Re: [compiz] Re: Beryl Merger
Hello, Let me please enter the debate about compiz-extra. It's me who named and packaged the whole plugins beryl under compiz. I am not responsible of plugins development, I only made an aggregate work of Mike Dransfield. For me It was relevant to offer a packaging and a simple installation for the users. I took this name of compiz-extra because I thought that the name compiz-community or some thing similar, was too polemical, at the moment of its creation. My principal problem currently i don't know enough compiz development and I miss some time. I was not able to maintain a version for last GIT. I think, that with the recent merge between Beryl/Compiz, the tarball itself becomes rather problematic. The Community idea of the plugins, in my opinion, would be to provide the plugins one by one, thus making possible to the user to install a particular plugin, like themes or desktop applets. This way it's also possible to maintain plugins separately. I currently work on gnome-compiz-manager to propose a gui interface of binary plugins installation, one by one, like that apt or yum proposed. The major problem i found currently is the plugins version management which mainly depends on the ABI VERSION of compiz. Until a few days, I'll propose some concrete things. To finish, and from what I know from the compiz-extra maintenance, I think that risk to maintain a compiz-extra tarball (or under its new name), will bring more problem, than to propose the plugins one by one. The problems that I saw occurring for compiz-extra, are: - plugins version synchronisation in each compiz-extra version - management of the deprecated plugins - selection criteria of inclusion or exclusion of one or more plugins - and the lack of time . Well, that was my 2 cents contribution. Nicolas -- gandalfn [EMAIL PROTECTED] ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] gtk-window-decorator segfault
I have same issue and after apply latest commit, gtk-window-decorator now hangs. When i attach with gdb, i have this backtrace : #0 0xb7f27410 in __kernel_vsyscall () #1 0xb75ae6bb in poll () from /lib/tls/i686/cmov/libc.so.6 #2 0xb773ef19 in _XWaitForReadable (dpy=0x806d300) at ../../src/XlibInt.c:500 #3 0xb773f2ff in _XRead (dpy=0x806d300, data=0xbfa4531c \034\033[\234 , size=32) at ../../src/XlibInt.c:1089 #4 0xb773fcd4 in _XReply (dpy=0x806d300, rep=0xbfa4531c, extra=0, discard=0) at ../../src/XlibInt.c:1716 #5 0xb7721176 in XGetWindowProperty (dpy=0x806d300, w=93, property=351, offset=0, length=2147483647, delete=0, req_type=33, actual_type=0xbfa453a8, actual_format=0xbfa453a4, nitems=0xbfa453a0, bytesafter=0xbfa4539c, prop=0xbfa45398) at ../../src/GetProp.c:64 #6 0xb7efd163 in _wnck_get_window_list (xwindow=93, atom=351, windows=0xbfa454b8, len=0xbfa454b4) at xutils.c:399 #7 0xb7ee914c in do_update_now (screen=0x80851c0) at screen.c:1186 #8 0xb7ee9fc8 in update_idle (data=0x80851c0) at screen.c:1936 #9 0xb7655ee1 in g_idle_dispatch (source=0x810f570, callback=0x, user_data=0x80851c0) at gmain.c:3928 #10 0xb7657c42 in IA__g_main_context_dispatch (context=0x807bb78) at gmain.c:2045 #11 0xb765ac1f in g_main_context_iterate (context=0x807bb78, block=1, dispatch=1, self=0x8095dc8) at gmain.c:2677 #12 0xb765afc9 in IA__g_main_loop_run (loop=0x8109260) at gmain.c:2881 #13 0xb7b9bf84 in IA__gtk_main () at gtkmain.c:1171 It's seem gtk-window-decorator hang when libwnck get _NET_CLIENT_LIST_STACKING window property. Nicolas -- gandalfn [EMAIL PROTECTED] signature.asc Description: This is a digitally signed message part ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] gtk-window-decorator segfault
I revert on compiz-0.3.6 with i have same problem, to catch gdk_x_error backtrace and valgrind, i hope that can you help. Nicolas -- gandalfn [EMAIL PROTECTED] ==7570== Memcheck, a memory error detector. ==7570== Copyright (C) 2002-2006, and GNU GPL'd, by Julian Seward et al. ==7570== Using LibVEX rev 1658, a library for dynamic binary translation. ==7570== Copyright (C) 2004-2006, and GNU GPL'd, by OpenWorks LLP. ==7570== Using valgrind-3.2.1-Debian, a dynamic binary instrumentation framework. ==7570== Copyright (C) 2000-2006, and GNU GPL'd, by Julian Seward et al. ==7570== For more details, rerun with: -v ==7570== ==7570== Invalid read of size 4 ==7570==at 0x4014757: (within /lib/ld-2.5.so) ==7570==by 0x4887C2C: (within /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x48882AB: (within /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x4887B63: dlopen (in /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x47C382E: open_object (XlcDL.c:334) ==7570==by 0x47C4523: _XlcDynamicLoad (XlcDL.c:423) ==7570==by 0x47D8D42: _XOpenLC (lcWrap.c:296) ==7570==by 0x47D8E82: _XlcCurrentLC (lcWrap.c:357) ==7570==by 0x47D9146: XSupportsLocale (lcWrap.c:101) ==7570==by 0x45379D2: _gdk_x11_initialize_locale (gdkim-x11.c:65) ==7570==by 0x453C806: _gdk_windowing_init (gdkmain-x11.c:99) ==7570== Address 0x4B503EC is 76 bytes inside a block of size 77 alloc'd ==7570==at 0x4021380: malloc (vg_replace_malloc.c:149) ==7570==by 0x400D0A2: (within /lib/ld-2.5.so) ==7570==by 0x4008164: (within /lib/ld-2.5.so) ==7570==by 0x4010D74: (within /lib/ld-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x40108C8: (within /lib/ld-2.5.so) ==7570==by 0x4887C2C: (within /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x48882AB: (within /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x4887B63: dlopen (in /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x47C382E: open_object (XlcDL.c:334) ==7570==by 0x47C4523: _XlcDynamicLoad (XlcDL.c:423) ==7570== ==7570== Conditional jump or move depends on uninitialised value(s) ==7570==at 0x4014761: (within /lib/ld-2.5.so) ==7570==by 0x4010AB0: (within /lib/ld-2.5.so) ==7570==by 0x4887C2C: (within /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x48882AB: (within /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x4887B63: dlopen (in /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x47C382E: open_object (XlcDL.c:334) ==7570==by 0x47C4523: _XlcDynamicLoad (XlcDL.c:423) ==7570==by 0x47D8D42: _XOpenLC (lcWrap.c:296) ==7570==by 0x47D8E82: _XlcCurrentLC (lcWrap.c:357) ==7570==by 0x47D9146: XSupportsLocale (lcWrap.c:101) ==7570==by 0x45379D2: _gdk_x11_initialize_locale (gdkim-x11.c:65) ==7570== ==7570== Conditional jump or move depends on uninitialised value(s) ==7570==at 0x401476C: (within /lib/ld-2.5.so) ==7570==by 0x4010AB0: (within /lib/ld-2.5.so) ==7570==by 0x4887C2C: (within /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x48882AB: (within /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x4887B63: dlopen (in /lib/tls/i686/cmov/libdl-2.5.so) ==7570==by 0x47C382E: open_object (XlcDL.c:334) ==7570==by 0x47C4523: _XlcDynamicLoad (XlcDL.c:423) ==7570==by 0x47D8D42: _XOpenLC (lcWrap.c:296) ==7570==by 0x47D8E82: _XlcCurrentLC (lcWrap.c:357) ==7570==by 0x47D9146: XSupportsLocale (lcWrap.c:101) ==7570==by 0x45379D2: _gdk_x11_initialize_locale (gdkim-x11.c:65) ==7570== ==7570== Invalid read of size 4 ==7570==at 0x4014729: (within /lib/ld-2.5.so) ==7570==by 0x4005B69: (within /lib/ld-2.5.so) ==7570==by 0x4007995: (within /lib/ld-2.5.so) ==7570==by 0x4010D74: (within /lib/ld-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x40108C8: (within /lib/ld-2.5.so) ==7570==by 0x4A24F01: (within /lib/tls/i686/cmov/libc-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x4A250B4: __libc_dlopen_mode (in /lib/tls/i686/cmov/libc-2.5.so) ==7570==by 0x4A002CE: __nss_lookup_function (in /lib/tls/i686/cmov/libc-2.5.so) ==7570==by 0x4A003BF: (within /lib/tls/i686/cmov/libc-2.5.so) ==7570==by 0x4A02125: __nss_passwd_lookup (in /lib/tls/i686/cmov/libc-2.5.so) ==7570== Address 0x4B61F7C is 36 bytes inside a block of size 38 alloc'd ==7570==at 0x4021380: malloc (vg_replace_malloc.c:149) ==7570==by 0x4007F33: (within /lib/ld-2.5.so) ==7570==by 0x4010D74: (within /lib/ld-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x40108C8: (within /lib/ld-2.5.so) ==7570==by 0x4A24F01: (within /lib/tls/i686/cmov/libc-2.5.so) ==7570==by 0x400CF95: (within /lib/ld-2.5.so) ==7570==by 0x4A250B4: __libc_dlopen_mode (in /lib/tls/i686/cmov/libc-2.5.so) ==7570==by 0x4A002CE
Re: [compiz] keybindings patch, gnome-keybinding-bridge and vignettes plugins
Please do that unless you plan do update the patch again before recommend that I include it. here workspaces keybindings patch, you were faster for terminal patch ;) I'm a bit concerned about a plugin moving a window like that when it's minimized. It can definitely cause viewport switching issues. The input-transform branch that will be available soon is going to provide a proper way to do this. Sure,vignettes plugin isn't really ready to be to use seriously, it's only a beginning. I have already removed the hack for minimized windows and await patiently for the input-transform. In this moment, I play with so that it functions with the Xembed protocol, I have a small question in addition on this subject. It's normal that I can intercept only ClientMessage Xembed when they are sent in broadcast ? Perhaps it's a mistake, wanting to intercept ClientMessage Xembed in compiz ? - David #0; Nicolas -- gandalfn [EMAIL PROTECTED] commit 22ba938c602661ce7cfb7123c88ea5f7ce3a9a41 Author: Nicolas Bruguier [EMAIL PROTECTED] Date: Wed Feb 14 00:18:52 2007 +0100 Add workspaces keybindings diff --git a/include/compiz.h b/include/compiz.h index cee019d..fb88780 100644 --- a/include/compiz.h +++ b/include/compiz.h @@ -546,7 +546,39 @@ typedef int CompFileWatchHandle; #define COMP_DISPLAY_OPTION_TERMINAL 56 #define COMP_DISPLAY_OPTION_RUN_TERMINAL 57 #define COMP_DISPLAY_OPTION_PING_DELAY 58 -#define COMP_DISPLAY_OPTION_NUM 59 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_159 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_260 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_361 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_462 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_563 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_664 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_765 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_866 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_967 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_10 68 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_11 69 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_12 70 +#define COMP_DISPLAY_OPTION_SWITCH_TO_LEFT_DESKTOP 71 +#define COMP_DISPLAY_OPTION_SWITCH_TO_RIGHT_DESKTOP72 +#define COMP_DISPLAY_OPTION_SWITCH_TO_UP_DESKTOP 73 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DOWN_DESKTOP 74 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_1_WINDOW 75 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_2_WINDOW 76 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_3_WINDOW 77 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_4_WINDOW 78 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_5_WINDOW 79 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_6_WINDOW 80 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_7_WINDOW 81 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_8_WINDOW 82 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_9_WINDOW 83 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_10_WINDOW84 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_11_WINDOW85 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DESKTOP_12_WINDOW86 +#define COMP_DISPLAY_OPTION_SWITCH_TO_LEFT_DESKTOP_WINDOW 87 +#define COMP_DISPLAY_OPTION_SWITCH_TO_RIGHT_DESKTOP_WINDOW 88 +#define COMP_DISPLAY_OPTION_SWITCH_TO_UP_DESKTOP_WINDOW89 +#define COMP_DISPLAY_OPTION_SWITCH_TO_DOWN_DESKTOP_WINDOW 90 +#define COMP_DISPLAY_OPTION_NUM91 typedef CompOption *(*GetDisplayOptionsProc) (CompDisplay *display, int *count); @@ -2145,6 +2177,13 @@ getWindowProp32 (CompDisplay *display, Atom property, unsigned short defaultValue); +Bool +getWindowPropList (CompDisplay *display, + Window id, + Atom property, + unsigned long **list, + unsigned long *nb); + void setWindowProp32 (CompDisplay*display, Window id, diff --git a/src/display.c b/src/display.c index c20246d..5565744 100644 --- a/src/display.c +++ b/src/display.c @@ -46,6 +46,13 @@ #include compiz.h +typedef enum _CompDisplayDirection { + COMP_DISPLAY_DIRECTION_LEFT, + COMP_DISPLAY_DIRECTION_RIGHT, + COMP_DISPLAY_DIRECTION_UP, + COMP_DISPLAY_DIRECTION_DOWN +} CompDisplayDirection; + static unsigned int virtualModMask[] = { CompAltMask, CompMetaMask, CompSuperMask, CompHyperMask, CompModeSwitchMask, CompNumLockMask, CompScrollLockMask @@ -219,6 +226,85 @@ freeDisplayPrivateIndex (int index) } static Bool +getDesktopLayout(CompDisplay *d, + Window root, + unsigned int *cols, + unsigned int *rows) +{ + unsigned long *list, nb; + Atom net_desktop_layout; + + net_desktop_layout = XInternAtom (d-display, _NET_DESKTOP_LAYOUT, 0); + + if (getWindowPropList(d, root, net_desktop_layout, list, nb)) + { + if (nb = 3) + { + *cols = (unsigned int)list[list[0] ? 2 : 1]; + *rows = (unsigned
Re: [compiz] keybindings patch, gnome-keybinding-bridge and vignettes plugins
It doesn't compile with latest Compiz from GIT. And if I even get it to compile, it just crashes. :( Stjepan #0; Yes sorry I made it for compiz-0.3.6, i updated it to work also with latest git : http://gandalfn.club.fr/ubuntu/vignettes/vignettes-applets-0.0.2.tar.gz gandalfn -- gandalfn [EMAIL PROTECTED] signature.asc Description: This is a digitally signed message part ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] keybindings patch, gnome-keybinding-bridge and vignettes plugins
The vignettes plugin doesn't do anything all alone. It needs an external application, here “vignettes-applet” which controls it via dbus. To add applet, you should build and install it like this : ./configure --prefix=/usr --sysconfdir=/etc make sudo make install and then replace GNOME Window List applet by “Vignettes Window List”. If you have some trouble to add applet, try to kill bonobo-activation-server and gconfd-2 before add applet. gandalfn -- gandalfn [EMAIL PROTECTED] signature.asc Description: This is a digitally signed message part ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] Irritating reloading
It's due in fact of session manager restart style, which always restart compiz. Here a small patch which fix this. On dim, 2007-01-28 at 15:45 +, Mike Dransfield wrote: Stjepan Glavina wrote: Compiz reloads itself if it crashes or another app kills it. This can be sometimes very irritating and annoying. Can there at least be a command-line option to disable this? Are you sure? I dont think it does Maybe you have the crashhandler plugin enabled or something else is autorestarting it. ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz #0; commit c80b1ed9b651cc16eeed951edefc54af08992a5f Author: Nicolas Bruguier [EMAIL PROTECTED] Date: Sun Jan 28 16:51:57 2007 +0100 Restart only if compiz running diff --git a/src/session.c b/src/session.c index 6f339be..b5a3800 100644 --- a/src/session.c +++ b/src/session.c @@ -93,7 +93,7 @@ setRestartStyle (SmcConn connection) { SmProp prop, *pProp; SmPropValue propVal; -charhint = SmRestartImmediately; +charhint = SmRestartIfRunning; prop.name = SmRestartStyleHint; prop.type = SmCARD8; signature.asc Description: This is a digitally signed message part ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
[compiz] keybindings patch, gnome-keybinding-bridge and vignettes plugins
Hi folks, For ubuntu specs and for a better desktop integration of compiz, I made some patchs and plugins. Focus prevention is now in main stream, and really better can i made :) First the workspace and terminal keybinding patch. Since my last email I updated the patch. http://gandalfn.club.fr/ubuntu/compiz-patch/91-workspaces-terminal-keybindings.patch It adds the support for vertical workspaces. If you want i re-split them, say me. You can found all patchs, which included in ubuntu feisty package here : http://gandalfn.club.fr/ubuntu/index.php?dir=compiz-patch%2F Some of them are only for ubuntu. I also modified de main stream restart patch, to restart compiz only if it's the current window manager : https://launchpad.net/ubuntu/+source/compiz/+bug/80764 Now gnome-keybinding-bridge plugin is avaible on git.compiz.org repos (Thanks David). You get it like this : git clone http://git.compiz.org/compiz/plugins/gnome-keybinding-bridge If you have improvements or patchs for it, you can post them here, it's with pleasure i'll integrate them. To finish, I also have made a small plugin: “vignettes”, based on miniwin and switcher plugin. http://gandalfn.club.fr/ubuntu/vignettes/vignettes-0.0.1.tar.gz This plugin can create clone window and destroy them with dbus. You can create a mini clone window like this: dbus-send --type=method_call --dest=org.freedesktop.compiz \ /org/freedesktop/compiz/vignettes/allscreens/create \ org.freedesktop.compiz.activate :window int32:window xid to clone \ :x int32:optional clone window x position \ :y int32:optional clone window y position \ :width int32:optional clone window width position \ :height int32:optional clone window height position you can close the clone window like this : dbus-send --type=method_call --dest=org.freedesktop.compiz \ /org/freedesktop/compiz/vignettes/allscreens/release \ org.freedesktop.compiz.activate :window int32:cloned window xid To illustrate the plugin use, I hacked GNOME applet window list, to have an thumbnail plugin equivalent. http://gandalfn.club.fr/ubuntu/vignettes/vignettes-applets-0.0.1.tar.gz Vignettes plugin manages also minimized windows, instead save the thumbnail in the core, I map simply the original window outer the visible space during clone windows is showed. You can see here a small video to see vignettes-applet in action: http://gandalfn.club.fr/ubuntu/vignettes/vignettes-applets-demo.ogg I made this plugin and applet only to demonstrate dbus interaction with compiz and without any modification in the core, not for replace thumbnail plugin ;). Mike Dransfield also made some interesting stuff on this subject : http://www.go-compiz.org/index.php?title=Scripting_Compiz Have fun gandalfn -- gandalfn [EMAIL PROTECTED] signature.asc Description: This is a digitally signed message part ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] ubuntu packages patch
http://gandalfn.club.fr/ubuntu/compiz-patch/90-fix-no-border-window-shadow.patch This patch fix white boxes instead of shadows for gnome-panel and other non-standard windows on AiGLX. That patch is funny. It would be nice to track that issue down. Have you checked that it's not just a width that is power-of-two that's causing problems? While testing with gdb it would seem that the phenomenon appears only when the width is a modulo of 32 (perhaps a relation with the depth) , I think it's perhaps a AIGLX xorg bug, unfortunately, I could not determine from where that came. http://gandalfn.club.fr/ubuntu/compiz-patch/91-workspaces-terminal-keybindings.patch: This patch add workspaces keybindings and launch terminal keybinding. I'd like to include this patch. Can you separate the terminal and desktop bindings into different patches? No problemo : http://gandalfn.club.fr/ubuntu/compiz-patch/add-workspaces-keybindings.patch http://gandalfn.club.fr/ubuntu/compiz-patch/add-terminal-keybinding.patch But I would like notify, calcDesktopIndex function supports only horizontal workspaces orientation, first element of the _NET_DESKTOP_LAYOUT atom. It should be improved to support vertical workspaces orientation. btw, we should move the viewport changing actions to the core as well. http://gandalfn.club.fr/ubuntu/compiz-patch/92-gnome-bridge-keybindings-plugin.patch: This patch synchronize metacity and compiz keybindings. It’s add a new keybinding plugin which when compiz is started update its keybindings with those of metacity, and when compiz is launched, it’ll update compiz keybindings when gnome keybindings are updated and vice versa. You can add full gnome keybindings support with this control center patch : http://gandalfn.club.fr/ubuntu/gnome-patch/36-compiz-keybindings-support.patch The keybinding bridge plugin is useful, do you want me to create a git repo on git.compiz.org for it? Why not :) it still needs some improvements and it would be the best mean to make them http://gandalfn.club.fr/ubuntu/compiz-patch/93-focus-user-event.patch: This patch allow the active window kept the focus, if user write in it, when a new window is open, like metacity do. We should add full focus prevention support to compiz but I guess this might be the most useful part of it so I'm OK with including it now. I am not sure it's ready for the upstream. It gives the focus in random way in some cases, it's seem the alone use of UserEvent to determine focus isn't best solution, I seek other conditions to give it more functional. http://gandalfn.club.fr/ubuntu/compiz-patch/94-fix-show-managed-window.patch: This patch add in client list window all windows in workspaces, it's fix window show in gnome applet workspace and in gnome applet tasklist. I don't think I understand exactly what this patch fixes. Could you describe a use case which is causing problems without this patch applied? With gnome task applet or workspace applet, all windows which aren't in active workspace are hidden except minimized window, it's due in fact that libwnck get stacked window list for draw or enumerate all windows in all workspace. When you switch between workspace, say me if wrong, you map all windows in current workscape and you unmap all others. On this unmap you set window state to CompWindowStateHiddenMask and wmstate to IconicState, if you test only state, window stacked list doesn't include all windows with Hidden state but in these windows you have all windows in others workspaces and then you don't show them in workspace applet or tasklist applet. If you include IconicState windows you kept them in stacked list. I assume that you're the author of all these patches and you're OK with me including the ones I'm interested in upstream. Let me know if that's not the case. Yes, I'm author of these patches, you can use them how do you want ;) i'm sure you improve them better can i do. Otherwise I include a fedora patch in ubuntu deb which restart compiz on crash using session manager, in this case i don't know author but it's work very well: http://gandalfn.club.fr/ubuntu/compiz-patch/06-fedora-restart-on-crash.patch Thanks, -David #0; Nicolas -- gandalfn [EMAIL PROTECTED] signature.asc Description: This is a digitally signed message part ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] ubuntu packages patch
hi sorry, i made a mistake in 94-fix-show-managed-window.patch, i updated it. On sam, 2007-01-06 at 14:01 +0100, gandalfn wrote: hi, For compiz ubuntu packages, I made some patchs to have a behavior of compiz nearer to metacity. All patch are made for compiz-0.3.6 and they are more hacks or draft than good fix, i hope they can help. http://gandalfn.club.fr/ubuntu/compiz-patch/90-fix-no-border-window-shadow.patch This patch fix white boxes instead of shadows for gnome-panel and other non-standard windows on AiGLX. http://gandalfn.club.fr/ubuntu/compiz-patch/91-workspaces-terminal-keybindings.patch: This patch add workspaces keybindings and launch terminal keybinding. http://gandalfn.club.fr/ubuntu/compiz-patch/92-gnome-bridge-keybindings-plugin.patch: This patch synchronize metacity and compiz keybindings. It’s add a new keybinding plugin which when compiz is started update its keybindings with those of metacity, and when compiz is launched, it’ll update compiz keybindings when gnome keybindings are updated and vice versa. You can add full gnome keybindings support with this control center patch : http://gandalfn.club.fr/ubuntu/gnome-patch/36-compiz-keybindings-support.patch http://gandalfn.club.fr/ubuntu/compiz-patch/93-focus-user-event.patch: This patch allow the active window kept the focus, if user write in it, when a new window is open, like metacity do. http://gandalfn.club.fr/ubuntu/compiz-patch/94-fix-show-managed-window.patch: This patch add in client list window all windows in workspaces, it's fix window show in gnome applet workspace and in gnome applet tasklist. ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz -- gandalfn [EMAIL PROTECTED] signature.asc Description: This is a digitally signed message part ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] Odd GTK-Window-Decorator bug
Hi, I can confirm this issue it's seem no border window shadow isn't drawing correctly in latest git, perhaps only on AiGLX. I made a little patch to fix it. Will, it's included in lastest git packages. gandalfn Le mardi 05 décembre 2006 à 15:54 -0500, Will Farrington a écrit : In the latest git compilation by gandalfn for his repositories, I receive an odd error regarding non-window items; but only under gtk-window-decorator. All non-window items (including panels, drop-down menus, etc.) have about a 10pixel system color stroke effect in place of the drop shadow, or what have you. This is experienced currently under Feisty Fawn testing, afaik. Latest kernel, et al. ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz #0; From ea52555b7a1c6f86d244650b6e74b0311f34822e Mon Sep 17 00:00:00 2001 From: Nicolas Bruguier [EMAIL PROTECTED] Date: Fri, 8 Dec 2006 21:08:07 +0100 Subject: [PATCH] Fix no border shadow --- gtk/window-decorator/gtk-window-decorator.c |8 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/gtk/window-decorator/gtk-window-decorator.c b/gtk/window-decorator/gtk-window-decorator.c index d22467b..45ba31e 100644 --- a/gtk/window-decorator/gtk-window-decorator.c +++ b/gtk/window-decorator/gtk-window-decorator.c @@ -4899,10 +4899,10 @@ update_shadow (void) no_border_shadow = decor_create_shadow (xdisplay, gdk_x11_screen_get_xscreen (screen), 1, 1, - 0, - 0, - 0, - 0, + 1, + 1, + 1, + 1, 0, 0, 0, 0, opt, shadow_context, -- 1.4.1 signature.asc Description: Ceci est une partie de message numériquement signée ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] reason for bad performance on i915 (AIGLX)
Hi, In my ubuntu packages, to compiz work properly on AiGLX and for performance issue i continue to apply two patchs of Kristian Høgsberg : http://people.freedesktop.org/~krh/compiz-on-aiglx/compiz-patches/06-glfinish.patch http://people.freedesktop.org/~krh/compiz-on-aiglx/compiz-patches/02-tfp-server-extension.patch It's perhaps a mistake but they resolve many issues on AIGLX gandalfn Le jeudi 30 novembre 2006 à 08:42 +0100, Mirco Müller a écrit : Greetings everybody! For about two weeks I'm now a happy owner of an intel-powered laptop (chipset, wifi, gfx) and wanted to give it a full test-drive of compiz and other GL-intensive programs. The laptop runs Xorg 7.1.1 and the i810 driver that comes with it (stock rpm from FC6). I'm forced to start compiz 0.3.4 with these options... --use-cow --strict-binding --indirect-rendering ... to merely get compiz to work. I cannot leave out any of those options, otherwise I don't see anything rendered by compiz. While I'm a bit spoiled regarding performance coming from my desktop-system (nvidia-driven, using nvidia's AIGLX), I did not expect the i915 (AIGLX) solution to be such a bad performer. At a resolution of 1280x800 I can barely use the cube-rotatation, switcher or drag-rotate while moving windows to different workspaces. Furthermore GL-apps leave artifacts on the screen when dragged around (even if ran under metacity) and I don't seem to be able to get transparent GL-apps working (e.g. try gl-cairo-cube from people.freedesktop.org/~macslow). I know that the i915 is an integrated graphics-chip solution using only system-memory, but I remember people showcasing screencaptures of such systems with much better performance under compiz. Is all this only a Xorg/driver problem or will I never get any decent performance (and feature-parity compared to low-end nvidia-hardware) from this laptop when running a composited desktop-environemnt? Also the screensaver seems to be to much for this setup as it only display a black nothing (only the mouse-cursor is visible and I have a hard time entering my password to unlock it again). What disappoints me also is the lack of support for transparent GL-apps under compiz on the i915. I'm not sure if this is a purely a driver issue or partly my fault (not selecting the correct visual in e.g. gl-cairo-cube). All my transparent cairo-hacks work fine. Thanks in advance for your time and kind advice! Best regards... Mirco MacSlow Müller ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
Re: [compiz] Compiling Compiz
I make a point all the same of announcing that there are official compiz packages , (I know for ubuntu), mine :P. http://gandalfn.club.fr/ubuntu Moreover I profit that you evoke the third-party plugins to raise a problem of packaging. I seek to create packages for those but I cannot find which version I put: compiz version ?, ABI compiz version ? I think that it would be important to put a number of version for each so that a modest packager can find himself there. Le mercredi 18 octobre 2006 à 15:20 +0100, Mike Dransfield a écrit : Johan Kotze wrote: Jigish This is the location of the old compiz-quinn packages. They are not updated anymore, now that they have forked into Beryl. I am specifically looking for an rpm of compiz-0.2.0 as was released by David. I think this is a major problem at the moment, there do not appear to be any packagers for the official compiz. I believe that all of them are involved in Beryl and have given up supporting compiz, which is a shame because it reduces the choice of the 'community'. It would be nice if someone could package compiz and then create an additional package with the third-party plugins in. ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz #0; ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz
[compiz] gtk-window-decorator metacity opacity theme
hi, I made a small a patch for gtk-window-decorator which define default border alpha to zero when it use metacity theme. That allows while playing with the alpha attributes of the metacity themes to have transparent borders. here too the modified “Human Ubuntu” theme and a screenshot to show the result. ps: sorry for my bad english --- compiz.orig/gtk/window-decorator/gtk-window-decorator.c 2006-09-30 12:34:36.0 +0200 +++ compiz/gtk/window-decorator/gtk-window-decorator.c 2006-10-01 18:45:08.0 +0200 @@ -2117,7 +2121,7 @@ region = meta_get_window_region (fgeom, clip.width, clip.height); -gdk_cairo_set_source_color (cr, style-bg[GTK_STATE_NORMAL]); +gdk_cairo_set_source_color_alpha (cr, style-bg[GTK_STATE_NORMAL], 0.0f); for (i = 0; i region-numRects; i++) { @@ -2497,6 +2501,7 @@ GdkPixmap *pixmap; GdkVisual *visual; GdkColormap *colormap; +GdkScreen * screen; visual = gdk_visual_get_best_with_depth (32); if (!visual) @@ -2506,15 +2511,13 @@ if (!pixmap) return NULL; -colormap = gdk_colormap_new (visual, FALSE); +screen = gdk_visual_get_screen(visual); + +colormap = gdk_screen_get_rgba_colormap(screen); if (!colormap) -{ - gdk_pixmap_unref (pixmap); - return NULL; -} +colormap = gdk_screen_get_rgb_colormap(screen); gdk_drawable_set_colormap (GDK_DRAWABLE (pixmap), colormap); -gdk_colormap_unref (colormap); return pixmap; } HumanAlpha.tar.bz2 Description: application/bzip-compressed-tar screenshot1.png Description: PNG image ___ compiz mailing list compiz@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/compiz