Re: [evolution-patches] [PATCH] prevent evolution from crashing if new-mail-notification is enabled but D-BUS session daemon is not running
- From: Not Zed <notzed ximian com>
- To: Timo Hoenig <thoenig suse de>
- Cc: Evolution Patches <evolution-patches lists ximian com>
- Subject: Re: [evolution-patches] [PATCH] prevent evolution from crashing if new-mail-notification is enabled but D-BUS session daemon is not running
- Date: Thu, 12 May 2005 10:14:36 +0530
On Wed, 2005-05-11 at 11:01 +0200, Timo Hoenig wrote:
Hi,
On Fri, 2005-05-06 at 14:58 +0530, Not Zed wrote:
> I've committed the newest version of this, plus a few changes to make
> the code a lot simpler. I removed the separate enabling of dbus events
> config, and it just sends them always if the plugin is enabled & it
> can find the bus.
[...]
Yes, it looks much cleaner now. But your modifications also lead to two
situations where Evolution will crash, which were covered by my original
patch.
* Since the connection to the session bus is only established once
at initialization of the plugin, Evolution will die once the
D-Bus session bus is killed.
How come? Are you saying the life of the object returned by get bus is only the scope of the call in which it was made?
Are you saying that dbus is so braindead that it doesn't handle these cases internally?
* If dbus_message_new_signal() fails (e.g. OOM), NULL is
returned. Once we send such an invalid message to the session
bus, Evolution will die.
Evolution will die automaticaly if it is out of memory. glib enforces this.
Please let me know if you're willing to accept a patch which helps to
prevent these situations.
Thanks,
Timo
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]