I think i applied the patch, how could i be sure ?
It would be very simple if you push #44504 comits (and all other 3.0 patches) in the need branches, so i just pull and recompile and test, and check git-rev
I can reproduce. Seems to be because terrain transforming activities are handled separately from other activities fixed in #44504 + irrigation & mine building activities are wrongly considered as terrain transformation activities in S3_0.
Correct fix for S3_0 probably has two parts: Stop considering irrigation / mine building as terrain transformations, and also handle the (remaining) terrain transformation activities correctly. For later branches former is probably already handled by the separation of Irrigation/Cultivate and Mine/Plant activities, but latter part is likely needed.
Reply To cazfi
Stop considering irrigation / mine building as terrain transformations ... For later branches former is probably already handled by the separation of Irrigation/Cultivate and Mine/Plant activities
Nope, the bug has been faithfully reimplemented in the new system.
Further investigation has also revealed that there's additional bug in that in this special cased terrain transformations handling *other* activities of *other* units turned illegal because of the terrain change is not handled at all. That seems like it has been a bug actually affecting the game for a long time (has not been only dormant bug, like the other issues turned active by the recent orders cancelling change) It seems to me that after fixing all these issues, terrain transforming actions are not special case at all (their handling is exactly the same code as the "else" branch)
Different patch for each branch:
- S3_0 differs already because the branch has the extra building and terraing transforming variants of the irrigation and mine activities in the same activities. Overall did the minimum changes to fix the bug (no unnecessary changes to a release branch)
- S3_1 fix the bugs, but does not refactor the special casing of terrain changing actitivies away. Difference to S3_0 in this respect is that S3_1 adds fc_assert() to check that indeed it the refactoring would be safe
- master does the full refactoring
Not tested much yet. Wanted to submit anyway, as waiting that I get to testing these could postpone the submission significantly.
in 3.01
same kind of message as in #44504
SHIFT-I stops at the first tile