See b/260945047.
Also refactored the corresponding tests, which should future updates easier.
This change should be deployed at or around 2023-02-15T16:00:00Z.
This PR removes all Ofy related cruft around `HistoryEntry` and its three subclasses in order to support dual-write to datastore and SQL. The class structure was refactored to take advantage of inheritance to reduce code duplication and improve clarity.
Note that for the embedded EPP resources, either their columns are all empty (for pre-3.0 entities imported into SQL), including their unique foreign key (domain name, host name, contact id) and the update timestamp; or they are filled as expected (for entities that were written since dual writing was implemented).
Therefore the check for foreign key column nullness in the various `@PostLoad` methods in the original code is an no-op as the EPP resource would have been loaded as null. In another word, there is no case where the update timestamp is null but other columns are not.
See the following query for the most recent entries in each table where the foreign key column or the update timestamp are null -- they are the same.
```
[I]postgres=> select MAX(history_modification_time) from "DomainHistory" where update_timestamp is null;
max
----------------------------
2021-09-27 15:56:52.502+00
(1 row)
[I]postgres=> select MAX(history_modification_time) from "DomainHistory" where domain_name is null;
max
----------------------------
2021-09-27 15:56:52.502+00
(1 row)
[I]postgres=> select MAX(history_modification_time) from "ContactHistory" where update_timestamp is null;
max
----------------------------
2021-09-27 15:56:04.311+00
(1 row)
[I]postgres=> select MAX(history_modification_time) from "ContactHistory" where contact_id is null;
max
----------------------------
2021-09-27 15:56:04.311+00
(1 row)
[I]postgres=> select MAX(history_modification_time) from "HostHistory" where update_timestamp is null;
max
----------------------------
2021-09-27 15:52:16.517+00
(1 row)
[I]postgres=> select MAX(history_modification_time) from "HostHistory" where host_name is null;
max
----------------------------
2021-09-27 15:52:16.517+00
(1 row)
```
* Allow anchor tenant creation via allocation token behavior
This also enforces that non-superusers cannot create registrations on
trademarked names prior to the sunrise period, even if they have an
allocation token with ANCHOR_TENANT behavior.
This includes:
- deletion of helper DB methods in tests
- deletion of various old Datastore-only classes and removal of any
endpoints
- removal of the dual-database test concept
- removal of 'ofy' from the AppEngineExtension
One of the more significant changes introduced in this PR is that we use
SQL as the backing database in all tests unless otherwise specified,
e.g. by using the TmOverrideExtension. We change various ofy-related
tests to use this.
This includes various changes:
- Deletion of SqlEntity/DatastoreEntity and related classes. Includes
any necessary changes because of that (e.g. getting a nice SQL key on
error in RegistryJpaIO).
- Deletion of classes that used libraries from the init-sql code
(RefreshDnsOnHostRenameAction)
- Removal of the JpaTransactionManager's backup implementation
- Modification of RegistryJpaWriteTest to not use init-sql code
- Removal of the Transaction class and related classes, however it does
not remove the TransactionEntity class as that would require DB
changes
- Removal of anything related to the actual usage of the database
migration schedule or read-only phases
- Various test changes and fixes to account for the differences in SQL
(like how foreign keys need to exist)
This deliberately doesn't do anything to alter the objects actually
stored in the DB yet, just how we use them
* Check PAK on domain create
* Add unit test
* update docs
* Remove unneccesary setup
* Fix blank line
* Add check and test to all relevant flows
* Change error message
* Add DS validation to match Cloud DNS
* Add checks to flows
* Add some flow tests
* Add tests for DomainCreateFlow
* Add tests for UpdateDomainCommand
* Fix docs test
* Small fixes
* Remove builder from tests
* Add NotLoggedInException tests to flows and flow docs
This wasn't included in flows.md before because the test existed in
ResourceFlowTestCase. So even though the exception could be thrown and
even though this was tested, it wasn't picked up in the documentation
because the documentation is picked up from the corresponding concrete
test class.
We want to keep the read-only-mode-exception as an unchecked exception,
so we introduce a temporary check in the EppController that provides a
specific error message for this situation (rather than letting it fall
through to the generic "command failed" messaging
* Rename client ID to registrar ID in most places
This is a code-only change, that shouldn't require any sort of data
migration. Correspondingly, there are some existing uses of clientId that are
not migrated (e.g. Datastore fields, task queue payloads, URL parameters for
actions that might be hit from task queues, etc.). And it of course doesn't
modify any fields in EPP XML. Note that the Cloud SQL schema fields are
already named using the registar_id pattern.
This also doesn't yet touch on the -c parameters in nomulus tools; that will be
coming later (since that is an external manual touch-point, it will require a
lot more in the way of changes to various meta scripts and documentation).
* Change more client IDs
* Merge branch 'master' into clientid-to-registrarid
* Migrate the documentation package to Java 11
The old Doclet API is deprected and removed in Java 12. This commit
changes the documentation package to use the new recommended API.
However it is not a drop-in replacement and there are non-idiomatic
usages all over the place. I think it is eaiser to keep the current code
logic and kind of shoehorn in the new API than starting afresh as the
return on investment of a do-over is not great.
Also note that the docs package is disabled as of this commit because we
are still using Java 8 to compile which lacks the new API. Once we
switch our toolchains to Java 11 (but still compiling Java 8 bytecode)
we can re-enable this package.
TESTED=ran `./gradlew :docs:test` locally with the documentation package
enabled.
* Always validate domain name on allocation token
This is in response to a client-reported error, where they accidentally sent the
wrong domain name on a domain create that included an allocation token. What
should have happened (and that now happens as of this commit) is an error being
thrown that the allocation token does not match the domain name being created.
What happened instead was that, since the incorrectly submitted domain name was
not reserved, the create succeeded (as it would for all creates of unreserved
domains in GA) and the allocation token was redeemed, which is not what you'd
expect.
* Fix tests to reflect changed check behavior
We should communicate to the users why this command failed, that they
are not allowed to use discounted allocation tokens on premium domains.
Currently it still fails, but we don't yet tell them why.
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
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
Adding it to one test is sufficient because we use the same loading logic across all flows.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=239506003
This removes the "create Domain/Host/Contact" forms that were supposed to be used instead of regular EPPs for CC-TLD that wanted to support it.
We're removing it because we don't use it and want to reduce unneeded code for the registry 3.0 migration.
Also, this is a security risk, as it allowed to do "billable actions" (creating a new domain for example) with the only authentication being access to the registrar's G Suite account.
This bypassed the certificate, IP whitelist, and EPP password, which is bad.
PUBLIC:
Remove the web console EPP endpoint
This removes the "create Domain/Host/Contact" forms that were supposed to be used instead of regular EPPs for CC-TLD that wanted to support it.
We're removing it because we don't use it and want to reduce unneeded code for the registry 3.0 migration.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=236244195
This also deletes the associated commands and domain application specific
entities.
We haven't used any of these TLD phases since early 2015 and have no
intent to do so in the future, so it makes sense to delete them now so we
don't have to carry them through the Registry 3.0 migration.
Note that, while there are data model changes, there should be no required
data migrations. The fields and entities being removed will simply remain
as orphans. I confirmed that the removed types (such as the SUNRUSH_ADD
GracePeriodType) are no longer used in production data, and left types
that are still used, e.g. BillingEvent.Flag.LANDRUSH or
HistoryEntry.Type.ALLOCATE.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=228752843
Registrars that are PENDING or SUSPENDED should not have access to
EPP commands that cost money, as in either case it's not likely we'd
actually be able to get payment from said registrar. For this reason
we already prevented access to the domain create flow for non-active
registrars. This commit extends that to other commands that cost
money, including renewals, restores, and transfer requests.
Note that implicit autorenews will still occur for suspended
registrars, as in our point-in-time data model there's no good way
to prevent them. So when a registrar is suspended for non-payment,
the game plan is to get all of their domains transferred out to a
registrar that will pay as soon as possible.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=223173677
EppConsoleAction still "manually" checks access by going over the
RegistrarContacts. We need it to use AuthenticatedRegistrarAccessor just like
every other part of the registrar console.
We still need to remove the (now unneeded) login EPP sent by the console, but that's left for a followup CL.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=222404208
This is only useful when we used the [] proxy because the GFE requires SNI during handshake in order to request the client certificate. The GCP proxy does not need this (it always requests the client certificate). We do not need to check for its existence.
Also removed the checking of internal headers for ssl cert hash used only by the [] proxy.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=213059027
We never used it and don't have any plans to use it going forward. All
conceivable parts of its functionality that we might use going forward have
already been subsumed into allocation tokens, which are a simpler way of
handling the same use case that are also standards-compliant.
Also gets rid of the hideous ANCHOR_ prefix on anchor tenant EPP authcodes
that was only ever necessary because of overloading the authcode for
anchor tenant creation. Going forward it'll be based on allocation tokens,
so there's no risk of conflicts.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=209418194
Also prevents signed marks from being used in non-sunrise TldStates.
Currently, we send out a Lordn update only when there's a ClaimNotice, or if
we're in end-date sunrise.
But EPPs can contain a SignedMark instead of a ClaimsNotice for trademarked
domains - in which case we aren't sending out Lordn update. This also applies
to start-date sunrises.
We also change the SignedMark behavior for superusers. Currently, if a
mismatched signed mark is given as superuser, we accept it. That causes
problems when we want to send the Lordn update.
Instead - we no longer allow superusers to give a mismatched SignedMark (just
as we don't allow users to give a bad ClaimNotice). A super user can still
create a domain WITHOUT a signed mark - but if one is provided, it must match.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=199783411
Superuser can also execute contact info commands. AuthInfo is no longer checked in the input and always displayed in the output as the only ones who can get a response are the sponsoring registrar and super user.
Also corrected a Javadoc in which '@' should have been escaped (see https://reflectoring.io/howto-format-code-snippets-in-javadoc/)
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=199521153
This verifies anchor tenant creates by registrars are for exactly 2 years, to simplify billing down the line (anchor tenants get 2 years of free domain creates).
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=193931858
A "mark" tells us that the holder owns the trademark for a given domain name. It is signed for authentication.
If the signature's certificate is either "not yet valid" or "expired", we return explicit errors to that effect.
But in addition to the signature's certificate, the mark itself might not be valid yet or already expired. Right now if that happens - we return an error saying "the mark doesn't match the domain name".
That is wrong - as the mark can match the domain name, just be expired. Returning "the mark doesn't match the domain name" in that case is misleading.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=190069976
Currently we validate the fee extension by summing up all fees present in the extension and comparing it against the total fee to be charged. While this works in most cases, we'd like the ability to individually validate each fee. This is especially useful during EAP when two fees are charged, a regular "create" fee that would also be amount we charge during renewal, and a one time "EAP" fee.
Because we can only distinguish fees by their descriptions, we try to match the description to the format string of the fee type enums. We also only require individual fee matches when we are charging more than one type of fees, which makes the change compatible with most existing use cases where only one fees is charged and the description field is ignored in the extension.
We expect the workflow to be that a registrar sends a domain check, and we reply with exactly what fees we are expecting, and then it will use the descriptions in the response to send us a domain create with the correct fees.
Note that we aggregate fees within the same FeeType together. Normally there will only be one fee per type, but in case of custom logic there could be more than one fee for the same type. There is no way to distinguish them as they both use the same description. So it is simpler to just aggregate them.
This CL also includes some reformatting that conforms to google-java-format output.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=186530316
The START_DATE_SUNRISE phase allows registration of domains only with a signed mark. In all other respects - it is identical to the GENERAL_AVAILABILITY phase.
Note that Anchor Tenants bypass all checks, and are hence able to register domains without a signed mark.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=185534793
Next up is adding custom logic so that the results of these checks can be
more meaningful.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=181660956
The next step is to add them for domain checks as well (which is simpler
because it doesn't involve validation).
This requires the addition of a TrimWhitespaceAdapter for XML JAXB objects,
which will prove useful for other @XmlValue attributes in the future.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=181526726
The extension itself isn't used for anything yet; allocation tokens can be
passed for domain creates and checks but are ignored if present. This will
be changed in a subsequent CL that adds AllocationToken entities and related
logic. Usage of this extension in any other EPP flow will throw an
UnsupportedExtensionException.
The relevant spec is https://tools.ietf.org/html/draft-gould-allocation-token-04
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=181343188
Failing to use the fee extension during EAP can result in charges to registrars
that are radically different than what they may have been expecting.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=177597883
This way it is consistent with the rest of our domain-related flows, which
consistently use the Domain* prefix. Note that claims checks are just a
special case of domain checks anyway, which run under DomainCheckFlow. This
will make dashboards looking at domain commands "just work" with a regexp of
Domain.*, without having to special-case in ClaimsCheck.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=172608964
Normally, if a domain is in the auto-renew grace period, a transfer will cancel the auto-renew billing event. In the event of a transfer with no change to registration end date, the auto-renew billing event should not be cancelled and the gaining registrar should not be charged for the transfer.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=170576726
Specifically, this prevents suspended registrars from creating domains or applications. Pending registrars already can't perform these actions because they get an error message when attempting to log in.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=170481338
