#tde-devs < 2025/06/12 > |
|---|
| [00:37] | micheleC has joined |
| [00:38] | micheleC: @denk: any way to replicate kmail problem systematically? |
| [03:28] | micheleC has quit (Read error: Connection reset by peer) |
| [09:21] | SlavekB has quit (Quit: Kopete 0.12.7 : http://trinitydesktop.org) |
| [10:30] | SlavekB has joined |
| [15:42] | denk: micheleC, this happens from time to time, it depends of my activity in kmail |
| [15:43] | denk: yes, it is broken, but I can't exactly reproduce it |
| [15:44] | denk: when a list of mail is loading from a server I just click to a next message (or something like this) |
| [15:44] | denk: and it just crashes |
| [15:44] | denk: the problen is in the private library |
| [15:45] | denk: I rebuilt it with the debug info and I can see the symbols I've shown |
| [15:46] | denk: I'm sure it is a race condition, but how to make it reprodussible... |
| [15:47] | SlavekB: I can confirm that I also observe the crashes from time to time. It seems to me that most cases are when I delete a message - or more often a number of selected messages. |
| [15:47] | denk: maybe sparc or arm can help to do it |
| [15:48] | denk: SlavekB: usually I move messages to another folder and it will crash if I faster than the code :) |
| [15:49] | SlavekB: Yes, you are right that delete => actually moving into the trash is as common as moving to the archive, and that this is actually the same case of crash. |
| [15:49] | denk: btw, I use IMAP protocol from my CGP server, I did not try it with local folders |
| [15:50] | SlavekB: In my case it is also accounts on IMAP server. |
| [15:50] | denk: there is something wrong in the imap protocol implementation |
| [15:52] | denk: or in the api between IMAP and a local list |
| [15:52] | SlavekB: In any case, your note was very correct - I also observe the crashes when moving messages - sometimes the target is the archive, sometimes the trash, but in both cases it is moving. |
| [15:52] | denk: need more statistics |
| [15:53] | denk: or steps to reproduce it |
| [15:54] | SlavekB: In a crash report, I usually found that unfortunately this is not cases of null pointer dereference, which could be easy to fix. |
| [15:54] | denk: it is called use-after-free |
| [15:55] | denk: one thread calls free/delete, but some other thread tries to use the same pointer |
| [15:55] | denk: and boom! |
| [15:55] | SlavekB: Yes, exactly, I assume tha this is exactly the case. |
| [20:31] | SlavekB has quit (Quit: TDE Kopete : https://trinitydesktop.org) |
| [23:19] | SlavekB has joined |
#tde-devs < 2025/06/12 > |
|---|
