Re: [Bug-gnubg] The match and game data structure

[Øystein Johansen]

> Perhaps I've coded too much C++... 
> I think a better exercise is to separate the logic
> from the data and think more in terms of
> modules/interfaces and let this drive the data structures.

mmmhhmm...

> In practice this would mean calling functions to
> work with matches and games rather than using
> globals/structs throughout the code.

Yes! Exactly!

> As an example look at this code from show.c:
>
> for( pl = lMatch.plNext; pl != &lMatch; pl = pl->plNext, ++n )
{
>      pmr = (moverecord *) ( (list *) pl->p )->plNext->p;
>      pmgi = &pmr->g;
>      assert( pmr->mt == MOVE_GAMEINFO );
>       ...

> It's pretty hard to work out what the code in this
> example does, which kind of says something...

Yes, you're pinpointing the problem with this example code. 

> Anyway it could look something like:
>
> for (i = 1; i < getNumGames(); i++)
> {
>      game *g = getGame(i);
>      pmgi = getGameInfo(g);
>      ...

Here's my (first) suggestion:

GList game = match->games;

for ( ; game ; game = game->next ){
  GList move = game->moves;
  for ( ; move ; move = move->next )
    analyse(move);  /* or do something else */
}

and a data structure for game and match like in the attached file.

-Øystein

game.h
Description: Text document