Stop using allocation trick in fastpath functions (OCaml 4.02).

[ocaml-bitstring.git] / TODO

diff --git a/TODO b/TODO
index 74551a6..ff2a556 100644 (file)
--- a/TODO
+++ b/TODO
@@ -9,7 +9,7 @@ Major to-do items.
 
 (4) Provide UInt32 and UInt64 types.
 
-(5) Allow for specific offsets and alignment.  Something like this:
+(5) DONE - Allow for specific offsets and alignment.  Something like this:
 
     { start : 16;
       another : 16 : offset(256);   (* location would be 256 bits from start *)
@@ -33,7 +33,7 @@ Major to-do items.
     Make the error locations fine-grained, particularly so they point to
     individual fields, not the whole match.
 
-(10) Cross-module, persistent, named patterns, see:
+(10) DONE - Cross-module, persistent, named patterns, see:
   http://caml.inria.fr/pub/ml-archives/caml-list/2008/04/25992c9c9fa999fe1d35d961dd9917a2.en.html
 
 (11) DONE -
@@ -49,11 +49,88 @@ Major to-do items.
 
 (12) DONE - More constant field lengths.
 
-(13) Implement native endian functions.
+(13) PARTLY DONE - Implement native endian functions.
 
-(14) A proper test suite.
+(14) PARTLY DONE - A proper test suite.
 
-(15) More examples:
+(15) DONE - More examples:
 
     ELF binaries
     GIF images
+
+(16) We now know the offset of the current field relative to the
+     whole match.  This may allow more efficient aligned versions
+     of functions to be called (at compile time).  However do note
+     that the offset in the bitstring is usually not known.
+
+(17) PARTLY DONE - Fix the META file.  Current one is very broken.
+
+(18) DONE - check() qualifier:
+
+     { field : 16 : check (field > 100) }
+
+     The check(expr) qualifier will abort the rest of the match if the
+     expression is false.
+
+(19) DONE - bind() qualifier:
+
+     { field : 16 : bind (field * 3) }
+       ^pattern           ^new value
+
+     The bind(expr) qualifier binds the pattern to a new value,
+     equivalent to doing:
+
+     let field = field * 3 in
+     (* remainder of match *)
+
+     There is a question of whether bind() should run before or
+     after when() [best is probably when() first, then bind()].
+
+(20) DONE - save_offset_to() qualifier:
+
+     { field : 16 : save_offset_to (offset), bind (field - offset) }
+
+     or:
+
+     { field : 16 : save_offset_to (field_offset) } ->
+       printf "the offset of field (%d) is %d\n" field field_offset
+
+     save_offset_to(patt) binds the current match offset to
+     the variable, for later use within bind(), when() or
+     any later parts of the match.
+
+(21) derive() so you can add your own variable decls:
+
+     { field : 32 : derive (field_as_int, Int32.to_int field) }
+
+     This would add a let derivation, equivalent to:
+
+     let field_as_int = Int32.to_int field
+
+     allowing you to use both the original field and field_as_int
+     as variables.
+
+     Note you can do this clumsily using bind():
+
+     { field : 32 : bind (field, Int32.to_int field) }
+
+     which redefines 'field' as a pair of (old value, derived value).
+
+(22) Allow constant 0 to be used in constructors to mean a zero-length
+     bitstring of the right length, eg:
+
+     BITSTRING { 0 : 32*8 : bitstring }
+
+     which would be equivalent to:
+
+     BITSTRING { zeroes_bitstring (32*8) : 32*8 : bitstring }
+
+(23) Add predicate Bitstring.is_zero_bitstring : bitstring -> bool
+
+(24) Add a function to turn bitstrings into printable strings.
+
+(25) Folding over bitstrings.  A narrow usage for this is to generate
+     checksums and hashes, where usually you want to fold over every
+     8/16/32-bit word.  So a set of functions which just enabled this
+     would be useful.  (However you still need le/be/ne variations so
+     it involves at least 7 functions).