test(spanner): optimize and fix system tests and enable in CI

[cy-yun]

This PR addresses and resolves the Spanner emulator system test flakiness and optimizations outlined in Issue #9230, enabling them to run more reliably in CI environments.

Key Changes:

• BackupTest.php Optimizations:
  • Added the @group flakey annotation so these tests can be safely skipped in standard CI runs.
  • Removed redundant Long Running Operation polling in testCancelBackupOperation and testDeleteBackup to significantly reduce overall execution time.
  • Fixed testCreateBackupRequestFailed flakiness where a pending backup from previous tests threw a FailedPreconditionException.
  • Fixed multiple PHPCS linter formatting errors (multi-line function arguments & trailing whitespace).
• Idempotent Mutations for Emulator Resilience:
  • Swapped insert() to insertOrUpdate() in TransactionTest.php and PgTransactionTest.php.
  • Changed insert() to insertOrUpdateBatch() in PartitionedDmlTest::testPdml().
  • Reason: The Spanner emulator occasionally drops the connection and returns an ABORTED error after it has already internally committed a mutation. When our client's runTransaction loop catches the abort and automatically retries the transaction closure, it was failing with a ConflictException (ALREADY_EXISTS). Making the mutations idempotent handles this retry edge-case gracefully.
• Core LRO Deprecation Fix:
  • Fixed a PHP 8.1 deprecation warning ("using null as array offset" / passing null to string parameter) in LongRunningClientConnection::resumeOperation by properly omitting the $method argument when it's not explicitly provided. This resolves the warning without breaking any internal mock expectations (like those in InstanceConfigurationTest).
• CI Configuration:
  • Removed Spanner from the phpunit-system.xml.dist exclusion list so that non-emulator system tests can be successfully executed in Kokoro.

🔗 Related Issues

• Fixes Improvements to Spanner System Tests #9230

[Hectorhammett]

Did the linter suggested this change? This change feels wrong to me. I mean it still is valid, but I would expect either the previous iteration or something like:

$op->pollUntilComplete(
    [
        'timeoutMillis' => ...
    ]
)

As it is right now, having the braces on the same level as the key feels off.

[Hectorhammett]

But I still think the previous one should be valid 🤷

[cy-yun]

I fixed the formatting of the code such that it is consistent with the original & phpcbf is not upset

[Hectorhammett]

Wait, isn't this testing that if you execute a backup deletion and cancel it, it shouldn't be still available? Or am i misreading the test?

[cy-yun]

Actually, this test is meant to test canceling a backup creation operation, not a deletion. Previously, the test was calling $op->pollUntilComplete() before $op->cancel(). That meant it was waiting for the backup to fully finish creating, and then issuing a cancel command on a completed operation (which basically does nothing), hence why it used to assert that the backup still existed.

By removing the polling, we are now properly canceling the operation while it is still in-progress. Because we cancelled it mid-flight, Spanner usually drops the incomplete backup, so asserting that it still exists makes the test extremely flaky.