Hi. Thanks, I was looking for that entry-point since the beginning...somehow I managed to completely miss the subfolders with further scripts there.
resize-window-with-hints seems to take not the size in pixels, but in rows and columns which are being calculated for each window (don't what that is, though). But resize-window-with-hints* takes the pixel values. The options dialogs seems also to assume pixels...and that's where they clashed for me, as resize-window-with-hints was called with the pixel values, which screwed the size completely. So basically I just removed the calculations and directly handed the values down to resize-window-with-hints. I'm not sure if this is a nVidia issue, as I was able to reproduce this on my Laptop with an Intel chipset. As it seems, the match-window.jls seems to be the only script/part which calls resize-window-with-hints, everything else calls resize-window-with-hints*...I wonder if that's a typo or a simple remain from previous versions. So, I came up with an even easier fix thanks to you: diff lisp/sawfish/wm/ext/match-window.jl 448c450 < (resize-window-with-hints w (car value) (cdr value)))) --- > (resize-window-with-hints* w (car value) (cdr value)))) Works perfectly, too. If somebody has something to debug this are some further insides, I'll do what I can to help. Thanks so far! On Thu, 05 Jul 2012 13:48:52 +0900 (JST) Teika Kazura <[email protected]> wrote: > On Tue, 3 Jul 2012 20:59:49 +0200, "Robert 'Bobby' Zenz" wrote: > > After not knowing what else to do, I ended up with this...well, > > hack: > > > > diff > > tmp/sawfish/sawfish-1.9.0/lisp/sawfish/wm/windows.jl > > /usr/local/share/sawfish/lisp/sawfish/wm/windows.jl > > 382,384c382,386 < (resize-window-to > > < w (constrain-dimension-to-hints (+ x-base (* x-inc cols)) > > 'x hints) < (constrain-dimension-to-hints (+ y-base (* y-inc > > rows)) 'y hints)))) --- > >> (resize-window-to w cols rows hints) > >> ;;;(resize-window-to > >> ;;; w (constrain-dimension-to-hints (+ x-base (* x-inc > >> cols)) 'x hints) ;;; (constrain-dimension-to-hints (+ y-base (* > >> y-inc rows)) 'y hints)) )) > > > > That fixes the problem completely... > > Hi. Glad to know your problem is gone, but this is not a bug of > Sawfish. Maybe due to rare race conditions? (IIRC there were trouble > reports only for nVidia drivers a couple of years ago.) > > But if the same issue is frequent, it may be better to provide a > workaround option. > > The involved function is resize-window-with-hints, which is only used > from match-window.jl. All other functions use > resize-window-with-hints*. (Notice the last asterisk.) I don't know > what's the difference. Could anyone help? > > Teika (Teika kazura) > > -- Sawfish ML
