On Fri, Dec 11, 2009 at 11:11:24AM -0500, John J. Foerch wrote: > On Fri, Dec 11, 2009 at 11:28:54PM +1030, David Kettler wrote: > > >What about having conkeror always use enterLastWindowClosingSurvivalArea, > > >instead of only in daemon mode? Daemon-mode would need to be rewritten to > > >accomodate this, but it would give us a chance to always run quit_hook, no > > >matter in what manner the program was exited. This change would also > > >allow us to fix the problem in sessions.js where sessions do not get > > >written in every circumstance. > > > > > > > Thanks, but I don't see how to make use of that. When that call is > > in effect none of the quit-application observer functions seem to be > > called. > > > > If I use an xpcom-shutdown observer function, that does get called > > for both my cases a and b. It's probably too late in the processing > > for some things one might want to put on the quit_hook though. It > > seems to prevent daemon mode exiting. > > > My line of thinking is an alternative to using quit-application observer. > If conkeror enters a last-window-closing-survival-area at startup, it > would give us complete control over exit behavior all in one central > place. The things which immediately come to mind, that need to be taken > account of exit-time are: > > * daemon-mode > > * walnut summary > > * session writing
Oh, now I see the case I was not considering. If you never open any windows in the first place, as is typical when running walnut, last-window-closing-survival-area does not do us any good. Sorry for the confusion. So to take this forward, what I'm really asking is for a comprehensive design (in english) of what happens at exit-time, in all possible circumstances, and how to accomodate all known cases. -- John Foerch _______________________________________________ Conkeror mailing list [email protected] https://www.mozdev.org/mailman/listinfo/conkeror
