Re: [Freeciv-Dev] [patch #1349] Rework (+fix) auto-settler code
Pepeto writes: > Le dimanche 01 novembre 2009 à 16:16 +0100, Goswin von Brederlow a > écrit : >> pepeto writes: >> >> > Update of patch #1349 (project freeciv): >> > >> > Status: In Progress => Ready For Test >> > >> > >> > ___ >> > >> > Follow-up Comment #2: >> > >> > Updated version of the patch against current svn. >> > >> > Also: >> > * Remove potential crash on tile_index(NULL). >> > * Cleanup in build city stuff: >> > * All settlers use the same pf_parameter. >> > * Units dying are not considered as a failure. >> >> If you have too many auto settler some will remain idle. If one dies >> then it should retask one of the idle ones. > > Yes, it doesn't change that behaviour. > >> > * Units which don't reach the position at the turns are not considered >> > as a >> > failure. >> >> If a unit can't reach its destination it should pick some other >> task. The failure makes the outer loop repeat assigning tasks. > > Yes, I agree, and it does that. I am talking about the settlers able to > reach their destination but the path is too long to be reached in one > turn. In your patch for building city, they were considered as failed > mission. Should have read the diff between the two patches. The patch currently never releases auto-worker when there is nothing to do. The old code made them not auto-workers so human players could assign them new roles. I would like to use something inbetween. Have them keep their "auto" flag but bring them into focus in the client. If the client assigns no new activity to them they stay auto-worker and do more work whenever some comes along. Any idea how to do that? MfG Goswin ___ Freeciv-dev mailing list Freeciv-dev@gna.org https://mail.gna.org/listinfo/freeciv-dev
Re: [Freeciv-Dev] [patch #1349] Rework (+fix) auto-settler code
Le dimanche 01 novembre 2009 à 16:16 +0100, Goswin von Brederlow a écrit : > pepeto writes: > > > Update of patch #1349 (project freeciv): > > > > Status: In Progress => Ready For Test > > > > ___ > > > > Follow-up Comment #2: > > > > Updated version of the patch against current svn. > > > > Also: > > * Remove potential crash on tile_index(NULL). > > * Cleanup in build city stuff: > > * All settlers use the same pf_parameter. > > * Units dying are not considered as a failure. > > If you have too many auto settler some will remain idle. If one dies > then it should retask one of the idle ones. Yes, it doesn't change that behaviour. > > * Units which don't reach the position at the turns are not considered as > > a > > failure. > > If a unit can't reach its destination it should pick some other > task. The failure makes the outer loop repeat assigning tasks. Yes, I agree, and it does that. I am talking about the settlers able to reach their destination but the path is too long to be reached in one turn. In your patch for building city, they were considered as failed mission. ___ Freeciv-dev mailing list Freeciv-dev@gna.org https://mail.gna.org/listinfo/freeciv-dev
Re: [Freeciv-Dev] [patch #1349] Rework (+fix) auto-settler code
pepeto writes: > Update of patch #1349 (project freeciv): > > Status: In Progress => Ready For Test > > ___ > > Follow-up Comment #2: > > Updated version of the patch against current svn. > > Also: > * Remove potential crash on tile_index(NULL). > * Cleanup in build city stuff: > * All settlers use the same pf_parameter. > * Units dying are not considered as a failure. If you have too many auto settler some will remain idle. If one dies then it should retask one of the idle ones. > * Units which don't reach the position at the turns are not considered as a > failure. If a unit can't reach its destination it should pick some other task. The failure makes the outer loop repeat assigning tasks. > * Don't use unit_list_safe_iterate() when not needed. > * Style cleanup. MfG Goswin ___ Freeciv-dev mailing list Freeciv-dev@gna.org https://mail.gna.org/listinfo/freeciv-dev
