#307506* (?/31) ⚐Flag
< Frost> Is there a theory of why the file write might get truncated (or corrupted) on what appears to be a normal shutdown?
< Frost> Is it just that the files are so large that maybe the write didn't complete before the process exited?
<~Freeyorp> I guess. It's at least more sane than the files being filled with data from the other servers, seemingly at random.
<&o11c> Frost: the problem is probably that fprintf() failed for some reason, and it doesn't check that before calling lock_fclose
< Frost> oh! Now I understand where that discussion came from. Thanks
<&o11c> I'm not sure *why* fprintf would fail ...
<~Freeyorp> I'm not sure why a save database would be filled with data from what should have been a distinct process either, but it's happened more than once in the past
<&o11c> yeah, that's simply impossible
<&o11c> I hate it wheen that happens
<~Freeyorp> Yet it happened; tmwAthena has some black magic in there somewhere
<&o11c> but I haven't *written* the code that shares file descriptors between processes yet
* o11c blames the kernel
* Mistergrey blames Ali-G, as this seems to be what most devs do as a default :P |