Collection context in the cache should always include indexes #138
Comments
apkar
added a commit
to apkar/fdb-document-layer
that referenced
this issue
Apr 4, 2019
…ays include indexes. getUnboundCollectionContext() should always return context with indexes included. As this function plays with cache, its good to not play with how it creates collection context. Instead client of this function can make a copy and do anything with it. Otherwise, we will find interesting inconsistencies in the metadata cache.
apkar
added a commit
that referenced
this issue
Apr 4, 2019
Resolves #138: Cache should always include collection context with indexes
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
getUnboundCollectionContext()takes a parameterincludeIndexto decide whether to include index to the collection context or not. Having a collection context without all the indexes could be useful, #137. But, this function also deals with the metadata cache. Even though the current caller doesn't care about indexes future callers might, as this function caches the context even the future callers wouldn't see the indexes. This is very bad for index consistency. We should just removeincludeIndexparameter altogether and the caller can always make a copy of collection context and do anything they want to that copy.The text was updated successfully, but these errors were encountered: