Hi Trent, Glad you are finally on board darcs. I don't see the attachment here for the diffs or patch file. Do you have a website or somesuch? Jimmy --- Trent Piepho <xyzzy at speakeasy.org> wrote: > Here is a pile of patches for the robot. Most of > have been discussed here > before, a few are new ones. For some reason, darcs > insisted that I include > the glib Makefile patch in this bundle. Just ignore > that one or kill the glib > patch I sent before and use this one instead. > > The patches, from oldest to newest: > > Rollback robotd-fixes1.dpatch, all these fixes will > be re-done here, but differently > > Mon May 22 15:40:42 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * fix mfprintf() > When mfprintf() was changed from varargs to > stdarg, it wasn't done > correctly. Fix this. > > Tue May 23 00:45:25 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * adapt when assaulting > The assault planet code didn't re-check the target > planet after the > assult started. If the planet was taken by > another player on a friendly > team, or modified by god, the robot would still > try to bomb/take it, even > when that action was no longer possible. The > robot will now check for > planets is can not drop on, and abort the assault > if so. It will not try > to bomb planets it can not bomb. It will also > stop reinforcing planets > when they get to 4 armies. > > Tue May 23 00:49:46 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * fix neutral planet check > Fix several bugs with check_take(). It didn't > handle the case when no > enemy planets were left. It looked at the last > enemy planet in the list, > rather than the target enemy planet when deciding > if it should take neut > planets. It would skip looking for neuts when it > found an enemy agri and > _didn't_ have enough armies to take it. > > Now it handles no enemy planets left. It only > skips looking for neut > planets in preference of enemy planets if it has > found an enemy agri and > it _does_ have enough armies to take it. In that > case, the agri take > will take precedence. > > Tue May 23 00:55:48 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * fix closest_planet > closest_planet() would usually return NULL every > other time it was > called. For speed, it would check the previous > closest planet first, > then look for a _closer_ planet. If it didn't > find one, it returned > NULL, rather than the previous (and current) > closest planet (which > it now returns). > > Tue May 23 01:00:23 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * fix crash in RCD generation > The code to generate the RCD message didn't handle > the case when there > were no visible friendly and/or enemy ships. In > this case, > _state.closest_{e,f} would be NULL. Do what > clients do in this case, and > use me->p_no when no suitable other player is > known. > > Tue May 23 01:08:45 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * fix use of un-initialized variable > Fix a warning about a un-initialized use of pldist > in update_players(). > It didn't matter, but fix the warning anyway. > > Tue May 23 01:54:09 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * fix calls to req_cloak_off() > Some calls to req_cloak_off() were missing the > function's argument > (reason string). > > Tue May 23 03:08:06 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * fix res danger > The check for res danger would make the robot > cloak when near _any_ home > planet, so it would cloak even when bombing it's > own core. Unfortunately, > it is not possible to tell from the planet flags > whose home world a > planet is. Changed the code to only check the > normal home planet > indexes, skipping it's own HW, for proximity. > This makes the code much > smaller and faster. Realistically, the HWs are > always at the standard > indexes, even with non-standard planet layouts. > Included a bit of code, > #if'ed out, that will find home worlds by name, > the way the bot does when > it wants to refit. > > Tue May 23 04:09:21 PDT 2006 Trent Piepho > <xyzzy at speakeasy.org> > * torp dir for robot > Hadley's bot expects to get torp direction from > the server, but the > server only sends it for your own torps. Someone > probably changed this > years ago to combat borgs. Changed back to send > it for other > torps too. Without it, the bot thinks all torps > are going straight up, > making it a very bad dodger (unless you attack > from below). > > This should be fixed to only turn this on for the > robot, so as not to > help borgs. This will have to wait on a better > way to detect robots. > > _______________________________________________ > netrek-dev mailing list > netrek-dev at us.netrek.org > http://mailman.us.netrek.org/listinfo/netrek-dev >