Use specific builder functions for encoding.

Instead of keeping internal function for decoding we can reuse much more efficient builders from the bytestring library. It increases speed and improves memory usage and allows more code sharing between the libraries. Specialization pragmas were removed, they didn't provide much benefit anyway as the code is recursive and optimizer does not inline it. However internal functions were not removed so the user of cassava can use them for semi-efficient numeric types encoding in case if there are no efficient builders available directly for those types.
haskell-hvr · Jul 5, 2020 · a2c416a · a2c416a
1 parent 2348030
commit a2c416a
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 47 deletions.
diff --git a/src/Data/Csv/Conversion.hs b/src/Data/Csv/Conversion.hs
@@ -79,6 +79,7 @@ import qualified Control.Monad.Fail as Fail
 import Data.Attoparsec.ByteString.Char8 (double)
 import qualified Data.Attoparsec.ByteString.Char8 as A8
 import qualified Data.ByteString as B
+import qualified Data.ByteString.Builder as Builder
 import qualified Data.ByteString.Char8 as B8
 import qualified Data.ByteString.Lazy as L
 #if MIN_VERSION_bytestring(0,10,4)
@@ -885,7 +886,7 @@ instance FromField Int where
 
 -- | Uses decimal encoding with optional sign.
 instance ToField Int where
-    toField = decimal
+    toField = viaBuilder . Builder.intDec
     {-# INLINE toField #-}
 
 -- | Accepts a signed decimal number. Ignores whitespace.
@@ -895,7 +896,7 @@ instance FromField Integer where
 
 -- | Uses decimal encoding with optional sign.
 instance ToField Integer where
-    toField = decimal
+    toField = viaBuilder . Builder.integerDec
     {-# INLINE toField #-}
 
 -- | Accepts a signed decimal number. Ignores whitespace.
@@ -905,7 +906,7 @@ instance FromField Int8 where
 
 -- | Uses decimal encoding with optional sign.
 instance ToField Int8 where
-    toField = decimal
+    toField = viaBuilder . Builder.int8Dec
     {-# INLINE toField #-}
 
 -- | Accepts a signed decimal number. Ignores whitespace.
@@ -915,7 +916,7 @@ instance FromField Int16 where
 
 -- | Uses decimal encoding with optional sign.
 instance ToField Int16 where
-    toField = decimal
+    toField = viaBuilder . Builder.int16Dec
     {-# INLINE toField #-}
 
 -- | Accepts a signed decimal number. Ignores whitespace.
@@ -925,7 +926,7 @@ instance FromField Int32 where
 
 -- | Uses decimal encoding with optional sign.
 instance ToField Int32 where
-    toField = decimal
+    toField = viaBuilder . Builder.int32Dec
     {-# INLINE toField #-}
 
 -- | Accepts a signed decimal number. Ignores whitespace.
@@ -935,7 +936,7 @@ instance FromField Int64 where
 
 -- | Uses decimal encoding with optional sign.
 instance ToField Int64 where
-    toField = decimal
+    toField = viaBuilder . Builder.int64Dec
     {-# INLINE toField #-}
 
 -- | Accepts an unsigned decimal number. Ignores whitespace.
@@ -945,7 +946,7 @@ instance FromField Word where
 
 -- | Uses decimal encoding.
 instance ToField Word where
-    toField = decimal
+    toField = viaBuilder . Builder.wordDec
     {-# INLINE toField #-}
 
 -- | Accepts an unsigned decimal number. Ignores whitespace.
@@ -969,7 +970,7 @@ instance FromField Word8 where
 
 -- | Uses decimal encoding.
 instance ToField Word8 where
-    toField = decimal
+    toField = viaBuilder . Builder.word8Dec
     {-# INLINE toField #-}
 
 -- | Accepts an unsigned decimal number. Ignores whitespace.
@@ -979,7 +980,7 @@ instance FromField Word16 where
 
 -- | Uses decimal encoding.
 instance ToField Word16 where
-    toField = decimal
+    toField = viaBuilder . Builder.word16Dec
     {-# INLINE toField #-}
 
 -- | Accepts an unsigned decimal number. Ignores whitespace.
@@ -989,7 +990,7 @@ instance FromField Word32 where
 
 -- | Uses decimal encoding.
 instance ToField Word32 where
-    toField = decimal
+    toField = viaBuilder . Builder.word32Dec
     {-# INLINE toField #-}
 
 -- | Accepts an unsigned decimal number. Ignores whitespace.
@@ -999,7 +1000,7 @@ instance FromField Word64 where
 
 -- | Uses decimal encoding.
 instance ToField Word64 where
-    toField = decimal
+    toField = viaBuilder . Builder.word64Dec
     {-# INLINE toField #-}
 
 instance FromField B.ByteString where

diff --git a/src/Data/Csv/Conversion/Internal.hs b/src/Data/Csv/Conversion/Internal.hs
@@ -2,6 +2,8 @@ module Data.Csv.Conversion.Internal
     ( decimal
     , scientific
     , realFloat
+    , viaBuilder
+    , formatBoundedSigned
     ) where
 
 import Data.ByteString.Builder (Builder, toLazyByteString, word8, char8,
@@ -12,7 +14,6 @@ import Data.Array.Base (unsafeAt)
 import Data.Array.IArray
 import qualified Data.ByteString as B
 import Data.Char (ord)
-import Data.Int
 import qualified Data.Monoid as Mon
 import Data.Scientific (Scientific)
 import Data.Word
@@ -29,26 +30,6 @@ decimal = toStrict . toLazyByteString . formatDecimal
 -- TODO: Add an optimized version for Integer.
 
 formatDecimal :: Integral a => a -> Builder
-{-# RULES "formatDecimal/Int" formatDecimal = formatBoundedSigned
-    :: Int -> Builder #-}
-{-# RULES "formatDecimal/Int8" formatDecimal = formatBoundedSigned
-    :: Int8 -> Builder #-}
-{-# RULES "formatDecimal/Int16" formatDecimal = formatBoundedSigned
-    :: Int16 -> Builder #-}
-{-# RULES "formatDecimal/Int32" formatDecimal = formatBoundedSigned
-    :: Int32 -> Builder #-}
-{-# RULES "formatDecimal/Int64" formatDecimal = formatBoundedSigned
-    :: Int64 -> Builder #-}
-{-# RULES "formatDecimal/Word" formatDecimal = formatPositive
-    :: Word -> Builder #-}
-{-# RULES "formatDecimal/Word8" formatDecimal = formatPositive
-    :: Word8 -> Builder #-}
-{-# RULES "formatDecimal/Word16" formatDecimal = formatPositive
-    :: Word16 -> Builder #-}
-{-# RULES "formatDecimal/Word32" formatDecimal = formatPositive
-    :: Word32 -> Builder #-}
-{-# RULES "formatDecimal/Word64" formatDecimal = formatPositive
-    :: Word64 -> Builder #-}
 {-# NOINLINE formatDecimal #-}
 formatDecimal i
     | i < 0     = minus Mon.<>
@@ -58,11 +39,6 @@ formatDecimal i
     | otherwise = formatPositive i
 
 formatBoundedSigned :: (Integral a, Bounded a) => a -> Builder
-{-# SPECIALIZE formatBoundedSigned :: Int -> Builder #-}
-{-# SPECIALIZE formatBoundedSigned :: Int8 -> Builder #-}
-{-# SPECIALIZE formatBoundedSigned :: Int16 -> Builder #-}
-{-# SPECIALIZE formatBoundedSigned :: Int32 -> Builder #-}
-{-# SPECIALIZE formatBoundedSigned :: Int64 -> Builder #-}
 formatBoundedSigned i
     | i < 0     = minus Mon.<>
                   if i == minBound
@@ -71,16 +47,6 @@ formatBoundedSigned i
     | otherwise = formatPositive i
 
 formatPositive :: Integral a => a -> Builder
-{-# SPECIALIZE formatPositive :: Int -> Builder #-}
-{-# SPECIALIZE formatPositive :: Int8 -> Builder #-}
-{-# SPECIALIZE formatPositive :: Int16 -> Builder #-}
-{-# SPECIALIZE formatPositive :: Int32 -> Builder #-}
-{-# SPECIALIZE formatPositive :: Int64 -> Builder #-}
-{-# SPECIALIZE formatPositive :: Word -> Builder #-}
-{-# SPECIALIZE formatPositive :: Word8 -> Builder #-}
-{-# SPECIALIZE formatPositive :: Word16 -> Builder #-}
-{-# SPECIALIZE formatPositive :: Word32 -> Builder #-}
-{-# SPECIALIZE formatPositive :: Word64 -> Builder #-}
 formatPositive = go
   where go n | n < 10    = digit n
              | otherwise = go (n `quot` 10) Mon.<> digit (n `rem` 10)
@@ -297,3 +263,8 @@ i2d i = fromIntegral (ord '0' + i)
 -- | Word8 list rendering
 word8s :: [Word8] -> Builder
 word8s = BP.primMapListFixed BP.word8
+
+-- | Helper function that build strict bytestring via a provided builder.
+viaBuilder :: Builder -> B.ByteString
+viaBuilder = toStrict . toLazyByteString
+{-# INLINE viaBuilder #-}