Re: Proposed change to desktop-entry-spec: make StartupWMClass a "string(s)" field

2015-02-25 Thread David Faure
On Tuesday 27 January 2015 14:10:25 Jasper St. Pierre wrote: > Hi Andy, > > StartupWMClass usage should be considered legacy. In order to compute a > mapping from window class to .desktop file, we need to read *all* files > from disk, which is super expensive to do first thing on boot. But the sa

Re: Proposed change to desktop-entry-spec: make StartupWMClass a "string(s)" field

2015-01-27 Thread Jasper St. Pierre
Hi Andy, StartupWMClass usage should be considered legacy. In order to compute a mapping from window class to .desktop file, we need to read *all* files from disk, which is super expensive to do first thing on boot. I wouldn't be too happy modifying it so that we have to do *more* work to make it

Re: Proposed change to desktop-entry-spec: make StartupWMClass a "string(s)" field

2015-01-27 Thread Andy
Hi, I'm just following up with some more use cases. The KDE System Settings can be launched as a monolithic item, or alternatively each module can be launched independently. The full application window's class is "systemsettings" while the components' are "kcmshell4" -- both of these should be gro

Proposed change to desktop-entry-spec: make StartupWMClass a "string(s)" field

2015-01-20 Thread Andy
Hi, Some applications make use of multiple windows of different types, with different values for StartupWMClass. If the desktop entry spec supported a semicolon-delimited list of values for the StartupWMClass field, then libraries like bamf could make use of that information to better match window