This increases the test coverage of AbandonUtil from 15% to 75%.
Change-Id: Ifc1c3c98a0e9557f568a0ea4643e572dd0af200d
Signed-off-by: Edwin Kempin <ekempin@google.com>
ChangeIT is growing too big which is why some tests should be moved out.
Change-Id: I95789af43d2c782a29b78ad079152f063cc3b29c
Signed-off-by: Edwin Kempin <ekempin@google.com>
There was a formatting issue caused by delete buttons in a draft change.
It really does not make sense for the button to be there for drafts
anyway. This change hides the delete button for admins on draft changes.
Bug: Issue 6257
Change-Id: I236eff9db1f0ca76ed658f3048132c9c675ee407
On keyboard completion, the autocomplete did not re-focus on the input.
This change fixes that, and prevents any regression by adding asserts
to the test suite.
Bug: Issue 6253
Change-Id: I76bae30e069d28c00d3029fe36ff9d731fd7631b
During the migration step of the ESLint project, there were a few rules
that were stricter than need be -- specifically, the rules that force
devs to not use indexOf and Array.forEach.
There are several legitimate cases for using these functions, and as
all of the instances where ES6 features like Array.includes and for...of
would work better have been addressed, these rules ought to be removed.
Bug: Issue 6179
Change-Id: Ib8c730a98c8d4a22999b1669dc4ce1a5a85fd749
* stable-2.14:
UploadValidationListener: Use Java 8 default on interface methods
Enhance UploadValidators to allow listening to negotation start
Change-Id: I1fddbf0448c182e3a73052afbe5891a64eb79c0a
* changes:
SubmoduleOp: Preserve exception from constructor
SubmoduleOp: TODO format cleanup
SubmoduleException: Minor cleanup
RetryHelper: Make time limits configurable
RetryHelper: Disable retrying when updates are non-idempotent
Exceptions from assisted-inject constructors get wrapped in
ProvisionException, so there's no guarantee that a higher-level handler
will be able to reasonably unwrap the exception message that's intended
for the user. These messages are important, and
SubmoduleSubscriptionsWholeTopicMergeIT explicitly checks for them, but
it's basically an accident that this was working currently.
SubmoduleOp really wants to do real work in its constructor to
initialize the sortedBranches map. Simply moving that map out of the
constructor doesn't work, because it depends in subtle ways on some of
the other fields which are mutable over the course of the instance's
lifetime. In an ideal world this subtle logic would get cleaned up, but
for now, do the simplest thing that fixes the exception-throwing
behavior, since this is preventing us from converting MergeOp to use
RetryHelper.
Change-Id: I4f6a22a95ecc533166ebf6c1518c2a488fe3422b
Use an arbitrarily high limit during NoteDbOnlyIT, where we explicitly
want retries to happen, in order to avoid spurious timeouts caused by
scheduling delays on overloaded test machines. For other tests,
LockFailureExceptions generally shouldn't crop up during updates anyway,
and if they do, we would rather the updates time out in finite amount of
time so we can see the resulting stack trace.
Change-Id: I5bad32749b994e6473176cdd45cc80d9d4e195eb
This includes everything except NoteDb fused mode, matching the old
behavior. In practice, RetryHelper wouldn't have retried in
ReviewDb-only mode previously, because BatchUpdate wouldn't throw
LockFailureException. But this is essentially an implementation detail;
we shouldn't count on that exception type never being thrown. Safer to
explicitly disable retries in this case.
Change-Id: Ibf502ac8157a1571b796d910862e3e69e6a0560c
In the change object, the revision number is represented as an integer.
In the download dialog, patchNum is passed as a string. Due to strict
equality comparison, the download schemes for a particular patchset were
never found.
This change parses the patchNum to an int every time an attempt to
compare it is made.
To protect against regressions, all of the mocked patchNum values in the
test suite have been changed to string values to more accurately reflect
the actual data.
Bug: Issue 6224
Change-Id: I3626cda9eaad48c05fb97efd97ba313f20ab5473
Previously, if a change status was 'NEW' the function that determines
if 'Merge Conflict' should be displayed was not called. This change
ensures that it is not bypassed if a change is 'NEW'
Change-Id: Ia014452be5d7b8a24abb6a8b29483dd5289cc515
Previously, if a change status was 'NEW' the function that determines
if 'Merge Conflict' should be displayed was not called. This change
ensures that it is not bypassed if a change is 'NEW'
Change-Id: Ia273aa846682e7e4453370314044888c8e8c7f99
