[Bug 28793] New: consider having PendingOperation ref itself and its parent until finished

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=28793

           Summary: consider having PendingOperation ref itself and its
                    parent until finished
           Product: Telepathy
           Version: git master
          Platform: Other
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: medium
         Component: tp-qt4
        AssignedTo: telepathy-bugs at lists.freedesktop.org
        ReportedBy: simon.mcvittie at collabora.co.uk
         QAContact: telepathy-bugs at lists.freedesktop.org
                CC: andrunko at gmail.com


PendingOperation and its subclasses don't guarantee to stay alive until the
operation finishes.

The equivalent constructs in the GLib world (GAsyncResult, TpProxyPendingCall
etc.) tend to keep a reference to themselves and the object they belong to for
the duration of the call, for greater predictability. We should consider doing
the same in telepathy-qt4.

I haven't investigated whether there are precedents either way in Qt/KDE.

13:35 < albanc> smcv, andrunko :  I don't keep a reference on the 
                PendingChannelRequest object returned by 
                Tp::Account->ensureChannel. I guess the object is released too 
                soon. I also have this warning:
13:35 < albanc> Telepathy-Qt4 (version 0.3.4) WARNING: 
                Tp::PendingOperation(0x326b40) still pending when it was 
                deleted - finished will never be emitted
13:35 < smcv> albanc: yeah don't do that :-P
...
13:38 < smcv> andrunko: would it make sense for the PendingOperation to keep 
              itself alive artificially until it finishes?
13:39 < smcv> andrunko: (a GAsyncResult keeps itself and its parent object 
              alive)
13:40 < andrunko> smcv, maybe, the problem is that if we have a bug and nobody 
                  calls finish we will leak
13:41 < smcv> andrunko: welcome to reference counting. I personally think a 
              leak is less bad than objects disappearing before they ought to
13:41 < smcv> and it's a bug either way

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.