RE: [Civlog-list] CivLog-0.6.1

[Michael G. <mic...@gr...>]

Hi!

Raimar wrote:
> On Mon, Sep 25, 2000 at 05:03:22PM +0200, Raimar Falke wrote:
> > On Mon, Sep 25, 2000 at 10:32:33PM +0800, Wong TM (Huang Deming) wrote:
> > > 
> > > Changes:
> > > 
> > >     * Ulrich have made a lot of changes to restructure the data structure:
> > >         - .......wait till I read through the mail.
> > 
> > I think you made the inclusion a bit to fast. The work of Ulrich has made
> > aware some questions wrt Score. We now think we how to do it the "right way
> > [tm]" so wait for a final version which should come in the next days.
> 
> As promised here is the updated version. It is the result of collaboration
> between Ulrich and me. We came to the conclusion that a second loop over all
> lines is unnecessary. This also kills the huge "String lines[];" thing.

Nice.

> A remaining problem are "extra player" (player which aren't declared in the
> header). The current solution is to create an array which is big enough to
> hold every number of players. Fortunate freeciv is currently limited to 32
> in this respect.
> 
> Another remaining problem is the current implemenation of the relative modus
> of PolyLineGroup. It may be possible to let extra players start somewhere in
> the graph (see line 380 in Graph) however this doesn't work in the relative
> modus. So unknown values are currently set to 0.

Another problem is what I reported earlier: the player-numbering is not always 
continuous. I speculated that this may be a bug in freeciv. It is not.
from report.c:
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
  for (i = 0; tags[i]; i++)
    {
      for (n = 0; n < game.nplayers; n++)
	{
	  if (is_barbarian (&(game.players[n])))            <---------------- NOTE!
	    {
	      continue;                                     <---------------- NOTE!
	    }
	  switch (i)
	    {
	    case 0:
	      fom = total_player_citizens (&(game.players[n]));
	      break;
	    case 1:
	      fom = game.players[n].score.bnp;
	      break;
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

Barabarians are not added to the log, so they generate gaps in the player 
numbering.

When I open a score-file that has such a gap, 0.6.1-mod breaks:

Exception occurred during event dispatching:
java.lang.RuntimeException: Incorrect order
        at org.freeciv.civlog.data.AbstractScore.addPlayer(AbstractScore.java:87)
        at org.freeciv.civlog.data.SimpleScore.loadScoreFile(SimpleScore.java:246)
        at org.freeciv.civlog.data.SimpleScore.loadScoreFile(SimpleScore.java:55)

I don't know if it still works if you just remove throwing that exception?

  Michael 

-- 
eMail: mic...@gr...
PGP key available on request