frosch wrote:

The hardest part is missing :(
The location in the industry layout should be signed just as for type 0xFF.

Edit:
Also "/* it->gfx is stored in the map. But the translated ID cur_gfx is the interesting one */" might need some care.

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602#comment7555

andythenorth wrote:

I modified newgrf.cpp to allow signed offsets. Currently all tiles can be built with signed offset, which will cause problems. I need to prevent a tile with signed offset from being constructed, I haven't figured that out yet.
There was also some exciting irc debate about the format of itt[k].ti.x = (int8)GB(itt[k].ti.x, 0, 8);

Attachments

• tile_special_dontbuild_flag_signed_offsets.diff (1.70 KiB)

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602#comment7556

andythenorth wrote:

I've modified industry_cmd.cpp further to prevent tiles with signed offsets being built.
I also moved a couple of lines as suggested by frosch in resp. of /* it->gfx is stored in the map. But the translated ID cur_gfx is the interesting one */

Not building is handled silently - no errors are reported.
IMO the newgrf author has done something stupid if they use negative offsets without prop 12 bit 1 set (i.e. don't build this tile).

EDIT: I wonder if CheckIfIndustryTilesAreFree needs to exclude tiles with negative offsets unless prop 12 bit 1 is set for tile?

Attachments

• tile_special_dontbuild_flag_signed_offsets.diff (2.28 KiB)

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602#comment7557

andythenorth wrote:

Update - CheckIfIndustryTilesAreFree now returns false if offsets are < 0 and tile prop 12 bit 1 is not set. (This is now checked in two places which seems redundant)

Attachments

• tile_special_dontbuild_flag_tile_checks.diff (6.07 KiB)

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602#comment7560

andythenorth wrote:

frosch was pretty adamant that negative offsets need checking in a better way with AfterLoadGRFs

This is my attempt to do that. It doesn't quite work, and I'm not sure why. It's not reading the special flag for tile prop 12 bit 1 correctly. I think I'm probably checking the wrong tile?

Attachments

• industry_tlle_negative_offset_check.diff (3.64 KiB)

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602#comment7564

Yexo wrote:

The latest change in src/industry_cmd.cpp is wrong. GetIndustryTileSpec(GFX_WATERTILE_SPECIALCHECK) is not valid, it's never done in trunk because the check for GFX_WATERTILE_SPECIALCHECK is first, but you move those calls around so GetIndustryTileSpec is called for that id.

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602#comment7565

andythenorth wrote:

Yexo - I've added some conditional code to check tile id before calling GetIndustryTileSpec. Also seemed easiest to set a little flag to control wherever the tile is built or not.

Attachments

• industry_tlle_negative_offset_check.diff (4.00 KiB)

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602#comment7566

andythenorth wrote:

For some reason I had ruled out using industry tile cb 2F and industry tile var 60 to solve this. That was a mistake.

Using cb 2F and var 60, I can simply check the contents of a tile's 8 neighbours. If they contain another industry (or other unremovable object), then I can prevent the industry from being built.

This should render my patch attempts for this redundant.

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602#comment7567

Rubidium closed the ticket.

Reason for closing: Won't implement

Not needed; can be done using other callback/variable.

This comment was imported from FlySpray: https://bugs.openttd.org/task/3602