The TestServerRule object is shared between retries in the same test,
so the testServer object constructed in TestServerRule's constructor
will be shared as well. This should be the reason why the test retry
carries over some state. (The log in the test proves that the
testServer object is shared because it listens to the same port in
all retries, which should not happen if its constructor is invoked
every time. You can find multiple "TestServerRule is listening on:
[]in this test
[]
So, this CL delayed the construction to rule.before() method which is
invoked before every retry.(You can see each retry has a test server
listening to different port and the error is "It differed by <16> pixels."
for all attempts instead of not clickable button []
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246904446
Bazel's GenTestRules does not support the "timeout" parameter, introduced in []
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246830344
Check the timing (that is, whether or not we're in a promotion), the allowed registrar client IDs, and the allowed TLDs.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246824080
The two behave identically, and containsAllOf is being removed.
More information:
[]
Tested:
TAP --sample for global presubmit queue
[]
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246791491
This CL looks good? Just LGTM and Approve it!
This CL doesn't look good? File a bug using []
What else can you do?
* Suggest a fix on the CL ([]
* Revert this CL, by replying "REVERT: <provide reason>".
* Revert this CL and not get a CL that cleans up this file in the future by replying "BLACKLIST: <provide reason>". This action is not reversible!
This CL was automatically generated by ATTT ([] using CL Shepherd ([] Please contact [] for questions.
//third_party/java_src/gtld/javatests/google/registry/model:registry/label/GenrulePremiumListTest: We're adjusting the test target timeout because it has a 1.248050 percent time out rate based on last week's (2019-04-27 - 2019-05-04) tap postsubmit running statistics. Its original timeout_sec is 60, and we're adjusting it to moderate, which is 300 seconds.
Some detailed statistics for the target:
Num of invocation: 641
Num of passed invocations: 633
Num of timeout invocations: 8
****** Stats for passing runs ******
Min run time in secs: 14
Max run time in secs: 57
Mean run time in secs: 28.955766
Sd run time in secs: 9.104767
#codehealth
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246791200
The overload that accepts a description generally produces better failure messages:
- The first line of the message it produces is something like: "value of: myProto.getResponse()" (where "getResponse()" is taken from the provided description)
- The last line of the message it produces is something like: "myProto was: response: query was throttled" (the full value of myProto)
- And the existing text goes in between.
Additional motivation: We are deleting the no-arg overload externally (and probably internally thereafter).
Tested:
TAP --sample ran all affected tests and none failed
[]
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246582329
The two behave identically, and containsAllIn is being removed.
More information:
[]
Tested:
TAP --sample for global presubmit queue
[]
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246565417
They behave identically, and the old names are being removed.
Open-source note: The new methods are available in Truth as of version 0.44.
END_PUBLIC
More information:
[]
Tested:
TAP --sample ran all affected tests and none failed
[]
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246550001
Overriding getter methods to change values is a bit overkill when these values
are static (don't change based on internal state).
Just setting them in the base class' constructor is simpler.
Also, we can read the PATH of an Action based on the Annotation instead
returning it manually for each Action.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246135754
Part of the attempt to remove or suppress warnings for deprecated
API use, which will make the Gradle project usable with Intellij.
Currently in the Intellij/Gradle setup, deprecation warnings cause
Intellij build process to fail. Passing -Werror:none flags to javac\
does not have any effect.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=246135737
(The exact change is slightly different in some cases, like when using custom subjects or check(), but it's always a migration from named(...) to [assert]WithMessage(...).)
named(...) is being removed.
This CL may slightly modify the failure messages produced, but all the old information will still be present.
More information:
[]
Tested:
TAP --sample for global presubmit queue
[]
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=245762634
For example:
assertThat(actual().foo()).isEqualTo(expected);
becomes:
check("foo()").that(actual().foo()).isEqualTo(expected);
More information:
[]
Tested:
TAP --sample for global presubmit queue
[]
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=245750144
We will need to be able to find all redemptions associated with a particular token and this will allow us to do that.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=245292865
The latest version of the selenium/standalone-chrome docker image appears to
be out of sync with the API that we're using and there is no more recent
version of the API available in maven central.
This change pins us to the earlier version and fixes our CI build.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=245286532
As it turns out, using Maps::transformValues does not allow us to change the
resulting map--calling Map::put throws an UnsupportedOperationException. As a
result, we have to do this roundabout stream-collect to do a group-by.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=244852373
Migrate Truth subjects from the old fail(String, Object) to the new failWithActual(String, Object), tweaking verbs for the new grammar.
Before:
fail("has foo", expected);
After:
failWithActual("expected to have foo", expected);
Open-source note: The fail*() methods used by this CL were added in Truth 0.41.
More information:
[]
Tested:
TAP --sample for global presubmit queue
[]
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=244418884
The test changes are perhaps a little bit more involved than expected, because I
had to add a second RegistrarContact which had knock-on effects on other tests.
This does make the other tests better though, in that we're now testing registrars
with multiple contacts (we weren't much before).
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=243827803
This also removes the related setting on the Registry entity. We never used
either of these, and keeping them around in our codebase just adds complexity
for no purpose. We already achieve the goals of this feature by restricting
nameservers on entire TLDs and by registry-locking important domain names.
This is also two fewer things we'll have to worry about carrying over to the new
schema in Registry 3.0.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=243816241
We haven't started dealing with timing or discounts yet, but unlimited use tokens should actually be unlimited use
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=243318266
Collecting by key leads to exceptions if there are multiple client IDs with the
same email address (if we group by client ID in the pipeline). Using
Multimaps::index means that if we're grouping by email, all matches with the
same email get concatenated together
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=242858112
From to the RDAP response profile:
2.4.5. Abuse Contact (email, phone) - an RDAP server MUST include an *entity*
with the *abuse* role within the registrar *entity* which MUST include *tel*
and *email*, and MAY include other members
Even though this is a MUST - this field will only be shown if the registrar has a *visible* abuse contact.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=242684303
NewRegistrar will still have the email address "new.registrar@example.com" and TheRegistrar will now have the email address "the.registrar@example.com".
I noticed this when looking at the Spec11 testing code and this will make it easier to test that code later when we retrieve email addresses from the registrar datastore objects themselves.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=242676543
It isn't used now, and as far as I can tell it was never used for anything, so
get rid of it. We won't be using it in our new Cloud SQL schema anyway.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=242501877
We also ignore events that happened before the domain was created (for example, in a previous incarnation of the same domain name) and we set the last changed event to be the later of the last EPP change and any other event that happened before "now".
From RDAP response profile
2.3.2 The domain object in the RDAP response MAY contain the following events:
2.3.2.3 An event of *eventAction* type *transfer*, with the last date and time that the
domain was transferred. The event of *eventAction* type *transfer* MUST be omitted if
the domain name has not been transferred since it was created.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=242461310
Since we're now sending these to actual registrars, change the text to make it
a little bit clearer what's going on.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=241992717
The Registrar object's @ID is on its clientId field, so that's what we're actually using here, not the name.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=241936005
This is a follow-up to [] We can't set registrars as DISABLED until
this is deployed.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=241767990
Previously we had a few customized Gradle build task to manage
the Docker container for provisioning browser and ChromeDriverService
used by WebDriver tests. This CL changed to use a java library
from testcontainers.org to achieve the same purpose. The main
benefit of it is that we can expect to run the WebDriver tests
from IDE going forward.
Also, this CL refactored the structure of WebDriver related classes
to have JUnit rule to manage the lifecycle of WebDriver instance,
this is also compatible with the API from testcontainers library.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=241539861
We're changing rdapConformance from rdap_level_0 to icann_rdap_response_profile_0, and changing the EPP status notices title from "EPP Status Codes" to "Status Codes" to conform with the changing spec.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=241527211
Because we make use of previous reports, it's easier to add this in and then use it later as two steps, rather than trying to add this and use it in one change. This way we don't really need to deal with backcompat issues.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=241378853
*** Reason for rollback ***
The inconsistent class loading is breaking the tests
*** Original change description ***
Validate provided email addresses when creating a Registrar
***
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=241014945
The JSON object toString isn't necessarily deterministic in its ordering of the keys, so we can only rely on length
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=241003060
Obviously this is a bad thing and would fail if it ever happened. If this does occur, we will send a warning email.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=240977242
This CL is to address the public static field in RepeatableRunner
for caller to get the current attempt number. We tried to have
a JUnit TestRule to achieve the purpose but it ended up with having
a RuleChain in each class where we already have multiple rules and
need to add the retry rule. This is because we have to make sure
the retry rule is the last one to wrap the test statement so that
the actual retry can include the actions defined in other rules.
Having a rule chain is not scalable and confuses engineer so we
gave it up.
Instead, we decided to expand the current RepeatableRunner to
use reflection to inject the attempt number to the test class.
Doing it this way can reduce the burden from the caller and it also
gets rid of the global state from the previous public static field.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=240789045
This way the error messages are more sensible when a registrar doesn't exist
(which realistically shouldn't happen in the typical case anyway).
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=240376239
We'll have a separate change to make sure we're not actually trying to email these folks, but this will make it so that the entire pipeline doesn't crash. The test makes sure that we can run the pipeline properly with these empty strings.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=240346954
