diff options
| author |  Joey Hess <joeyh@joeyh.name> | 2015-06-16 18:38:12 -0400 |
|---|---|---|
| committer | Joey Hess <joeyh@joeyh.name> | 2015-06-16 18:50:13 -0400 |
| commit | 2a4a641a3c940a652f62371f417433b47f7f3e79 (patch) | |
| tree | e610e549f25dd4cd6719fcce9dcde14a69c46bd2 /Annex/BloomFilter.hs | |
| parent | e3da28295e11972bcb14749ef294d1f39fb03efa (diff) | |
use bloom filter in second pass of sync --all --content
This is needed because when preferred content matches on files,
the second pass would otherwise want to drop all keys. Using a bloom filter
avoids this, and in the case of a false positive, a key will be left
undropped that preferred content would allow dropping. Chances of that
happening are a mere 1 in 1 million.
Diffstat (limited to 'Annex/BloomFilter.hs')
| -rw-r--r-- | Annex/BloomFilter.hs | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/Annex/BloomFilter.hs b/Annex/BloomFilter.hs index 3ac81fa58..5773a88ee 100644 --- a/Annex/BloomFilter.hs +++ b/Annex/BloomFilter.hs @@ -40,11 +40,11 @@ bloomBitsHashes = do - Once the action completes, the mutable filter is frozen - for later use. -} -genBloomFilter :: Hashable v => ((v -> Annex ()) -> Annex b) -> Annex (Bloom v) +genBloomFilter :: Hashable v => ((v -> Annex ()) -> Annex ()) -> Annex (Bloom v) genBloomFilter populate = do (numbits, numhashes) <- bloomBitsHashes bloom <- lift $ newMB (cheapHashes numhashes) numbits - _ <- populate $ \v -> lift $ insertMB bloom v + populate $ \v -> lift $ insertMB bloom v lift $ unsafeFreezeMB bloom where lift = liftIO . stToIO |