http://bugzilla.moblin.org/show_bug.cgi?id=7711 --- Comment #1 from yongsheng zhu <yongsheng.zhu(a)intel.com&gt; 2010-01-06 01:41:42 PST --- the logic I think should be: 1) If any client attaches or any connection is active, do not any of checkings for termination. 2) Otherwise, add a timer to track duration in glib main loop if it is not unlimited. After a method call occurs, the timer is reset. Once the timer is timeout, syncevo-dbus-server is terminated. To track method calls, I'd like to add the checking in the gate of each dbus call('handle_message' in gdbus.c) but not know whether it is suitable. 3) The command line option is '--duration xxx'. 'unlimited' - run forever 'xxx' - a number means a given duration in seconds -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching someone on the CC list of the bug.

http://bugzilla.moblin.org/show_bug.cgi?id=7711 --- Comment #3 from yongsheng zhu <yongsheng.zhu(a)intel.com&gt; 2010-01-17 20:06:03 PST --- A callback is per-interface not the global one to re-use the gdbus InterfaceData mechanism and the callback is stored in interface data for each interface. APIs are not changed, but newly added. -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching someone on the CC list of the bug.

http://bugzilla.moblin.org/show_bug.cgi?id=7711 --- Comment #5 from pohly <patrick.ohly(a)intel.com&gt; 2010-01-18 01:37:40 PST --- (In reply to comment #4) Yes, looks good to me. -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching someone on the CC list of the bug.

http://bugzilla.moblin.org/show_bug.cgi?id=7711 --- Comment #7 from pohly <patrick.ohly(a)intel.com&gt; 2010-01-22 08:11:33 PST --- gdbus: add interface-level callback function Should mention that we want to use this for tracking whether the server is still in use. GDBus C++ binding: changes according to gdbus changes When reading this summary by itself it is not clear what these changes are. The code itself adds a C-style callback. DBusObjectHelper is meant to be more C++ like. Please use a Boost function object instead of the "void *" + function pointer. That'll allow you to turn AutoTerm::checkCallback() into a normal member function in the next patch. DBus server: automatic termination (MB#7711) [...] Allow one dbus client attaching or detaching more than one time. Each will be treated as only one. Also we detect the client gone without detaching itself. This is not clear to me. Does it mean that each attach has to be matched by one detach, or does "attach + attach + detach" detach the client? According to the code, it is the latter. I think we should implement the former, because attach/detach might be called in different, independent modules of the caller. It would be unexpected if one module was allowed to detach the whole process. Our documentation is silent on this. Please add a comment there. AutoTerm::checkCallback() is not documented. A short comment from where it is called would help to understand the code. + at->m_elapsed += TERM_INTERVAL; You assume that checkCallback() will be called in a timely fashion, while in reality it will be called after TERM_INTERVAL + delta, with delta being large if the D-Bus server is busy. Therefore m_elapsed will underestimate the real elapsed time. But because this only happens in a busy server and a busy server doesn't have to auto-terminate, this code probably really works. A comment explaining this reasoning would be useful. + static GDBusInterfaceFunction getCallback() Callback? Which callback? ;-) Please choose a better name. Please put the code into a feature branch mb7711 and remove it from your personal branch, so that I can merge it more easily. -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching someone on the CC list of the bug.

http://bugzilla.moblin.org/show_bug.cgi?id=7711 pohly <patrick.ohly(a)intel.com&gt; changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #9 from pohly <patrick.ohly(a)intel.com&gt; 2010-01-26 00:41:40 PST --- (In reply to comment #8) Merged, plus some minor language fixes. -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching someone on the CC list of the bug.