[freeciv-ai] coreai: assess_danger()

["Per I. Mathisen" <per@xxxxxxxxxxx>]

This is big one.
 - Much rewriting. Almost every line of code touched in some way, since I
unnested several ifs by reversing their condition and doing a continue; on
the loop instead. Much, much cleaner.
 - Fixed case where less dangerous situation would result in lower want
for city walls than a slightly more dangerous situation (danger > defense,
for high values of each, would give less want than the default want in
danger situation). Now gives AI slightly more wall want in some cases.
 - Wrapped up the convoluted military building want algorithms into a more
consise form, and put it into a separate function. Use the more
sophisticated wall algorithm also for coastal, sam and sdi.
 - Did not follow Ross' code all the way in giving all cities a default
danger. This I think is fundamentally wrong, and I will change this later
to use the cached information created by ai_eval_* to only give cities
situated dangerously geographically default danger.
 - Left the "will this enemy city complete a dangerous unit next turn"
cheesiness in the code. I wanted to remove it but had second thoughts.

Yours
Per

coreai2.diff
Description: Text document