Can't seem to get the addressing right.

[GuruSR]

Buuut, Mov Bl, [Edi +2]  does not

I've tested it: mov bl, [edi+2] works fine with ML 6.15, 8.0, 10.0 at least, plus JWasm, HJWasm and AsmC. None of them requires the byte ptr.

Alright, it didn't, but it is now.  Only thing I can think of is the line had a ctrl code stuck in there somewhere somehow that was causing it to not compile, kept giving me a syntax error on that line, but after I re-wrote the entry from the Mov onward, using the Byte Ptr [Edi+2] it worked, so I'm thinking a ctrl code got in there somehow.

GuruSR.

[rrr314159]

Only thing I can think of is the line had a ctrl code stuck in there somewhere somehow that was causing it to not compile, kept giving me a syntax error on that line, but after I re-wrote the entry from the Mov onward, using the Byte Ptr [Edi+2] it worked, so I'm thinking a ctrl code got in there somehow.

Yes, that can happen. One way it can occur, copying and pasting from a rich-text file or similar can introduce an invisible control code. If you'd known for certain the line was supposed to work you would have quickly realized it. But when you're not certain, something like that can easily put you on the wrong track

[jj2007]

copying and pasting from a rich-text file or similar can introduce an invisible control code.

Yes, that can happen. For this scenario, the developers of assemblers have introduced "error messages" some time ago:

Code Select Expand

  mov bl, [edi+1] ; OK
  mov bl, [edi+1] ; [edi+hiddentext+1]: Error A2102: Symbol not defined : hiddentext
  and [edi+2], 0 ; ML: error A2070:invalid instruction operands
  and [edi+2], 0 ; HJWasm: Warning A4073: Size not specified, assuming: BYTE


[GuruSR]

Well, the error it was giving me was Syntax, which made no sense, which is why I was wondering what I didn't have right (syntax basically means the format of how it works and when you get an error, means you didn't format it correctly).

Code Select Expand


Mov Edi, DWord Ptr [SecOne] ; Ready output.
Mov Ecx, 00H ; Length of Section.
Cmp Edi, Ecx
Je NoOne


And another weirdness, it crashed, I debugged and I looked at the Edi and Ecx, both had 0's in them (Edi went to 1 after the Stosb, which caused the crash), but that Je never went to NoOne, passed it by.  ZF should have been 1 in that situation and the two should have matched and Je should have bypassed the writing of section 1 because the storage for that section was invalid.

GuruSR.

[jj2007]

Code Select Expand


Mov Edi, DWord Ptr [SecOne] ; Ready output.
Mov Ecx, 00H ; Length of Section.
Cmp Edi, Ecx
Je NoOne


And another weirdness, it crashed, I debugged and I looked at the Edi and Ecx, both had 0's in them (Edi went to 1 after the Stosb, which caused the crash)

Nope, edi can't go to one using a stosb. Check it in the debugger:

Code Select Expand

  xor edi, edi   ; set to zero
  int 3
  stosb

Btw:
   Mov Edi, DWord Ptr [SecOne]   ; Ready output.
   Mov Edi, SecOne         ; the assembler knows that edi is dword sized, and that SecOne is a memory location
   Mov Ecx, 00H         ; Length of Section.
   xor ecx, ecx         ; same effect, 2 bytes instead of 5

Similar:
  mov SecOne, 0  ; 10 bytes
  and SecOne, 0  ; 7 bytes, same effect

[GuruSR]

Nope, edi can't go to one using a stosb. Check it in the debugger:

Code Select Expand

  xor edi, edi   ; set to zero
  int 3
  stosb

Btw:
   Mov Edi, DWord Ptr [SecOne]   ; Ready output.
   Mov Edi, SecOne         ; the assembler knows that edi is dword sized, and that SecOne is a memory location
   Mov Ecx, 00H         ; Length of Section.
   xor ecx, ecx         ; same effect, 2 bytes instead of 5

Similar:
  mov SecOne, 0  ; 10 bytes
  and SecOne, 0  ; 7 bytes, same effect

I tracked that 1 down to a null I was using, apparently it was... 1!  So that include had a null set to 1, so I fixed it, solved the issue.

SecOne is the memory pointer to a buffer space, Ecx checks against the length of that (pre-supplied value of the length), just used to seeing the Mov dest, Dword Ptr [source] and figured since most were doing that, that it meant it was the best way to accomplish it.

I know about the xor trick, I use it elsewhere, but where I was doing that, I wanted to visually see what it was.  I know the xor is faster and smaller, I actually haven't streamlined the code much, wasn't sure if there was anything out there to offer such.  I remember doing it manually on my CRC library.

GuruSR.