On Mon, Feb 29, 2016 at 10:45 AM, Tom Hacohen <t...@stosb.com> wrote: > Yes, I already told him that on IRC. The cost of caching is most likely > higher than the cost of calculating every time, and anyway, it's rare for > an object to need to have information about the mainloop, adding a cache on > every object is a bad idea, calculating on creation is the way to go.
Yes, we can figure that out when the problem arise, which I doubt it will. > On Mon, Feb 29, 2016 at 6:38 PM, Cedric BAIL <cedric.b...@free.fr> wrote: >> Hello, >> >> On Fri, Feb 26, 2016 at 4:32 PM, Carsten Haitzler <ras...@rasterman.com> >> wrote: >> > we were discussing this and i don't remember an adequate solution. >> > >> > so we're needing to allow multiple event loops (main loop). >> > >> > ok = first loop created == main loop. fine. store that in a global. >> > >> > but after this... what loop does an object belong to? ok - timers: >> > >> > create timer. set parent to the loop it belongs to. that works. simple. >> what >> > about more complex things? a whole widget tree with some edje or button >> object >> > deep down. their parent != loop. sure you can walk the tree to a toplevel >> > window and window has to belong to a loop. >> > >> > i remember tls being discussed. we store a loop created in the tls for >> the >> > thread. this requires that all objects be created in the thread they >> belong to. >> > they all just get their loop from the tls storage. >> > >> > was this the solution? then that is the limitation - must create in the >> thread >> > you belong to. that's not a great limitation... unless we can "fix it" >> eg by >> > creating in one place and transferring to another loop. how? what api? it >> > likely has to be universal or almost (an iface every object that needs a >> loop >> > would inherit?) >> > >> > basically -> i don't remember a nice solution to this. did we have one? >> >> We concluded it wasn't a problem as only at object creation it would >> likely need to cache that information. So overriding eo_parent_set + >> walking the tree of the new parent to find the mainloop seem to be a >> sufficiently efficient way to do it and if not we can always add a >> special function later to get the main loop on object type where it >> matters. >> -- >> Cedric BAIL >> >> >> ------------------------------------------------------------------------------ >> Site24x7 APM Insight: Get Deep Visibility into Application Performance >> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month >> Monitor end-to-end web transactions and take corrective actions now >> Troubleshoot faster and improve end-user experience. Signup Now! >> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 >> _______________________________________________ >> enlightenment-devel mailing list >> enlightenment-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel >> > ------------------------------------------------------------------------------ > Site24x7 APM Insight: Get Deep Visibility into Application Performance > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month > Monitor end-to-end web transactions and take corrective actions now > Troubleshoot faster and improve end-user experience. Signup Now! > http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 > _______________________________________________ > enlightenment-devel mailing list > enlightenment-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel > -- Cedric BAIL ------------------------------------------------------------------------------ Site24x7 APM Insight: Get Deep Visibility into Application Performance APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month Monitor end-to-end web transactions and take corrective actions now Troubleshoot faster and improve end-user experience. Signup Now! http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 _______________________________________________ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
