Unnecessary use of strict bytestrings in Data.Acid.Log

[adamgundry]

I happened across a couple of places where strict bytestrings are used seemingly unnecessarily (now):

acid-state/src/Data/Acid/Log.hs

Lines 186 to 189 in 194f0ab

	-- XXX: Strict bytestrings are used due to a performance bug in
	-- cereal-0.3.5.2 and binary-0.7.1.0. The code should revert back
	-- to lazy bytestrings once the bug has been fixed.
	archive <- liftM Lazy.fromChunks $ mapM (Strict.readFile . snd) relevant

acid-state/src/Data/Acid/Log.hs

Lines 320 to 323 in 194f0ab

	-- XXX: Strict bytestrings are used due to a performance bug in
	-- cereal-0.3.5.2 and binary-0.7.1.0. The code should revert back
	-- to lazy bytestrings once the bug has been fixed.
	archive <- fmap Lazy.fromStrict $ Strict.readFile logFile

This is somewhat similar to the issue from #125:

acid-state/src/Data/Acid/Log.hs

Line 342 in 194f0ab

encoded = Lazy.fromChunks [ Strict.copy $ Lazy.toStrict $

[ddssff]

Maybe a version #ifdef would be the right thing here? Edit: actually the cabal file requires cereal >= 0.4.1, so nvm about that ifdef.