Migrate TMCH, SMD, and Fee models to java.time

[CydeWeys]

This commit completes the java.time migration for the Trademark Clearinghouse (TMCH), Signed Mark Data (SMD), and Mark models, the EPP Response & InfoData Objects, and the Fee Extension ecosystem, transitioning them from Joda-Time's DateTime to java.time.Instant.

Changes made:

1. Core Models Migration:

   • ClaimsList, TmchCrl, SignedMark, Trademark, BaseFee and its extensions (FeeCreateCommandExtension, FeeUpdateCommandExtension, FeeTransferCommandExtension, FeeRenewCommandExtension, FeeCheckResponseExtension etc.) have all been migrated to natively store, parse, and return Instant.
   • Deprecated toDateTime bridge methods (e.g. BaseFee's getValidDateRangeInstant) have been removed.

2. Flow Integrations & Utilities:

   • DomainFlowUtils, DomainPricingLogic, and CheckApiAction seamlessly integrate with java.time.Instant for fee requests, effective dates, pricing lookups, and premium checks.
   • java.time.Duration is natively utilized instead of Joda-Time's Duration.
   • Added plusMonths(Instant, int) and minusMonths(Instant, int) to DateTimeUtils to safely handle monthly temporal additions on Instant boundaries.

3. XML Binding and Serializers:

   • Extensively applied the updated UtcInstantAdapter in package-info.java definitions under fee*, smd, mark, contact, and eppoutput to securely marshal and unmarshal ISO-8601 strings into Instant representations.

4. Parsers:

   • Both ClaimsListParser and SmdrlCsvParser now strictly validate and build objects utilizing Instant.parse().

5. Testing Ecosystem:

   • Resolved [TruthIncompatibleType] warnings where DateTime comparisons against Instant were previously triggering ErrorProne. Added Instant assertions and toDateTime wraps across test suites including DomainSubject and AbstractEppResourceSubject.
   • Simplified fully qualified imports (java.time.Duration, START_INSTANT, END_OF_TIME) and replaced FakeClock.nowUtc() with FakeClock.now().
   • Replaced toDateTime(clock.now()) with clock.nowUtc() and toDateTime(END_INSTANT) with END_OF_TIME across tests.

This change is 

[CydeWeys]

PTAL, ready for code review. The approach I'm taking here is a little different than in previous PRs; you might see some method calls that perform a toInstant() or toDateTime() conversion from the call site. I'd rather do it this way than expand the scope of the PR and add in more function overloads that will only need to be later removed. It seems cleaner to either fully migrate any given function over, or to convert it at callsites as necessary, rather than overloading everything with two versions and adding additional churn. This PR is already creaking at the seams; expanding the scope any further is unlikely to lead to good results.

[CydeWeys]

This PR is already large enough so I won't do it here, but I am considering adding some methods to the Clock interface to make the following pattern nicer. E.g. instead of:

plusDays(clock.now(), 5)

You'd be able to do:

clock.nowPlusDays(5)

Or maybe something like a fluent method:

clock.plusDays(5).plusHours(5).now()

(crucially, none of these would change the clock's state, just make it easier to perform date arithmetic on the current time)

[gbrodman]

possibly but people are going to still have the instinct to do something like clock.now().plusDays(1) just because the usual pattern is to get the "current" datetime in some way then modify it. It's probably not good to have both intermixed.

actually yeah i'm not a big fan of having to use DateTimeUtils to do any sort of big datetime addition/subtraction. Is it necessary?

@gbrodman reviewed 101 files and all commit messages, and made 10 comments.
Reviewable status: all files reviewed, 16 unresolved discussions (waiting on CydeWeys).

---

-- commits line 1 at r3:
this commit message is too long and verbose to the point where it's counterproductive

it's TOTALLY SHOCKING that AI wrote something too overly verbose

---

core/src/main/java/google/registry/flows/domain/DomainFlowUtils.java line 591 at r3 (raw file):

      @Nullable BillingRecurrence billingRecurrence)
      throws EppException {
    DateTime now = currentDate;

make this into an instant here so that we only need to do the conversion once

---

core/src/main/java/google/registry/flows/domain/DomainTransferApproveFlow.java line 166 at r3 (raw file):

                                Tld.get(tldStr),
                                targetId,
                                toInstant(transferData.getTransferRequestTime()),

get as an instant instead

---

core/src/main/java/google/registry/flows/poll/PollRequestFlow.java line 81 at r4 (raw file):

        .setMessageQueueInfo(
            new MessageQueueInfo.Builder()
                .setQueueDate(toInstant(pollMessage.getEventTime()))

pls fix

---

core/src/test/java/google/registry/flows/domain/DomainTransferRequestFlowTest.java line 343 at r4 (raw file):

    assertThat(domainAutorenewEvent.getRegistrarId()).isEqualTo("TheRegistrar");
    assertThat(domainAutorenewEvent.getRecurrenceEndTime())
        .isEqualTo(toDateTime(implicitTransferTime));

get as instant

---

core/src/test/java/google/registry/flows/domain/DomainTransferRequestFlowTest.java line 384 at r4 (raw file):

        getOnlyPollMessage(
            "NewRegistrar", toDateTime(expectedExpirationTime), PollMessage.Autorenew.class);
    assertThat(transferApprovedPollMessage.getEventTime())

can get as instant

---

core/src/test/java/google/registry/flows/domain/DomainTransferRequestFlowTest.java line 386 at r4 (raw file):

    assertThat(transferApprovedPollMessage.getEventTime())
        .isEqualTo(toDateTime(implicitTransferTime));
    assertThat(autorenewPollMessage.getEventTime()).isEqualTo(toDateTime(expectedExpirationTime));

can get as instant

---

core/src/test/java/google/registry/flows/domain/DomainTransferRequestFlowTest.java line 414 at r4 (raw file):

            .collect(onlyElement());
    assertThat(losingTransferPendingPollMessage.getEventTime()).isEqualTo(clock.nowUtc());
    assertThat(losingTransferApprovedPollMessage.getEventTime())

can get as instant

---

core/src/test/java/google/registry/testing/DatabaseHelper.java line 481 at r4 (raw file):

            getDomainRenewCost(
                domain.getDomainName(),
                google.registry.util.DateTimeUtils.toInstant(costLookupTime),

lol import

[gbrodman]

also it looks like there are many cases where we replace "clock.nowUtc()" with "toDateTime(clock.now())" which doesn't seem to really help things

@gbrodman made 1 comment.
Reviewable status: all files reviewed, 16 unresolved discussions (waiting on CydeWeys).

[CydeWeys]

@CydeWeys made 9 comments and resolved 7 discussions.
Reviewable status: 87 of 102 files reviewed, 9 unresolved discussions (waiting on gbrodman).

---

-- commits line 1 at r3:

Previously, gbrodman wrote…

this commit message is too long and verbose to the point where it's counterproductive

it's TOTALLY SHOCKING that AI wrote something too overly verbose

Done.

---

core/src/main/java/google/registry/flows/domain/DomainFlowUtils.java line 591 at r3 (raw file):

Previously, gbrodman wrote…

make this into an instant here so that we only need to do the conversion once

Done.

---

core/src/main/java/google/registry/flows/domain/DomainTransferApproveFlow.java line 166 at r3 (raw file):

Previously, gbrodman wrote…

get as an instant instead

Done.

---

core/src/main/java/google/registry/flows/poll/PollRequestFlow.java line 81 at r4 (raw file):

Previously, gbrodman wrote…

pls fix

Done.

---

core/src/test/java/google/registry/flows/domain/DomainTransferRequestFlowTest.java line 343 at r4 (raw file):

Previously, gbrodman wrote…

get as instant

Done.

---

core/src/test/java/google/registry/flows/domain/DomainTransferRequestFlowTest.java line 384 at r4 (raw file):

Previously, gbrodman wrote…

can get as instant

Done.

---

core/src/test/java/google/registry/flows/domain/DomainTransferRequestFlowTest.java line 386 at r4 (raw file):

Previously, gbrodman wrote…

can get as instant

Done.

---

core/src/test/java/google/registry/flows/domain/DomainTransferRequestFlowTest.java line 414 at r4 (raw file):

Previously, gbrodman wrote…

can get as instant

Done.

---

core/src/test/java/google/registry/testing/DatabaseHelper.java line 481 at r4 (raw file):

Previously, gbrodman wrote…

lol import

Done.

[CydeWeys]

Agreed that I don't love the common pattern of static import and then call plusDays(..., 5) ... but the alternatives also aren't amazing. We don't have to address it in this PR (in fact I'd rather not, as it's more easily changed through IntelliJ), but another alternative would be something like .plus(Duration.ofDays(5)). ... would it be dumb to make a statically importable wrapper function around Duration.ofDays() that could make that look like .plus(days(5))? The problem with Duration.ofDays() is that in a lot of files that are still using the Joda version it actually has to look like .plus(java.time.Duration.ofDays(5)), which is obviously sub-ideal.

Anyway, we can solve this later -- it's trivial to change the implementation of plusDays() to whatever we like, and then use IntelliJ to inline it everywhere and delete it.

@CydeWeys made 1 comment.
Reviewable status: 87 of 102 files reviewed, 9 unresolved discussions (waiting on gbrodman).

[CydeWeys]

PTAL

@CydeWeys made 1 comment.
Reviewable status: 87 of 102 files reviewed, 9 unresolved discussions (waiting on gbrodman).