Interesting idea for macro repetition

Index > Macroinstructions > Interesting idea for macro repetition

Goto page Previous 1, 2, 3 Next

Author

Thread

Ancient One

Joined: 28 Feb 2005
Posts: 55

Ancient One 05 Mar 2005, 06:43

i suggest we use breed like this :

Code:

breed [j, i name] line

which means :
repeat line j times, starting the counter from i; name represents the counter count (any occurence of name in line will be replaced by the current count number). i or/and name can always be omitted so we can always write the simple breed [j] line which means repeat line j times, or breed[j counter] line. if i is not specified, its value defaulted to 1 (or 0?). only j should always be specified here (even if it's 0).

example :

Code:

breed [0] include "debug.mac"
;repeat line include "debug.mac" 0 times.
;j=0, i is assumed to be 0, no counter name.

Code:

macro foo x {
    display `x,13,10
}

breed [3, 10 %c] foo %c
;foo 10
;foo 11
;foo 12

;iterative use of breed

breed [5 %c1] breed [%c1 %c] foo %c

;firstly :
;breed [1 %c] foo %c
;breed [2 %c] foo %c
;breed [3 %c] foo %c
;breed [4 %c] foo %c
;breed [5 %c] foo %c

;then :
;foo 1
;foo 1
;foo 2
;foo 1
;foo 2
;foo 3
;foo 1
;foo 2
;foo 3
;foo 4
;foo 1
;foo 2
;foo 3
;foo 4
;foo 5

and maybe we can make the pp to accepts negative value of "j" so we can count backwards. e.g

Code:

breed [-3,3 %c] foo %c
;foo 3
;foo 2
;foo 1

we can set the min/max (internal) value to certain value, so the increment/decrement stop when reaching this limit. e.g, if we set i to default at 1 and minimum value of i is also 1. then

Code:

breed [-3 %c] foo %c

will become

Code:

foo 1
foo 1
foo 1

05 Mar 2005, 06:43

Tomasz Grysztar

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

Tomasz Grysztar 13 Mar 2005, 12:16

Good idea, though I think I would still prefer something like:

Code:

breed [name,i] j line

with the brackets-surrounded parameters being optional (i - first value, j - count of duplicates). Possible extended form:

Code:

breed [name,i,step] j line

where step would be the value added to i after each duplicate.

13 Mar 2005, 12:16

vid
Verbosity in development

Joined: 05 Sep 2003
Posts: 7105
Location: Slovakia

vid 13 Mar 2005, 22:54

looks good

13 Mar 2005, 22:54

MCD

Joined: 21 Aug 2004
Posts: 602
Location: Germany

MCD 14 Mar 2005, 19:10

looks good so far, but please, I really don't like the worg breed...

14 Mar 2005, 19:10

Tomasz Grysztar

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

Tomasz Grysztar 14 Mar 2005, 20:14

Do you prefer "dupe"? For some reason do not. Wink

But seriously: "clone" is also a good alternative. Its meaning is quite similar in this context.
I preferred "breed", because it sounds almost like an English phrase: "breed 5 instructions", etc.

14 Mar 2005, 20:14

MCD

Joined: 21 Aug 2004
Posts: 602
Location: Germany

MCD 15 Mar 2005, 08:10

okay, this is only a question of taste. I think someone should start a poll on this.

15 Mar 2005, 08:10

revolution
When all else fails, read the source

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

revolution 18 Mar 2005, 06:23

Yes, the "breed" looks like a good idea.

But how about something similar to the way the assembler uses "repeat". Like this:

Code:

breed 20
call @f
dd %
@@:
end breed

And the normal arithmetic should allow using "(%-1)*4" type of coding to generate whatever range of values is needed.

Regarding the name: Hmm, somehow breed does not seem appropriate. Basically it is a preprocessor type of repeat so how about "prepeat" or "pre-repeat" or "multiple-generate" or "mgen". I don't think that trying to find a suitable English word is necessary.

Of course, the names suggested above are a little quirky, don't take them too seriously. But, instead is it possible to move "repeat" into the preprocessor stage? The % can still be replaced as usual and I think the code generated would not change but it would allow constructions just like I show above to work intuitively.

18 Mar 2005, 06:23

Tomasz Grysztar

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

Tomasz Grysztar 18 Mar 2005, 06:30

revolution wrote:

But, instead is it possible to move "repeat" into the preprocessor stage? The % can still be replaced as usual and I think the code generated would not change but it would allow constructions just like I show above to work intuitively.

No, it wouldn't be the same. Consider just:

Code:

repeat 1000-$
 db %
end repeat

And "breed" is rather a preprocessor's "times" directive than preprocessor's "repeat". And since you can use macro inside a line, which you duplicate, there is no need to implement any "repeat" equivalent.

18 Mar 2005, 06:30

Ancient One

Joined: 28 Feb 2005
Posts: 55

Ancient One 19 Mar 2005, 07:51

Privalov wrote:

Good idea, though I think I would still prefer something like:
Code:
breed [name,i] j line    
with the brackets-surrounded parameters being optional (i - first value, j - count of duplicates). Possible extended form:
Code:
breed [name,i,step] j line    
where step would be the value added to i after each duplicate.

yeah, it looks good to me.

i think the word 'breed' is a good name for this directive.

19 Mar 2005, 07:51

vid
Verbosity in development

Joined: 05 Sep 2003
Posts: 7105
Location: Slovakia

vid 21 Mar 2005, 14:05

breed is allright

21 Mar 2005, 14:05

Tomasz Grysztar

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

Tomasz Grysztar 10 May 2005, 11:30

Another new idea is to make something more similar to macro syntax, like:

Code:

prerep 15 
{
  ; source that will be repeated 15 times
}

prerep 15 counter
{
  ; in this source "counter" will be replaced with number of repetition
}

and also:

Code:

preif symbol eq other_symbol
{
  ; conditionally preprocessed code
}

The only problem it has that "breeed" has not is that using it within marcos would require fixing some symbols for the braces again (as with standard macros within macros).

10 May 2005, 11:30

UCM

Joined: 25 Feb 2005
Posts: 285
Location: Canada

UCM 10 May 2005, 21:06

I think 'breed' is ok.

_________________
This calls for... Ultra CRUNCHY Man!
Ta da!! *crunch*

10 May 2005, 21:06

revolution
When all else fails, read the source

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

revolution 11 May 2005, 01:06

Privalov. I think your suggestion of using the braces is much better than the single line approach mentioned for breed.

11 May 2005, 01:06

MCD

Joined: 21 Aug 2004
Posts: 602
Location: Germany

MCD 11 May 2005, 09:36

revolution wrote:

Privalov. I think your suggestion of using the braces is much better than the single line approach mentioned for breed.

I can only agree with you, especially because preprocessor in fasm is not (completely) line based, but also block based. (stuff like macors, structures etc...)

_________________
MCD - the inevitable return of the Mad Computer Doggy

-||__/
.|+-~
.|| ||

11 May 2005, 09:36

vid
Verbosity in development

Joined: 05 Sep 2003
Posts: 7105
Location: Slovakia

vid 11 May 2005, 09:46

yes, i don't like declaring new macro for each repeated block.

about the problem: couldn't you add something like i mentioned you before - some symbol, let's say "@", which will be normal part of name like ".", but every symbol starting with "@" will have it removed when unrolling macro.

Something like

Code:

macro declare_macro_XXX
{ 
   macro XXX
   @{
      asdas
   @}
}
declare_macro_XXX

So "@" thingy will be kind of equivalent for fixes, but much easier to use, more readable, more comprehendable, will allow macros independent from particular fixes etc.

i quess you would have to rewrite prepreprocessor a little not to separate @ and { symbol, please think if it could be done.

11 May 2005, 09:46

Tomasz Grysztar

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

Tomasz Grysztar 11 May 2005, 10:38

I think about something like:

Code:

macro first
{
   macro second
   {
      macro third
      {
          
      \\}
   \}
}

The backlash character is already treated specially by line splitter, so this wouldn't complicate anything. And there shouldn't be confusion with line concatenation operator, since only comments after it are allowed.

11 May 2005, 10:38

JohnFound

Joined: 16 Jun 2003
Posts: 3499
Location: Bulgaria

JohnFound 11 May 2005, 11:04

Hm, I wonder, isn't it possible to make clear recurcive macro behaviour. I mean:

Code:

macro first
{
   macro second
   {
      macro third
      {
          
      }
   }
}

This will need only a counter that to be incremented on every "{" and decremented on every "}", so, if after decrement, the counter is 0, then the "}" character should be treated as an end of macro definition, else should be treated as a simple character.

Regards

11 May 2005, 11:04

Tomasz Grysztar

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

Tomasz Grysztar 11 May 2005, 11:17

But what when you want to begin macro definition with one macro and end with another (like the current "struct" macro does)?

11 May 2005, 11:17

Tomasz Grysztar

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

Tomasz Grysztar 11 May 2005, 11:26

Implementing such backslashes feature is quite simple - I'm attaching the modified preprocessor module that has it.

Attachment removed - now it's implemented in official releases

Last edited by Tomasz Grysztar on 19 May 2005, 12:16; edited 1 time in total

11 May 2005, 11:26

JohnFound

Joined: 16 Jun 2003
Posts: 3499
Location: Bulgaria

JohnFound 11 May 2005, 11:30

Privalov wrote:

But what when you want to begin macro definition with one macro and end with another (like the current "struct" macro does)?

Well, "fix" is still here. Smile

11 May 2005, 11:30

Goto page Previous 1, 2, 3 Next

< 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