Le 2018-11-08 12:02, Klemens Nanni a écrit :
On Thu, Nov 08, 2018 at 10:50:47AM +0000, Stuart Henderson wrote:
> ccam$ mbsync -a
> C: 0/1 B: 0/6 M: +0/0 *0/0 #0/0 S: +0/0 *0/0 #0/0
> Warning: lost track of 553 pulled message(s)
> C: 0/1 B: 0/6 M: +0/0 *0/0 #0/0 S: +0/556 *0/0 #0/0Abort trap (core dumped)
>
> And at the end of the dmesg, we can find the pledge error.
You mentioned in subject, but it would be clearer to include this here
too, e.g.
mbsync[96869]: pledge "fattr", syscall 76
Using your config (which was useful) I can confirm it.
This is triggered by the `CopyArrivalDate' option I did not account for
back when adding pledge.
syscall 76 is utimes() which is done here in maildir_store_msg()
in drv_maildir.c:
1636 if (data->date) {
1637 /* Set atime and mtime according to INTERNALDATE
or mtime of source message */
1638 struct utimbuf utimebuf;
1639 utimebuf.actime = utimebuf.modtime = data->date;
1640 if (utime( buf, &utimebuf ) < 0) {
1641 sys_error( "Maildir error: cannot set
times for %s", buf );
1642 cb( DRV_BOX_BAD, 0, aux );
1643 return;
1644 }
1645 }
This port diff should fix things.
Yes, OK kn.
Thanks Stuart for the quick handling!
Thanks you Klemens and Stuart, both for the patch and for the posting
tips !
Sorry to have missed the error line in the debug, it showed when I typed
"dmesg" but not in the "sendbug" and I failed to double check it.
