daemon: allow "eavesdrop" match rule

daemon: allow "eavesdrop" match rule

Recent dbus-monitor supplies "eavesdrop" match rule which ibus-daemon
does not accept.  This patch silently ignores the rule so ibus-daemon
behave the same as before.

BUG=none

Committed: f9ddda1

[Daiki Ueno, 2013-01-29 09:12:32]

Recent dbus-monitor specifies "eavesdrop=true" match rule which ibus-daemon does
not accept:

$ dbus-monitor --address $IBUS_ADDRESS
Error: Parse match rule [eavesdrop=true,type='signal'] failed
zsh: exit 1     dbus-monitor --address $IBUS_ADDRESS

[Peng, 2013-01-29 15:33:22]

https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c
File bus/matchrule.c (right):

https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c#newcode532
bus/matchrule.c:532: g_dbus_message_get_destination (message) != NULL)
Could you explain how eavesdrop works?

[Daiki Ueno, 2013-01-30 02:15:52]

https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c
File bus/matchrule.c (right):

https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c#newcode532
bus/matchrule.c:532: g_dbus_message_get_destination (message) != NULL)
On 2013/01/29 15:33:22, Peng wrote:
> Could you explain how eavesdrop works? 

From the D-Bus spec:
http://dbus.freedesktop.org/doc/dbus-specification.html#message-bus-routing-m...

"...match rules do not match messages which have a DESTINATION field unless the
match rule specifically requests this by specifying eavesdrop='true' in the
match rule. eavesdrop='false' restores the default behaviour."

This probably means that a client shall not receive any unicast messages whose
destination is set to other client.

[Peng, 2013-01-30 02:19:42]

On 2013/01/30 02:15:52, Daiki Ueno wrote:
> https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c
> File bus/matchrule.c (right):
> 
> https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c#newcode532
> bus/matchrule.c:532: g_dbus_message_get_destination (message) != NULL)
> On 2013/01/29 15:33:22, Peng wrote:
> > Could you explain how eavesdrop works? 
> 
> From the D-Bus spec:
>
http://dbus.freedesktop.org/doc/dbus-specification.html#message-bus-routing-m...
> 
> "...match rules do not match messages which have a DESTINATION field unless
the
> match rule specifically requests this by specifying eavesdrop='true' in the
> match rule. eavesdrop='false' restores the default behaviour."
> 
> This probably means that a client shall not receive any unicast messages whose
> destination is set to other client.

Thanks. lgtm. Please those info in git log or source code.

[Daiki Ueno, 2013-02-25 09:05:45]

On 2013/01/30 02:19:42, Peng wrote:
> On 2013/01/30 02:15:52, Daiki Ueno wrote:
> > https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c
> > File bus/matchrule.c (right):
> > 
> > https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c#newcode532
> > bus/matchrule.c:532: g_dbus_message_get_destination (message) != NULL)
> > On 2013/01/29 15:33:22, Peng wrote:
> > > Could you explain how eavesdrop works? 
> > 
> > From the D-Bus spec:
> >
>
http://dbus.freedesktop.org/doc/dbus-specification.html#message-bus-routing-m...
> > 
> > "...match rules do not match messages which have a DESTINATION field unless
> the
> > match rule specifically requests this by specifying eavesdrop='true' in the
> > match rule. eavesdrop='false' restores the default behaviour."
> > 
> > This probably means that a client shall not receive any unicast messages
whose
> > destination is set to other client.
> 
> Thanks. lgtm. Please those info in git log or source code.

Thinking again, strictly emulating "eavesdrop" rule makes little sense for ibus.
 So I've simplified the CL to just ignore the rule (and updated the commit log).
 If no objection, I'll commit this instead of the original one.

[Peng, 2013-02-25 15:17:50]

On 2013/02/25 09:05:45, Daiki Ueno wrote:
> On 2013/01/30 02:19:42, Peng wrote:
> > On 2013/01/30 02:15:52, Daiki Ueno wrote:
> > > https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c
> > > File bus/matchrule.c (right):
> > > 
> > > https://codereview.appspot.com/7225062/diff/1/bus/matchrule.c#newcode532
> > > bus/matchrule.c:532: g_dbus_message_get_destination (message) != NULL)
> > > On 2013/01/29 15:33:22, Peng wrote:
> > > > Could you explain how eavesdrop works? 
> > > 
> > > From the D-Bus spec:
> > >
> >
>
http://dbus.freedesktop.org/doc/dbus-specification.html#message-bus-routing-m...
> > > 
> > > "...match rules do not match messages which have a DESTINATION field
unless
> > the
> > > match rule specifically requests this by specifying eavesdrop='true' in
the
> > > match rule. eavesdrop='false' restores the default behaviour."
> > > 
> > > This probably means that a client shall not receive any unicast messages
> whose
> > > destination is set to other client.
> > 
> > Thanks. lgtm. Please those info in git log or source code.
> 
> Thinking again, strictly emulating "eavesdrop" rule makes little sense for
ibus.
>  So I've simplified the CL to just ignore the rule (and updated the commit
log).
>  If no objection, I'll commit this instead of the original one.

lgtm