#tde-devs < 2026/07/06 >
[02:14]micheleC has joined
[06:32]denk: micheleC: I think it is the same bug when I move a message to another folder (via imap)
[07:23]micheleC: can replicate it systematically? or it happens randomly?
[07:56]denk: randomly
[07:56]denk: crash on reply, crash on moving
[07:57]denk: by the core dump I see use-after-free on some pointers
[07:58]denk: probably some object freed but then some list uses it again
[08:00]Fat-Zer has quit (Ping timeout: 265 seconds)
[08:00]denk: I know that it happens in different threads
[08:01]denk: also I know it is somewhere in the private library
[08:01]denk: but it could be the finalof the problem
[08:01]denk: *final of
[08:03]denk: I tried to send a message via mail(x) and reply to the new message, it does not crash in the test
[08:04]denk: I have no good ideas how to reproduce it
[08:09]denk: maybe it requires some additional headers in the message
[08:13]Fat-Zer has joined
[08:35]micheleC has quit (Ping timeout: 264 seconds)
[11:11]SlavekB has quit (Quit: Kopete 0.12.7 : http://trinitydesktop.org)
[11:54]micheleC has joined
[11:55]micheleC: ok thanks. It's most likely some memory corruption that happens in specific sequence of events, resulting in a pointer being freed in one thread and still used in another. very tricky to troubleshoot
[11:56]micheleC: unless you profile the whole application and wait till it happens again and then are prepare to go through huge quantities of log info
[11:58]denk: https://privatebin.net/?ecbf36d59fae95ce#5pTa6GYzZKZFot6MfZpQigVAU4XL7rZcwmJDeFzRc6aM
[11:58]tde-bot: Page title: PrivateBin
[11:58]denk: need more statistics? :)
[11:59]micheleC: those crash dumps will probably all show a similar scenario of an invalid pointer still used somewhere. But I am pretty sure they won't tell you what happened before and what led to the invalid pointer in first place
[12:01]denk: well, google says new internally calls malloc(), it could be used to debug memory allocations
[12:06]micheleC: yeah, it requires a deep profiling session basically
[12:06]micheleC: you could try to Valgrind-it but probably it would run too slow and it would generate way too much info if the crash is random
[12:06]denk: I reran kmail with MALLOC_CONF=junk:true,utrace:true,xmalloc:true
[12:07]denk: but I have some doubts it will crash today
[12:09]micheleC: ok
[12:28]SlavekB has joined
[13:49]micheleC has quit (Quit: Kopete 0.12.7 : http://trinitydesktop.org)
[22:25]SlavekB has quit (Quit: Kopete 0.12.7 : http://trinitydesktop.org)
[23:05]SlavekB has joined

#tde-devs < 2026/07/06 >