struct (does not compile ?)

Index > Macroinstructions > struct (does not compile ?)

Author

Thread

ouadji

Joined: 24 Dec 2008
Posts: 1081
Location: Belgium

ouadji 09 Nov 2010, 21:12

"A" does compile, "B" does not compile

why ?

Code:

A) ;does compile
include 'win32a.inc'

struct BB
 union
  v1 dd ?
  v2 dd ?
 ends
ends

struct AA
v3 dd ?
v4 BB
ends

Code:

B) ;does NOT compile <------ ??
include 'win32a.inc'

struct BB
 union
  v1 dd ?
  v2 dd ?
 ends
ends


struct AA
v3 dd ?
BB    ;<----- B) = A) - "v4" .... here "BB" is alone.
ends       ;there are no other differences between A and B

that said, both below compile

C) D) == A) B) - union

Code:

C) ;does compile
include 'win32a.inc'

struct BB
  v1 dd ?
  v2 dd ?
ends

struct AA
v3 dd ?
v4 BB
ends

Code:

D) ;does compile
include 'win32a.inc'

struct BB
  v1 dd ?
  v2 dd ?
ends

struct AA
v3 dd ?
BB  ;<--- B) is alone, but here it does compile
ends

_________________
I am not young enough to know everything (Oscar Wilde)-

09 Nov 2010, 21:12

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8358
Location: Kraków, Poland

Tomasz Grysztar 09 Nov 2010, 23:09

Yes, this macro was unfinished in this aspect and I forgot about it. Here's the corrected "define@struct" macro:

Code:

macro define@struct name,[field,type,def]
 { common
    local list
    list equ
   forward
    if ~ field eq .
     name#field type def
     sizeof.#name#field = $ - name#field
    else
     label name#.#type
     rb sizeof.#type
    end if
    local value
    match any, list \{ list equ list, \}
    list equ list <value>
   common
    sizeof.#name = $
    restruc name
    match values, list \{
    struc name value \\{
    match any, fields@struct \\\{ fields@struct equ fields@struct,.,name,<values> \\\}
    match , fields@struct \\\{ label .
   forward
     match , value \\\\{ field type def \\\\}
     match any, value \\\\{ field type value
                            if ~ field eq .
                             rb sizeof.#name#field - ($-field)
                            end if \\\\}
   common \\\} \\}
    macro name value \\{
    match any, fields@struct \\\{ \\\local anonymous
                                  fields@struct equ fields@struct,anonymous,name,<values> \\\}
    match , fields@struct \\\{ label .
   forward
     match , value \\\\{ type def \\\\}
     match any, value \\\\{ \\\\local ..field
                           ..field = $
                           type value
                           if ~ field eq .
                            rb sizeof.#name#field - ($-..field)
                           end if \\\\}
   common \\\} \\} \} }

There may be some more fixing need to make the standalone "BB" work also outside the "struct", though.

09 Nov 2010, 23:09

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8358
Location: Kraków, Poland

Tomasz Grysztar 09 Nov 2010, 23:28

OK, I'm uploading the package with finished (I hope) corrections.

09 Nov 2010, 23:28

ouadji

Joined: 24 Dec 2008
Posts: 1081
Location: Belgium

ouadji 10 Nov 2010, 01:02

flat assembler 1.69.25 for Windows
size: 891 kilobytes
last update: 09 Nov 2010

now, with this last update,
this below does not compile ???

(with the previous version of "struct", it compiled without problem.)

Code:

include 'win32a.inc'

struct BB
  v1 dd ?
  v2 dd ?
ends

mov eax,dword[toto.v2.x]

toto BB

.x :

_________________
I am not young enough to know everything (Oscar Wilde)-

10 Nov 2010, 01:02

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20445
Location: In your JS exploiting you and your system

revolution 10 Nov 2010, 01:07

Code:

mov eax,dword[toto.x]

10 Nov 2010, 01:07

ouadji

Joined: 24 Dec 2008
Posts: 1081
Location: Belgium

ouadji 10 Nov 2010, 01:11

no, with the prévious version of "struct"
mov eax,dword[toto.v2.x] compiled without problem

I develop a full highlighting of struct macro ! (for next Wink)
I absolutely need a reference.

_________________
I am not young enough to know everything (Oscar Wilde)-

Last edited by ouadji on 10 Nov 2010, 01:25; edited 2 times in total

10 Nov 2010, 01:11

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20445
Location: In your JS exploiting you and your system

revolution 10 Nov 2010, 01:20

ouadji wrote:

no, with the prévious version of "struct"

mov eax,dword[toto.v2.x] compiled without problem

But this is a new version. It has different behaviour.

10 Nov 2010, 01:20

ouadji

Joined: 24 Dec 2008
Posts: 1081
Location: Belgium

ouadji 10 Nov 2010, 01:31

Tomasz, why this is no longer allowed?
This new version greatly limits the possibilities of "struct".
Is this really your wish? Confused

Code:

include 'win32a.inc'

struct BB
  v1 dd ?
  v2 dd ?
ends

mov eax,dword[toto.v2.x]

toto BB

.x :

_________________
I am not young enough to know everything (Oscar Wilde)-

10 Nov 2010, 01:31

baldr

Joined: 19 Mar 2008
Posts: 1651

baldr 10 Nov 2010, 23:57

Tomasz Grysztar,

Probably label should explicitly declare whether it wants to have following local labels to be attached to it. Something like

Code:

hello.:  …
goodbye: …
.world:  jmp     hello.world

10 Nov 2010, 23:57

ouadji

Joined: 24 Dec 2008
Posts: 1081
Location: Belgium

ouadji 11 Nov 2010, 01:35

it is true that your idea is interesting baldr, but I do not see the interest on a practical level. I do not think it would bring much more possibilities. That said, once again, it's a original idea.

_________________
I am not young enough to know everything (Oscar Wilde)-

11 Nov 2010, 01:35

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8358
Location: Kraków, Poland

Tomasz Grysztar 11 Nov 2010, 07:45

baldr wrote:

Probably label should explicitly declare whether it wants to have following local labels to be attached to it.

If I was to break the long-lived NASM compatibility on this one, I would probably just remove this feature completely.

11 Nov 2010, 07:45

bitRAKE

Joined: 21 Jul 2003
Posts: 4072
Location: vpcmpistri

bitRAKE 11 Nov 2010, 08:21

Tomasz, what method do you use to avoid name collision under SSSO principal? Please, don't remove dot labels - not all labels need some_descriptive_long_text.

11 Nov 2010, 08:21

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8358
Location: Kraków, Poland

Tomasz Grysztar 11 Nov 2010, 09:35

Did I say I was going to remove them? Wink

11 Nov 2010, 09:35

ouadji

Joined: 24 Dec 2008
Posts: 1081
Location: Belgium

ouadji 11 Nov 2010, 09:55

200% agree with baldr.
dot label are very useful.
Without them, fasm lose much of its interest.

_________________
I am not young enough to know everything (Oscar Wilde)-

11 Nov 2010, 09:55

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8358
Location: Kraków, Poland

Tomasz Grysztar 11 Nov 2010, 10:18

If it was not for NASM compatibility I would go for something more fitting the overall fasm's design, like nested local namespaces closed with "end" directive. But being compatible with NASM at least in this aspect doesn't usually hurt - though you have to be careful with macro design to take it into consideration. But this is all academic discussion - this feature is to well-established in fasm's 1.x architecture to change it now.

11 Nov 2010, 10:18

ouadji

Joined: 24 Dec 2008
Posts: 1081
Location: Belgium

ouadji 11 Nov 2010, 11:23

Quote:

this feature is to well-established in fasm's 1.x architecture
to change it now.

very good conclusion Tomasz !
don't touch anything, it's perfect like that
fasm does not have to be like this or like that ...
fasm can be simply himself. Razz

_________________
I am not young enough to know everything (Oscar Wilde)-

11 Nov 2010, 11:23

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum