nasm/Wishlist

NASM Wishlist
=============

Numbers on right hand side are version numbers that it would be nice to
have this done by. ? means I haven't looked at it yet.

- Create a binary RDF tools distribution. Should probably be distributed 0.98
  seperately.

- Check misc/ide.cfg into RCS as Watcom IDE enhancement thingy.		0.98
  (nop@dlc.fi)

- Package the Linux Assembler HOWTO.					0.98

- AMD 3dNow extensions need documenting.				0.98

- prototypes of lrotate don't match in test/*. Fix.			0.98

- Build djgpp binaries for 0.98 onwards. Look into PMODE/W as a stub	0.98
  - it might be a lot better than CWSDPMI. It's in PMW133.ZIP.

- Fix `%error' giving error messages twice.				0.99
  Not especially important, as changes planned for 1.1x below will make
  the preprocessor be only called once.

- Sort out problems with OBJ:						0.99
  * TLINK32 doesn't seem to like SEGDEF32 et al. So for that, we
    should avoid xxx32 records wherever we can.
  * However, didn't we change _to_ using xxx32 at some stage? Try
    to remember why and when.
  * Apparently Delphi's linker has trouble with two or more
    globals being defined inside a PUBDEF32. Don't even know if it
    _can_ cope with a PUBDEF16.
  * Might need extra flags. *sigh*

- Symbol table output may possibly be useful.				0.99
  Ken Martwick (kenm@efn.org) wants the following format:
	labelname	type	offset(hex)	repetition count
  Possibly include xref addresses after repetition count?

- There are various other bugs in outelf.c that make certain kinds	0.99
  of relocation not work. See zbrown.asm. Looks like we may have to do
  a major rewrite of parts of it. Compare some NASM code output with
  equivalent GAS code output. Look at the ELF spec. Generally fix things.

- NASM is currently using a kludge in ELF that involves defining	0.99
  a symbol at a zero absolute offset. This isn't needed, as the
  documented solution to the problem that this solves is to use
  SHN_UNDEF.

- Debug information, in all formats it can be usefully done in.		0.99
  * including line-number record support.
  * "George C. Lindauer" <gclind01@starbase.spd.louisville.edu>
    wants to have some say in how this goes through.
  * Andrew Crabtree <andrewc@rosemail.rose.hp.com> wants to help out.

- Think about a line-continuation character.				0.99

- Consider allowing declaration of two labels on the same line,
  syntax 'label1[:] label2[:] ... instruction'. Need to investigate
  feasibility.								0.99

- Quoting of quotes by doubling them, in string and char constants.	0.99

- Two-operand syntax for SEGMENT/SECTION macro to avoid warnings	0.99
  of ignored section parameters on reissue of __SECT__.
  Or maybe skip the warning if the given parameters are identical to
  what was actually stored. Investigate.				

- Apparently we are not missing a PSRAQ instruction, because it
  doesn't exist.  Check that it doesn't exist as an undocumented
  instruction, or something stupid like that.				0.99

- Any assembled form starting 0x80 can also start 0x82. ndisasm		1.00
  should know this. New special code in instruction encodings,
  probably.

- Pointing an EQU at an external symbol now generates an error. There	1.05
  may be a better way of handling this; we should look into it.
  Ideally, the label mechanism should be changed to cope with one
  label being declared relative to another - that may work, but could be
  a pain to implement (or is it? it may be easy enough that you just
  need to declare a new offset in the same segment...) This should be done
  before v1.0 is released. There is a comment regarding this in labels.c,
  towards the end of the file, which discusses ways of fixing this.

- nested %rep used to cause a panic. Now a more informative error	1.10
  message is produced. This problem whould be fixed before v1.0.
  See comment in switch() statement block for PP_REP in do_directive()
  in preproc.c (line 1585, or thereabouts)

- Contribution: zgraeme.tar contains improved hash table routines	?
  contributed by Graeme Defty <graeme@HK.Super.NET> for use in the
  label manager.

- Contribution: zsyntax.zip contains a syntax-highlighting mode for	?
  NASM, for use with the Aurora text editor (??).

- Contribution: zvim.zip contains a syntax-highlighting mode for	?
  NASM, for use with vim.

- Contribution: zkendal1.zip and zkendal2.zip contain Kendall		?
  Bennett's (<KendallB@scitechsoft.com>) alternative syntax stuff,
  providing an alternative syntax mode for NASM which allows a macro
  set to be written that allows the same source files to be
  assembled with NASM and TASM.

- Add the UD2 instruction.						?

- Add the four instructions documented in 24368901.pdf (Intel's own	?
  document).

- Some means of avoiding MOV memoffs,EAX which apparently the		1.10?
  Pentium pairing detector thinks modifies EAX. Similar means of
  choosing instruction encodings where necessary.

- The example of ..@ makes it clear that a ..@ label isn't just		?
  local, but doesn't make it clear that it isn't just global either.

- hpa wants an evaluator operator for ceil(log2(x)).			?

- Extra reloc types in ELF: R_386_16 type 20, PC16 is 21, 8 is 22, PC8 is 23.
  Add support for the 16s at least.					?


- Lazy section creation or selective section output, in COFF/win32	?
  at least and probably other formats: don't bother to emit a section
  if it contains no data. Particularly the default auto-created
  section. We believe zero-length sections crash at least WLINK (in
  win32).

- Make the flags field in `struct itemplate' in insns.h a long		?
  instead of an int.

- Implement %ifref to check whether a single-line macro has ever been	?
  expanded since (last re) definition. Or maybe not. We'll see.

- add pointer to \k{insLEAVE} and \k{insENTER} in chapters about	?
  mixed-language programming.

- Some equivalent to TASM's GLOBAL directive, ie something which	?
  defines a symbol as external if it doesn't end up being defined
  but defines it as public if it does end up being defined.

- Documentation doesn't explain about C++ name mangling.		?

- see if BITS can be made to do anything sensible in obj (eg set the	?
  default new-segment property to Use32).

- OBJ: coalesce consecutive offset and segment fixups for the same	?
  location into full-32bit-pointer fixups. This is apparently
  necessary because some twazzock in the PowerBASIC development
  team didn't deign to support the OMF spec the way the rest of the
  world sees it.

- Allow % to be separated from the rest of a preproc directive, for	?
  alternative directive indentation styles.

- __DATE__, __TIME__, and text variants of __NASM_MAJOR__ and		?
  __NASM_MINOR__.

- Warn on TIMES combined with multi-line macros. TIMES gets applied	1.00
  to first line only - should bring to users' attention.

- Re-work the evaluator, again, with a per-object-format fixup		1.10
  routine, so as to be able to cope with section offsets "really"
  being pure numbers; should be able to allow at _least_ the two
  common idioms
     TIMES 510-$ DB 0            ; bootsector
     MOV AX,(PROG_END-100H)/16   ; .COM TSR
  Would need to call the fixup throughout the evaluator, and the
  fixup would have to be allowed to return UNKNOWN on pass one if it
  had to. (_Always_ returning UNKNOWN on pass one, though a lovely
  clean design, breaks the first of the above examples.)

- Preprocessor identifier concatenation?				1.10

- Arbitrary section names in `bin'.					?
  Is this necessary? Is it even desirable?

- Ability to read from a pipe. Obviously not useful under dos, so	1.10
  memory problems with storing entire input file aren't a problem
  either.

  Related topic: file caching under DOS/32 bit...			1.10?
  maybe even implement discardable buffers that get thrown away
  when we get a NULL returned from malloc(). Only really useful under
  DOS. Think about it.

  Another related topic: possibly spool out the pre-processed		1.10?
  stuff to a file, to avoid having to re-process it. Possible problems
  with preprocessor values not known on pass 1? Have a look...

  Or maybe we can spool out a pre-parsed version...?			1.10
  Need to investigate feasibility. Does the results from the parser
  change from pass 1 to pass 2? Would it be feasible to alter it so that
  the parser returns an invariant result, and this is then processed
  afterwards to resolve label references, etc?

- Subsection support?							?

- A good ALIGN mechanism, similar to GAS's. GAS pads out space by	1.10?
  means of the following (32-bit) instructions:
          8DB42600000000    lea esi,[esi+0x0]
          8DB600000000      lea esi,[esi+0x0]
          8D742600          lea esi,[esi+0x0]
          8D7600            lea esi,[esi+0x0]
          8D36              lea esi,[esi]
          90                nop
  It uses up to two of these instructions to do up to 14-byte pads;
  when more than 14 bytes are needed, it issues a (short) jump to
  the end of the padded section and then NOPs the rest. Come up with
  a similar scheme for 16 bit mode, and also come up with a way to
  use it - internal to the assembler, so that programs using ALIGN
  don't knock over preprocess-only mode.
    Also re-work the macro form so that when given one argument in a
  code section it calls this feature.

- Possibly a means whereby FP constants can be specified as		?
  immediate operands to non-FP instructions.
  * Possible syntax: MOV EAX,FLOAT 1.2 to get a single-precision FP
    constant. Then maybe MOV EAX,HI_FLOAT 1.2 and MOV EAX,LO_FLOAT
    1.2 to get the two halves of a double-precision one. Best to
    ignore extended-precision in case it bites.
  * Alternatively, maybe MOV EAX,FLOAT(4,0-4,1.2) to get bytes 0-4
    (ie 0-3) of a 4-byte constant. Then HI_FLOAT is FLOAT(8,4-8,x)
    and LO_FLOAT is FLOAT(8,0-4,x). But this version allows two-byte
    chunks, one-byte chunks, even stranger chunks, and pieces of
    ten-byte reals to be bandied around as well.

- A UNION macro might be quite cool, now that ABSOLUTE is sane		?
  enough to be able to handle it.

- An equivalent to gcc's ## stringify operator, plus string		?
  concatenation, somehow implemented without undue ugliness, so as
  to be able to do `%include "/my/path/%1"' in a macro, or something
  similar...

- Actually _do_ something with the processor, privileged and		1.10
  undocumented flags in the instruction table. When this happens,
  consider allowing PMULHRW to map to either of the Cyrix or AMD
  versions?

- Maybe NEC V20/V30 instructions?					?

- Yet more object formats.
  * Possibly direct support for .EXE files?				1.10

- Symbol map in binary format. Format-specific options...		1.10?

- REDESIGN: Think about EQU dependency, and about start-point		1.20?
  specification in OBJ. Possibly re-think directive support.

- Think about a wrapper program like gcc? Possibly invent a _patch_	2.00?
  for gcc so that it can take .asm files on the command line?

- If a wrapper happens, think about adding an option to cause the	?
  resulting executable file to be executed immediately, thus
  allowing NASM source files to have #!... (probably silly)

- Multi-platform support? If so: definitely Alpha; possibly Java	?
  byte code; probably ARM/StrongARM; maybe Sparc; maybe Mips; maybe
  Vax. Perhaps Z80 and 6502, just for a laugh?

- Consider a 'verbose' option that prints information about the resulting ?
  object file onto stdout.