Hi.
running git commit id now:
GIT update -->
repository: http://git.dbmail.eu/paul/dbmail
at the commit: abac5b58c94f270e534320ff264054bc79530b78
branch: dbmail_3_1
storage directory: "/usr/portage/distfiles/egit-src/dbmail"
checkout type: bare repository
For testing purposes I did make a git live ebuild for the 3.1 branch now
so any Gentoo users feel free to add my dev overlay (lordvan):
My layman git overlay:
http://git.overlays.gentoo.org/gitweb/?p=dev/lordvan.git;a=summary
Regards
On 2013-09-26 17:10, Paul J Stevens wrote:
On 26-09-13 15:57, Thomas Raschbacher wrote:
Hi.
it just happened again:
read(15, 0xa0d150, 11) = -1 EAGAIN (Resource
temporarily unavailable)
read(15, 0xa0d150, 11) = -1 EAGAIN (Resource
temporarily unavailable)
So read-events on a network socket are being triggered.
#0 0x00007fc2fbf6732a in ci_write () from
/usr/lib64/dbmail/libdbmail.so.0
#1 0x00007fc2fbf6742c in ci_write_cb () from
/usr/lib64/dbmail/libdbmail.so.0
#2 0x0000000000408018 in imap_cleanup_deferred ()
But you're also in a deferred cleanup state. During that state, only
writes should happen, and only if the connected client can be written
to
succesfully. Imagine a stdin/stdout connected setup where stdin was
closed (EOF) but there is still data pending for stdout.
I've pushed a small change that might - hopefully will - help here.
http://git.dbmail.eu/paul/dbmail/commit/?h=dbmail_3_1&id=abac5b58c94f270e534320ff264054bc79530b78
If you don't like patching you can pick up a tar-ball there as well.
_______________________________________________
DBmail mailing list
[email protected]
http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail
