802cd777cff339d7
(cherry picked from commit 6915b8aa40620176313ec3b55aef90e28a3e6936)
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We don't need to trigger this as the model already updates the diffPref
config and triggers a refresh.
Release-Notes: Remove reload-diff-preference unneeded event
Bug: 335613081
Change-Id: I2413cee2169d7347c7354af6131db28411197251
(cherry picked from commit e81286c3421227ea1e2636c9cd97e8fd349130c6)
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
I'm not sure if there's a better fix to this (it's kinda ugly)
but it at least prevents the buttons from disappearing.
The login button was disappearing when the title was big enough.
The correct fix would be to redesign the header to use a nav bar
that works on mobile e.g. using a hamburger icon and having a menu
that comes out. That's a bigger change.
Release-Notes: Fix buttons on header disappearing
Change-Id: I9da124be1ecd911cb16803fe709fb23a0860cd9e
(cherry picked from commit 9debe8c497697f93abf4fb311789ac7bf4b2369c)
|
| | |
| | |
| | |
| | |
| | | |
Release-Notes: skip
Change-Id: Ie256fb5368cc1a3fcf900ed1c5bfdbb0d63c52bf
|
| | |
| | |
| | |
| | |
| | | |
Release-Notes: skip
Change-Id: Ic8c6c64ac2d3ef994ffd60a9e4b8cfcc09caca66
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The "is:watched" predicate matches changes that are being watched.
To match changes that are being watched the "is:watched" predicate is
expanded to an OR query that has one "project:<watched-project>
<project-watch-filter>" predicate per project watch
("project:<watched-project>" is omitted if the project watch is the
All-Projects project, "<project-watch-filter>" is omitted if the project
watch doesn't set a filter). This expansion happens when the
IsWatchedByPredicate is instantiated. Expanding the query requires
loading the project watches of the user and parsing the filter (to
convert the filter string into Predicates).
If the filter of a project watch used the "is:watched" predicate,
querying changes by "is:watched" or checking whether a change matches
the project watch triggered an endless loop: If "is:watched" is used
ChangeQueryBuilder.parse creates an IsWatchedByPredicate instance (in
the is(String) method), which is expanded to the OR query (in the
IsWatchedByPredicate constructor), which requires parsing the project
watch filters. Parsing the project watch filters was done by
ChangeQueryBuilder.parse which creates another IsWatchedByPredicate
instance (in the is(String) method), which is again expanded to the OR
query (in the IsWatchedByPredicate constructor), which again requires
parsing the project watch filters, starting the loop anew.
To fix this we:
1. Disallow using "is:watched" in ProjectWatch.WatcherChangeQueryBuilder
which is a subclass of ChangeQueryBuilder.parse used to check whether
a change matches a project watch.
2. Change IsWatchedByPredicate to use
ProjectWatch.WatcherChangeQueryBuilder to parse the project watch
filters instead of ChangeQueryBuilder.
Using ProjectWatch.WatcherChangeQueryBuilder in IsWatchedByPredicate
makes the matching logic for project watches when a change is updated
consistent with the matching logic for project watches when "is:watched"
is used in a regular change query.
By disallowing "is:watched" in ProjectWatch.WatcherChangeQueryBuilder
project watches that use "is:watched" in their filter do not match any
change now. Before this change they triggered an endless loop, affecting
the stability of the service.
Note, IsWatchedByPredicate did have a check to prevent an endless loop
in this case, but it didn't work since the endless loop happened before
this check was reached.
Bug: Issue 321784734
Release-Notes: Fix endless loop when using "is:watched" in project watches
Change-Id: Ie38535b2df123a62dfd6a6e4b4ee60a80b0254f3
Signed-off-by: Edwin Kempin <ekempin@google.com>
|
|\| |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
* stable-3.7:
Suppress deprecation warning for LabelType.Builder#setFunction(...)
Release-Notes: skip
Change-Id: I3bde59d5f6300dc48f38754b713e2650a25c1d1d
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
As long as we have Rest API and tests for label functions we will use
this method and the compiler warnings makes it difficult to check
if new changes introduces other compiler warnings.
Release-Notes: skip
Change-Id: I788ef20091b39d72712ecfaed287a15af5921001
|
|\ \ \
| | | |
| | | |
| | | |
| | | |
| | | | |
* changes:
Emit value change event even when no items are matched
Fix gr-dropdown-list not updating on items change
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
In I6b5457f71 we kept the original logic, where the event is
only emitted if value changed and one of the items matched. But that
creates an issue if value and items are updated on separate iterations.
1. Value updates, no items -> so no event
2. Items update, no value change -> so no event
Even though in the end both value and items satisfy the requirement, the
event was never sent.
Since gr-dropdown-list is a utility class it generally makes sense for
it to be as simple as possible, and not try to account for logic of
components using it. The issue in gr-admin-view that didn't work with
this previously was that it doesn't maintain it's own invariants:
It was possible for the value set on gr-dropdown-list to not be in the
subsectionLinks, due to render happening in the middle of `reload()`.
We fix this issue in this change.
Release-Notes: skip
Google-Bug-Id: b/296106236
Change-Id: I9dfe61d6f7a247842e5392976378915eec6775f2
(cherry picked from commit 13948dfbb562033992d5b5346970de755a2a574a)
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
handleValueChanged is dependent on contents of `items` therefore it
needs to be recalculated if it changes.
items were being received/set later than path when loading gr-diff-view
from a url, resulting in no path shown in the gr-dropdown-list
We only emit an event if the new value matches one of the items. It
matches the behaviour before the change. It might make sense to emit the
value anyway, but that broke gr-admin-view tests, because they don't set
up rest api calls. So we keep it at the old behaviour.
Also, spotted some small indexing errors in admin-view tests and fixed
that.
Google-Bug-Id: b/296106236
Release-Notes: skip
Change-Id: I6b5457f71b573c9c56c425372cb0f4ec242e10f0
(cherry picked from commit ae6aa0f26a985d7ac34b3eef5be6ef24c1bbda86)
|
|\| | |
| | | |
| | | |
| | | | |
stable-3.8
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
1. Don't call handleValueChange if gr-drop-down-list hasn't actually changed value
Analysis:
- Navigate to https://go-review.googlesource.com/c/go/+/494187/29/src/runtime/trace2buf.go
- gr-diff-view eventually ends up with this.path = src/runtime/trace2buf.go.
- Press Shift+m
- handleNextUnreviewedFile gets called
- this.setReviewed(true) gets called
- this.navigateToUnreviewedFile('next')
- this.naveToFile gets called which computes newPath src/runtime/trace2cpu.go
- this.getChangeModel().navigateToDiff gets called wit the newPath
- Before the model has time to finish updating gr-diff-view, the header gets re-rerendered due to this.setReviewed(true)
- this.path gets updated based on the listener for this.getViewModel().diffPath$
- gr-dropdown-list.willUpdate is called based on the re-render of #8, it still has the old path. It's items' have changed but not its `path.
- This triggers the spin-locking back and forth.
Google-Bug-Id: b/296388347
Release-Notes: skip
Change-Id: I88196b34ddb037809ece6952555cb82e41171ad2
(cherry picked from commit 687f1203adf46755149dbdcec0069fb6b7571dfa)
|
|\ \ \ \
| | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* stable-3.7:
Avoid work in Optional.orElse() calls
Fix a bug with paper-tooltip.
Change-Id: I7016b62f40df0b61a05b6b78878e84296292e6e5
Release-Notes: skip
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Any code inside the orElse() will always be called, even if the result
isn't used because the optional is present. This is wasteful at a
minimum and can be actively harmful (as seen in [1]) because of side
effects or performance impacts. Fix that by replacing all orElse() calls
that create new instances or do non-constant work with either a call to
orElseGet() or an if statement.
It's possible this will result in performance improvements, but I didn't
attempt to measure any, so I'm not including a release note.
It would be nice if there were an ErrorProne checker for this, but one
doesn't exist yet.
[1] https://gerrit-review.googlesource.com/c/gerrit/+/417915/comment/0ea287cd_bfecb7f2/
Change-Id: Icab4115998fb4f3787ca74c0890949cb3ce5c6a6
Release-Notes: skip
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
We need to set properties, not attributes for paper-tooltip.
Google-Bug-Id: b/333416564
Release-Notes: skip
Change-Id: I48d41184d10aa0b2fd719717c38664792fa8bc54
(cherry picked from commit 4f7449246c08c9d6066ebbf5dca54423d07a9afc)
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
In our logs we found out event details with over 750k characters
and daily max around 100k characters. This is caused only by
dashboard displayed that can accumulate large rpcList when it is
refreshing without router change location.
This special case was introduced in Change 350774.
Release-Notes: skip
Google-Bug-Id: b/330373302
Change-Id: I9052621367ce3efe4c73f321b13bce53b6ddc1e5
(cherry picked from commit 6e6f42e4e7fd59b6e248ed1f45e8d14b5bff2334)
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
We need to set properties, not attributes for paper-tooltip.
Google-Bug-Id: b/333416564
Release-Notes: skip
Change-Id: I48d41184d10aa0b2fd719717c38664792fa8bc54
(cherry picked from commit 4f7449246c08c9d6066ebbf5dca54423d07a9afc)
|
|\| | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* stable-3.7:
Reset pageResultSize when PaginationType NONE back-fill results
Add test paginationType NONE to query only if there are more results
Fix paginationType NONE to run further queries only if needed
Run tests in FakeQueryChangesTest for paginationType NONE
Fix detection of invalid SSH key algorithm
Demonstrate SshKeyUtil fails to validate invalid SSH keys
Move comment of PaginatingSource.read()
Fix visibility filter for changes when paginationType NONE
Add tests for pagination to back fill the results
Allow plugins to use PredicateArgs
Release-Notes: skip
Change-Id: I80dcc0e671fe46e9ca98c99f09bd6503ff5d6fab
|
| |\ \ \
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
* stable-3.6:
Reset pageResultSize when PaginationType NONE back-fill results
Add test paginationType NONE to query only if there are more results
Fix paginationType NONE to run further queries only if needed
Run tests in FakeQueryChangesTest for paginationType NONE
Fix detection of invalid SSH key algorithm
Demonstrate SshKeyUtil fails to validate invalid SSH keys
Move comment of PaginatingSource.read()
Fix visibility filter for changes when paginationType NONE
Add tests for pagination to back fill the results
Allow plugins to use PredicateArgs
Release-Notes: skip
Change-Id: I78775c62d60bbdb63eb19dd8683179091c6ca447
|
| | |\ \ \
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
* stable-3.5:
Reset pageResultSize when PaginationType NONE back-fill results
Add test paginationType NONE to query only if there are more results
Fix paginationType NONE to run further queries only if needed
Run tests in FakeQueryChangesTest for paginationType NONE
Fix detection of invalid SSH key algorithm
Demonstrate SshKeyUtil fails to validate invalid SSH keys
Move comment of PaginatingSource.read()
Fix visibility filter for changes when paginationType NONE
Add tests for pagination to back fill the results
Allow plugins to use PredicateArgs
Release-Notes: skip
Change-Id: I8e166780d069267e3f25c454dc947ebe1d32d949
|
| | | |\ \ \
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
* stable-3.4:
Reset pageResultSize when PaginationType NONE back-fill results
Release-Notes: skip
Change-Id: I32a22fcbd7c84d07ff96df44975dd4c53259b94e
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
This is a follow up of change 413358.
In particular, if some results are skipped because of
of the visibility constraints, more changes need to
be asked from the index, and pageResultSize needs to
be reset to know how many results are returned from each
query.
Without this change querying for more changes could end
up in an infinite loop as exposed in a test in 3.6 [1]
[1]: https://gerrit.googlesource.com/gerrit/+/refs/heads/stable-3.6/javatests/com/google/gerrit/server/query/change/FakeQueryChangesTest.java#126
Release-Notes: skip
Change-Id: I566010c6f5bfcb4fbc003bc6693aa25d4dd44a81
|