Hi, > Can you please give precise reproduction instructions -- you don't > even tell us what the window manager is. > I'd be willing to consider a workaround, but only if I can reproduce > the problem (which appears to be a false assumption elsewhere in your > setup).
Yes, sorry, it is one of my first bug report so please forgive me for being incomplete. The window manager I use is dwm, which can be found at : http://www.suckless.org/wiki/dwm It is a lightweight tiled window manager with configuration by compilation. You can change settings by modifying the config.h file and recompiling dwm.c. The setting which cause problem in my config.h is the following : Rule rules[] =3D { /* class instance title tags ref isf= loating */ { NULL, NULL, "R Graphics", NULL, True}, which makes every window whose title match the =C2=ABR Graphics=C2=BB subst= ring to be considered as a floating window. If you use dwm with this configuration rule, launch R and open a graphics window, it will not be set as floating but will remain tiled (which is the default). The problem is that dwm applies the rule given above as soon as the window is mapped. But in R 2.7.0, when the window is mapped she doesn't have any title (it is assigned later), so the rule doesn't apply because the window is considered having no instance name, no class name (the WM_CLASS attribute) and no title. > Which is perfectly valid in X11, so I don't see where the 'bug' is. > > R changes the title to reflect the window state, do the supplied > title' is never actually used without decoration. R 2.6.x put up one > title and then immediately changed it, but that was undocumented and > not safe to rely on (as you found). You're perfectly right by saying that it is not a bug per se, it is just a small change which shouldn't affect the vast majority of users but those who rely on windows attributes to manage them. But maybe I am the only one in that case... Another workaround could be to give a class or instance name specific to the graphics windows. Thanks for your answer and sorry for the incomplete bug report. Regards, --=20 Julien Barnier Groupe de Recherche sur la Socialisation ENS Lettres et Sciences humaines, Lyon T=C3=A9l : 04 37 37 62 92 ______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
