On Sun, 23 May 2010 17:26:04 +0200, Christopher Roy Bratusek wrote: > Well, if it does not cause any bugs/regressions or anything, there's > no real reason for rejection
Wait a minute. The objective is to support *one* specific application, "dockapp". (Or one class of applications?) So the code should be stored in *one* file, I think in wm.ext.dockapp, not scattered across several modules. In this case, doc in sawfish.texi is not necessary. Just a news item which says "look at wm/ext/dockapp.jl for the details". (Since there's no option. A 'require' is enough.) And C part does too much. "Withdrawn & iconic must be dockapp." crosses the line, and can affect other people. Isn't it possible to detect one with window name or class, from before / after-add-window-hook? If necessary, implement some mechanism which allows wm.ext.dockapp to react when a dockapp appears. C code should be general enough. Or, if you can't come up with a subsitution, at least check a new user option 'detect-dockapp' in that C detection code. ICCCM and ewmh defines window manager standards. ICCCM section 4 speaks a lot on withdrawn state, but I have never read it, so I can't ask you to understand it. ;) You may wonder how about mark-window-as-dock/desktop. In fact they're defined in ewmh, so common enough. By the way, Chris, you probably prefer these patches sent in one email, like me? It will be more readable, easier to apply, and they make sense only when put together. (And people with slower traffic can ignore big mails.) But this must break the night for dockapp users. Thanks a lot! Teika (Teika kazura)
