Follow-up Comment #2, patch #3856 (project freeciv):
Yes, the AI is looking for cities to defend with the reversed nativity check,
and yes, it just makes the AI suboptimal (doesn't try as hard to defend
non-coastal cities, oceanic-native units don't end up being used to defend if
their assigned city goes away for some reason (assuming that they happened to
be selected as defenders anyway, which is fairly unlikely (and even moreso
since revision 22826))). There are indeed no catastrophic consequences, and
no point backporting to 2_3.
I could open a separate ticket for only this issue, but I'm unsure which of
the other changes should be disincluded for that. 2_4 has the more complex
transport stuff, so I would think the change in wipe_unit() ought also be
included (consider that UTYF_UNDISBANDABLE units are killed outright for maps
without coastal cities when attempting to teleport with the current code: this
may be an entirely different bug). The editor thing is mostly just cleanup,
and doesn't need to be in 2_4, but if someone wanted to use any ruleset with
interesting nativity, there is a fair chance of unexpected failure to assign
homecities for editor-created units.
Of course, if there are changelog- or project-coordination- reasons to have
multiple commits, I don't mind separating them, but one of them will have the
entire implementation (and change every line in the previously attached
patch), and the others will be one-line changes for the precise arguments to
the extended function.
Reply to this item at:
Message sent via/by Gna!
Freeciv-dev mailing list