previously we just used a Go builtin map
this was not safe for concurrent read/write and upon upgrading
to Go 1.6 we were notified of the problem
fixes#349
our implementation uses: golang.org/x/net/context
New method SearchInContext() allows the user to run a search
in the provided context. If that context is cancelled or
exceeds its deadline Bleve will attempt to stop and return
as soon as possible. This is a *best effort* attempt at this
time and may *not* be in a timely manner. If the caller must
return very near the timeout, the call should also be wrapped
in a goroutine.
The IndexAlias implementation is affected in a slightly more
complex way. In order to return partial results when a timeout
occurs on some indexes, the timeout is strictly enforced, and
at the moment this does introduce an additional goroutine.
The Bleve implementation honoring the context is currently
very course-grained. Specifically we check the Done() channel
between each DocumentMatch produced during the search. In the
future we will propogate the context deeper into the internals
of Bleve, and this will allow finer-grained timeout behavior.
Is it fine to add codebeat badge to README?
codebeat is automated code review tool for Swift, Ruby & Go that helps get instant feedback on code quality.
"Quick wins" suggested by codebeat could be a nice candidate for a pull request and help other developers become contributors.
FYI. To be fully open and honest. I'm co-founder of codebeat.
previously we incorrectly identified matching buckets by
comparing string pointers. this worked in the unit test
but not in real applications since the strings result from
date parsing inside the facet collector, and are therefore
different pointers
having boost of 0 led to invalid scores of NaN
added integration test for wildcard query
added ability to run single integration test at a time
added assertion that scoare is not NaN/+Inf/-Inf
the Status section can report on the number of total/fail/success
indexes when querying across multiple indexes through IndexAlias
Further, searching an IndexAlias will now return partial results,
the burden is on the caller to check the number of failed
indexes and decide how to handle this situation.
This is somewhat unlikely, but if a term is (incredibly) popular, its
uvarint count value representation might go beyond 8 bytes.
Some KVStore implementations (like forestdb) provide a BatchEx cgo
optimization that depends on proper preallocated counting, so this
change provides a proper worst-case estimate based on the max-unvarint
of 10 bytes instead of the previously incorrect 8 bytes.
Performance optimization. Before this change, by using Merge()
instead of AllocMerge(), moss's internal batch buf's would be
wastefully, dramatically grown during append()'s to a mis-sized buf.
in strict parsing mode, unrecognized keys in the mapping will
cause errors. this is partial at the moment and does not
include the custom analysis components.
it would appear that a document lookup for an id fails
but that is a document id that was returned as a search hit
since we're using a stable snapshot, this should not happen
