Optimisations to time it takes git-annex to walk working tree and find files to work on. Sped up by around 18%.

key2file and file2key were top cost centers according to profiling.
The repeated use of replace was not efficient. This new approach is quite a
lot more efficient.

This commit was sponsored by Denis Dzyubenko on Patreon.
This commit is contained in:
Joey Hess 2016-09-26 16:47:59 -04:00
parent a8734698eb
commit 8794dcf27b
No known key found for this signature in database
GPG key ID: C910D9222512E3C7
2 changed files with 22 additions and 5 deletions

View file

@ -452,15 +452,25 @@ preSanitizeKeyName = concatMap escape
- can cause existing objects to get lost.
-}
keyFile :: Key -> FilePath
keyFile key = replace "/" "%" $ replace ":" "&c" $
replace "%" "&s" $ replace "&" "&a" $ key2file key
keyFile = concatMap esc . key2file
where
esc '&' = "&a"
esc '%' = "&s"
esc ':' = "&c"
esc '/' = "%"
esc c = [c]
{- Reverses keyFile, converting a filename fragment (ie, the basename of
- the symlink target) into a key. -}
fileKey :: FilePath -> Maybe Key
fileKey file = file2key $
replace "&a" "&" $ replace "&s" "%" $
replace "&c" ":" $ replace "%" "/" file
fileKey = file2key . unesc []
where
unesc r [] = reverse r
unesc r ('%':cs) = unesc ('/':r) cs
unesc r ('&':'c':cs) = unesc (':':r) cs
unesc r ('&':'s':cs) = unesc ('%':r) cs
unesc r ('&':'a':cs) = unesc ('&':r) cs
unesc r (c:cs) = unesc (c:r) cs
{- for quickcheck -}
prop_isomorphic_fileKey :: String -> Bool

View file

@ -1,3 +1,10 @@
git-annex (6.20160924) UNRELEASED; urgency=medium
* Optimisations to time it takes git-annex to walk working tree and find
files to work on. Sped up by around 18%.
-- Joey Hess <id@joeyh.name> Mon, 26 Sep 2016 16:46:19 -0400
git-annex (6.20160923) unstable; urgency=medium
* Rate limit console progress display updates to 10 per second.