View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0009761 | Dwarf Fortress | World Generation -- General | public | 2016-05-16 20:57 | 2017-03-15 20:50 |
| Reporter | TV4Fun | Assigned To | lethosor | ||
| Priority | normal | Severity | crash | Reproducibility | always |
| Status | assigned | Resolution | open | ||
| Platform | Windows | OS | Windows 8.1 | OS Version | 8.1 |
| Product Version | 0.43.02 | ||||
| Summary | 0009761: Dwarf Fortress crashes when exporting map from worldgen or legends mode | ||||
| Description | Every single time I press 'p' to export a map, either after worldgen or in legends mode, the game crashes. This is vanilla unmodded DF that I have tried under Windows 8.1 64-bit and under a virtual machine running Windows XP 32-bit. The crash is consistent. Specifically, it appears that DF is passing a null SDL_Surface pointer to SDL_SaveBMP_RW. This only seems to affect regular map export and not detail map export. | ||||
| Steps To Reproduce | 1. Either: a. Generate a new very large world or b. Open Legends mode 2. Press 'p'. | ||||
| Additional Information | I have a crash dump from one of these crashes at https://drive.google.com/file/d/0B7QYgbtpmzDPeFdwZDNxSW95aGs/view?usp=sharing | ||||
| Tags | No tags attached. | ||||
|
|
I couldn't get this to reproduce after a few tries. - Are you using a tileset? - Do you have specific worldgen parameters where this occurs? Did the regionX-world_gen_param.txt file get saved before DF crashed, maybe? - What PRINT_MODE are you using (check init.txt)? |
|
|
PRINT_MODE is 2D. Not using a tileset. region1-world_gen_param.txt was not saved, but the parameters I was using were: World Size: Large History: Very Long Number of Civilizations: Very High Maximum Number of Sites: Very High Number of Beasts: Very High Natural Savagery: Very High Mineral Occurrence: Frequent One other thing to note is that DF did successfully create an empty file named region1-01050-01-01-world_map.bmp before it crashed. |
|
|
This seems to be fixed in 43.05. If I had to guess, I'd say it's hitting the 32-bit memory limit and failing to allocate the buffer it needs for its image file in earlier versions. There should be better checks for failed heap allocations though. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2016-05-16 20:57 | TV4Fun | New Issue | |
| 2016-05-17 16:57 | lethosor | Note Added: 0035209 | |
| 2016-05-17 16:57 | lethosor | Assigned To | => lethosor |
| 2016-05-17 16:57 | lethosor | Status | new => feedback |
| 2016-05-18 11:08 | TV4Fun | Note Added: 0035228 | |
| 2016-05-18 11:08 | TV4Fun | Status | feedback => assigned |
| 2017-03-15 20:50 | TV4Fun | Note Added: 0036320 |
