hunk ./series 2 -todo-file-name.diff hunk ./todo-file-name.diff 1 ---- binary/debian/docs.orig 2010-09-28 12:02:28.000000000 -0700 -+++ binary/debian/docs 2010-10-09 07:09:15.000000000 -0700 -@@ -1,3 +1,3 @@ - README --todo -+TODO - index.html rmfile ./todo-file-name.diff hunk ./ghc7.diff 63 -diff -ruN binary-0.5.0.2/todo binary-0.5.0.3/todo ---- binary-0.5.0.2/todo 2009-09-16 10:50:42.000000000 -0700 -+++ binary-0.5.0.3/todo 1969-12-31 16:00:00.000000000 -0800 -@@ -1,28 +0,0 @@ --layer handling: -- -- bit packing -- state parameters -- string pools -- -- reading structures from the end of a stream, seek/tell behaviour -- --seek based protocols are too hard. -- hGetContents/ interleaving. -- --user requests: -- -- get remaining bytestring after a runGet -- -- some kind of lookahead, or restoring parsing state, or something with -- equal functionality. make it another layer on top? -- -- getLazyByteString takes an Int, which in Haskell98 is only guarantied to -- be 29 bits, ie. 512 mb. -- maybe we should have a readN64 for allowing reading of larger stuff? -- (which could be implemented with readN on 64bit machines) -- reference: bringerts tar archive decoder would be limitid to 0.5GB -- files, alt. 2GB in GHC -- --SYB-deriving -- --investigate the UArray instance, it does not seem to compile in GHC 6.4 -diff -ruN binary-0.5.0.2/TODO binary-0.5.0.3/TODO ---- binary-0.5.0.2/TODO 1969-12-31 16:00:00.000000000 -0800 -+++ binary-0.5.0.3/TODO 2010-09-25 10:10:10.000000000 -0700 -@@ -0,0 +1,28 @@ -+layer handling: -+ -+ bit packing -+ state parameters -+ string pools -+ -+ reading structures from the end of a stream, seek/tell behaviour -+ -+seek based protocols are too hard. -+ hGetContents/ interleaving. -+ -+user requests: -+ -+ get remaining bytestring after a runGet -+ -+ some kind of lookahead, or restoring parsing state, or something with -+ equal functionality. make it another layer on top? -+ -+ getLazyByteString takes an Int, which in Haskell98 is only guarantied to -+ be 29 bits, ie. 512 mb. -+ maybe we should have a readN64 for allowing reading of larger stuff? -+ (which could be implemented with readN on 64bit machines) -+ reference: bringerts tar archive decoder would be limitid to 0.5GB -+ files, alt. 2GB in GHC -+ -+SYB-deriving -+ -+investigate the UArray instance, it does not seem to compile in GHC 6.4