View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0011729 | Dwarf Fortress | Technical -- Saving/Loading | public | 2021-03-24 07:22 | 2021-03-24 16:28 |
| Reporter | Spiky | Assigned To | |||
| Priority | normal | Severity | major | Reproducibility | have not tried |
| Status | new | Resolution | open | ||
| Platform | Linux | OS | Debian 10.8 | ||
| Product Version | 0.47.05 | ||||
| Summary | 0011729: Savegame corruption | ||||
| Description | My fortress got corrupted and now won't load. I'd been playing for several days, and have no backups so don't really have a good idea where to start looking for a place to reproduce it beyond 'make a similar world and play for a few days and see what happens' The offending save is at https://spikycaterpillar.com/misc/region2.lockup.tar.bz2 I disabled the libstdc++ that shipped with the game because it conflicts with Debian's libGLU, so it's using the system libstdc++ instead. Several load/save cycles before the save became unloadable, I had loaded the fort, saved the fort, loaded it again without quitting the game, saved it again, and quit. When I quit, there was some sort of memory management error on the console - a double free, I think. However, as the game loaded and saved without incident for a few days after that, I don't know if it was in any way relevant. | ||||
| Additional Information | This fortress contains some of my favorite fun things - a high-savagery world with lots of necromancer activity. It looks like my civ was taken over by dark forces years ago, given that necromancers and undead can serve openly in my military and fragmentary recollections from surviving witnesses say the most popular temple was to the god of theft and death... I suspect, possibly irrationally, that the necromancy is in some way to blame for the corruption. This is the same save as 0011728. | ||||
| Tags | No tags attached. | ||||
|
|
I have tracked down WHERE the compression error is. It is chunk 1625 (zero-based) and it appears to be worked around by changing byte 837 of the compressed chunk to (decimal) 181. Also, after fixing it, load-and-resave does not recreate the problem. Which might be a hardware error on my computer or something... but I now have a script to try to brute-force recover compression errors if anyone with a similarly bugged save wants to fix it. https://spikycaterpillar.com/spiky_df_uncorruptor/ |
