JNI: Do not create 8M block cache for negative blockCacheSize values

[javeme]

As BlockBasedTableConfig setBlockCacheSize() said, If cacheSize is non-positive, then cache will not be used. but when we configure a negative number or 0, there is an unexpected result: the block cache becomes 8M.

• Allow 0 as a valid size. When block cache size is 0, an 8MB block cache is created, as it is the default C++ API behavior. Also updated the comment.
• Set no_block_cache true if negative value is passed to block cache size, and no block cache will be created.

[adamretter]

LGTM. Although is 0 actually a valid choice? If so we should preserve 0

[sagar0]

0 is a valid size to block cache and I feel in such a case we should fall back to the default behavior on the C++ API side, which is to create a 8MB block cache when block_cache is set to nullptr.
If you don't want the block cache to be used, you need to use setNoBlockCache(true).

rocksdb/include/rocksdb/table.h

Lines 129 to 136 in 6d113fc

	// Disable block cache. If this is set to true,
	// then no block cache should be used, and the block_cache should
	// point to a nullptr object.
	bool no_block_cache = false;
	// If non-NULL use the specified cache for blocks.
	// If NULL, rocksdb will automatically create and use an 8MB internal cache.
	std::shared_ptr<Cache> block_cache = nullptr;

[javeme]

@sagar0 @adamretter Thanks for the review, I have updated the code:

1. allowed 0 as a valid size and updated the comment.
2. set no_block_cache true if negative number is passed.

[adamretter]

@javeme LGTM

[facebook-github-bot]

@sagar0 has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

[facebook-github-bot]

@sagar0 merged this pull request in c92c58f.