On 01/10/15 02:59 PM, Jasper St. Pierre wrote: > For the present case, present_with_event will support focus-stealing > prevention if the serial is garbage or not up to date, whereas present > should always steal focus.
I thought the intent was that "present" would *never* steal focus, and that present_from_event could steal focus if certain conditions were met. I don't think we want the final decision to be up to the application... > The garbage serial and don't-have-a-serial cases are different, so I > believe we should have two different methods in that case for > different behavior. > What do you mean by a "garbage serial"? Just one that's really old? Or the application just made up some random numbers to be anti-social? > On Thu, Oct 1, 2015 at 12:52 PM, Bill Spitzak <spit...@gmail.com> wrote: >> The no-serial value really really really needs to be zero. Otherwise you get >> all kinds of annoyances with many language wrappers which map zero to false >> and all other values to true. I think most will have to resort to xor'ing >> the value with the no-serial value, meaning the numerical values are all >> different. >> >> >> >> On Wed, Sep 30, 2015 at 1:23 PM, Jasper St. Pierre <jstpie...@mecheye.net> >> wrote: >>> >>> >>> In cases where we have two behaviors for serial-aware and >>> non-serial-aware operations, I would rather have two different client >>> requests. >> >> >> Absolutely wrong. One of the main reasons for the no-serial value is so that >> it is encouraged that unrecognized serial numbers act exactly like the >> no-serial case. If sending garbage to the serial-taking message has some >> useful result that is not achieved by the no-serial version, then clients >> are going to do this. Having a no-serial value avoids this design mistake. >> >> >> > > > _______________________________________________ wayland-devel mailing list wayland-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/wayland-devel