import qualified Data.ByteString       as B
import qualified Data.ByteString.Short as B
         (ShortByteString, toShort, fromShort)

import qualified Data.ByteString.Short as B.Short

The ShortByteString type

data ShortByteString Source

A compact representation of a Word8 vector.

It has a lower memory overhead than a ByteString and and does not contribute to heap fragmentation. It can be converted to or from a ByteString (at the cost of copying the string data). It supports very few other operations.

It is suitable for use as an internal representation for code that needs to keep many short strings in memory, but it should not be used as an interchange type. That is, it should not generally be used in public APIs. The ByteString type is usually more suitable for use in interfaces; it is more flexible and it supports a wide range of operations.

Memory overhead

With GHC, the memory overheads are as follows, expressed in words and in bytes (words are 4 and 8 bytes on 32 or 64bit machines respectively).

• ByteString unshared: 9 words; 36 or 72 bytes.
• ByteString shared substring: 5 words; 20 or 40 bytes.
• ShortByteString: 4 words; 16 or 32 bytes.

For the string data itself, both ShortByteString and ByteString use one byte per element, rounded up to the nearest word. For example, including the overheads, a length 10 ShortByteString would take 16 + 12 = 28 bytes on a 32bit platform and 32 + 16 = 48 bytes on a 64bit platform.

These overheads can all be reduced by 1 word (4 or 8 bytes) when the ShortByteString or ByteString is unpacked into another constructor.

For example:

data ThingId = ThingId {-# UNPACK #-} !Int
                       {-# UNPACK #-} !ShortByteString

This will take 1 + 1 + 3 words (the ThingId constructor + unpacked Int + unpacked ShortByteString), plus the words for the string data.

Heap fragmentation

With GHC, the ByteString representation uses pinned memory, meaning it cannot be moved by the GC. This is usually the right thing to do for larger strings, but for small strings using pinned memory can lead to heap fragmentation which wastes space. The ShortByteString type (and the Text type from the text package) use unpinned memory so they do not contribute to heap fragmentation. In addition, with GHC, small unpinned strings are allocated in the same way as normal heap allocations, rather than in a separate pinned area.

Conversions

toShort :: ByteString -> ShortByteString Source

O(n). Convert a ByteString into a ShortByteString.

This makes a copy, so does not retain the input string.

fromShort :: ShortByteString -> ByteString Source

O(n). Convert a ShortByteString into a ByteString.

pack :: [Word8] -> ShortByteString Source

O(n). Convert a list into a ShortByteString

unpack :: ShortByteString -> [Word8] Source

O(n). Convert a ShortByteString into a list.

Other operations

empty :: ShortByteString Source

O(1). The empty ShortByteString.

null :: ShortByteString -> Bool Source

O(1) Test whether a ShortByteString is empty.

length :: ShortByteString -> Int Source

O(1) The length of a ShortByteString.

index :: ShortByteString -> Int -> Word8 Source

O(1) ShortByteString index (subscript) operator, starting from 0.