feat(storage): add full object checksum validation for grpc flow

[Dhriti07]

Adding full object checksum for grpc flow

Refer to: go/full_checksum_java

[gemini-code-assist]

Code Review

This pull request introduces cumulative checksum validation for full object reads in the GapicUnbufferedReadableByteChannel. It wraps the existing hasher in a CumulativeHasher when a read starts at offset 0 and triggers validation once the end of the stream is reached. The update includes several new test cases to verify validation behavior across different scenarios, such as multi-chunk reads and ranged reads. Review feedback suggests optimizing the implementation by only instantiating the CumulativeHasher when no read limit is specified, thereby avoiding unnecessary overhead for ranged reads where validation is skipped.

[gemini-code-assist]

The CumulativeHasher is currently instantiated for any read starting at offset 0, even if a readLimit is provided. However, the test case validateCumulativeChecksum_skippedForRangedRead demonstrates that validation is skipped when a limit is present. To avoid unnecessary object allocation and wrapping overhead for ranged reads, the CumulativeHasher should only be instantiated when both the offset is 0 and no limit is specified (i.e., a full object read).

Suggested change

	this.hasher = (req.getReadOffset() == 0)
	? new CumulativeHasher(hasher, 0, req.getReadLimit() <= 0 ? OptionalLong.empty() : OptionalLong.of(req.getReadLimit()))
	: hasher;
	this.hasher = (req.getReadOffset() == 0 && req.getReadLimit() <= 0)
	? new CumulativeHasher(hasher, 0, OptionalLong.empty())
	: hasher;