Follow-up Comment #5, bug #19029 (project freeciv): So that suggests on at least two platforms, including one popular one ;), there's no actual data loss, just permutation. I checked my savefile and it's consistent with the numbers from the test progam.
Meanwhile, a question. In the per-player part of the savefile, there's the saved terrain, playerX.map_tYYYY (saved in sg_save_player_vision()). For each map tile, this saves a single character representing what the player believes the terrain to be, *or 'u' for unknown terrain*. Specifically, it's 'u' if map_get_player_tile(ptile, plr)->terrain is NULL. Is there any circumstance where that condition isn't identical to map_is_known(ptile,plr) returning FALSE? On a quick look, I haven't found any sanity checks or similar asserting that this is so for every tile every turn, or on loading a savefile, or what have you, but I haven't thought of a way in which they can come adrift from each other. I think map_show_tile() is the primary way in which this state changes. If so, the state being saved here is redundant. Normally I think we'd remove it for the next major release, but here we could just switch to generating the known info from the player map and lose the need for backward compatibility code? _______________________________________________________ Reply to this item at: <http://gna.org/bugs/?19029> _______________________________________________ Message sent via/by Gna! http://gna.org/ _______________________________________________ Freeciv-dev mailing list Freecivfirstname.lastname@example.org https://mail.gna.org/listinfo/freeciv-dev