mate-desktop / marco Goto Github PK

Hello,

It seems theres a bug in marco with Unity3D engine games:
video to explain the issue visually
https://dl.dropboxusercontent.com/u/28788188/Game-0041.mkv

the problem is that the mouse just keeps moving up when its captured by the game, until it reaches the end of the window, in which case the game freezing.
when using alt+tab to bring it back it starts to slowly move to the top again until next freeze.

moving mouse faster (doesnt matter which direction) moves it to the top just faster.

replaced marco with xfwm4 and the problem was gone.

to replicate the problem:
Try one of the games: Verdun, Doorways, Rust, Legends Of Aetherus, Mirrormoon.

also free game: http://www.desura.com/games/haunt

I compil (with prefix /usr) and install the 1.7.0 branch of mate-window-manager but it doesn't work. When I put a window in right or left, it doesn't resize. But it works when I move to the top (maximize window).

Any solution ?

command-1 is actually present and working in the schema, but the others are not there. This means I can only have 1 global keybinding command.

I have tried adding command-2 in the schema XML, but I still get "No command 2 has been defined.".

Edit: running glib-compile-schemas /usr/share/glib-2.0/schemas after adding command-2 and friends fixed this issue for me, at least temporarily.

This is an issue since 1.6.

How to reproduce:
press PrintScreen/SysRq keyboard button
and you'll see a window
Marco:
"No command 33 has been defined."
in fact, the command is defined
in org.mate.Marco.keybinding-commands.command-screenshot
as 'mate-screenshot'
but, according to log messages
(MARCO_VERBOSE=1)
KEYBINDINGS: Command command-1 has no underscore?
KEYBINDINGS: Command command-screenshot has no underscore?
KEYBINDINGS: Command command-window-screenshot has no underscore?
this command (and every other keybinding command)
seems to be rejected by a chack that was intended for
org.mate.Marco.global-keybindings but has no sence for org.mate.Marco.keybinding-commands.
I guess it is an artifact of getting rid of matefonf.

This issue was brought up in irc. A patch is located here: https://bugs.launchpad.net/ubuntu/+bug/299330/comments/46

When a window is set to be sticky, alt-tabbing to it brings you back on the workspace that window was onto before making it sticky.

The expected (and usual) behaviour would be that the sticky window would be given focus while remaining on the same workspace.

This only happens when switching to the window with the alt-tab shortcut. Clicking on the window button in the window list (in the panel) does no do this.

I have window compositing turned on.

Repro steps:

1. Open System > Preferences > Keyboard
2. Click on the Layouts tab, then the Options button
3. Expand Alt/Win key behaviour
4. Select the Meta is mapped to Win keys option
5. Close the prefs windows that were opened
6. Open System > Preferences > Windows
7. Under the Movement Key section, select the Meta option
8. Hold down your Windows key, click and drag on a window, and confirm that the option has taken effect
9. Log out and back in
10. Repeat step 7, by holding down the Windows key and clicking and dragging a window around
    Expected: the setting should still be in effect
    Actual: the windows key no longer allows the user to move windows around

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

Orginal report at fedora bugzilla.
https://bugzilla.redhat.com/show_bug.cgi?id=1011869

glad08 2013-09-25 05:53:15 EDT

Description of problem:
I close my programs and logout, then log in again and marco crashes. In general, I use compiz, bat afaik marco starts before compiz and after that compiz replaces it.

Version-Release number of selected component:
mate-window-manager-1.6.2-3.fc19

Additional info:
reporter: libreport-2.1.7
backtrace_rating: 4
cmdline: marco
crash_function: meta_rectangle_edge_aligns
executable: /usr/bin/marco
kernel: 3.11.1-200.fc19.x86_64
runlevel: N 5
type: CCpp
uid: 1000
var_log_messages: Sep 25 13:33:13 localhost abrt[29721]: Saved core dump of pid 29090 (/usr/bin/marco) to /var/tmp/abrt/ccpp-2013-09-25-13:33:12-29090 (20348928 bytes)

Truncated backtrace:
Thread no. 1 (10 frames)
 #3 meta_rectangle_edge_aligns at core/boxes.c:1159
 #4 apply_edge_resistance at core/edge-resistance.c:391
 #5 apply_edge_resistance_to_each_side at core/edge-resistance.c:610
 #6 meta_window_edge_resistance_for_move at core/edge-resistance.c:1162
 #7 update_move at core/window.c:7010
 #8 meta_window_handle_mouse_grab_op_event at core/window.c:7450
 #9 event_callback at core/display.c:2039
 #10 filter_func at ui/ui.c:209
 #11 gdk_event_apply_filters at gdkevents-x11.c:356
 #12 gdk_event_translate at gdkevents-x11.c:946

backtrace:
https://bugzilla.redhat.com/attachment.cgi?id=802671
core_backtrace:
https://bugzilla.redhat.com/attachment.cgi?id=802673

This is a feature request.

If one has a window on Desktop 1 and a window on Desktop 2, there is currently no keyboard-based way to switch between the two.

I'm not sure if this is the right place to report it but I think that MATE should handle this situation.
A short introduction.
I have a Dell Latitdue E6500 nad it has fn+f8 combination for cycle through monitors. In 2.6 kernels on old Gnome it worked fine. However newer kernels report to BIOS as Windows 7 which handles this cycling different and now my BIOS sends Super Key+ P and after few seconds is sends return/enter. Here is a longer description of the problem
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/539477

I think that MATE should react on super+p somehow.
XFace for instance shows a dialog on which I can change the display (clone, 2 monitors...)

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

Problem found on F18 (64bit) With both included OpenJDK-1.7 (1.7.0_09-icedtea) and Oracle JDK 1.6 (java version "1.6.0_38")

Any SWING java program can be used to reproduce, ie Netbeans.

JDK 1.6 Problem: When Java window is maximized the menu will immediately disappear when mouse button is released. If mouse button is pressed, there will be huge offset between curser and menu position.

Open-JDK 1.7 Problem: menu will immediately disappear, however rather than having curser offset, the menu will popup wrong place.

It seems there exists some workarounds in java runtime for this problem, but the fixes will be applied depending on name of window manager, and MATE is not included. We need to get this included in Oracle JRE and OpenJDK

Cinnamon suffers from similar problem:
linuxmint/cinnamon#573

This has been an issue with both GNOME 2 and GNOME 3 until 3.4 come out. I recently tried MATE again. I hoped this issue was fixed. It wasn't.

Please see the bug report I filed with the GNOME devs as the problem with MATE is exactly the same.

How to reproduce this bug:

1. Have two monitors
2. Open Firefox or Chrome
3. Maximize the window
4. Press "Ctrl + N"

What happens:

The new window appears on the other monitor (You might have to test this with both monitors, as the new created one always appears on the same monitor for me. So when I do the steps above on the other monitor, it behaves as expected.)

What should happen:

The new window should appear on the same monitor as the old one. No matter which of both was the old one.

I have tried

marco --replace --no-force-fullscreen

But every window I open is maximized. And the title bar vanished if I manually maximize something.

note, compiled from sources.

One feature I miss from XFWM is the ability to set a flag so windows are not auto-raised when you click their contents. Clicking the border or alt-clicking in the window raises the window, but just clicking a control within leaves the window where it is in the stack.

This makes drag and drop operations (and series of copy-and-paste operations on smaller screens where windows can't be tiled side-by-side) that little bit easier.

It does exist after a fashion, but it's slightly unreliable and can only be enabled if 'focus-follows-mouse' is also enabled, something I don't want (I still want click to focus but not be forced to raise the window in the same operation).

I suggest to enable compositing manager by default on mate-window-manager; this should allow to have transparencies by default. I'm assuming most modern distributions provide hardware support to have this feature enabled by default.

make[2]: Entering directory /home/rave/rpmbuild/BUILD/marco-1.4.0/po' file=echo am | sed 's,./,,'.gmo \ && rm -f $file && /usr/bin/msgfmt -o $file am.po file=echo ar | sed 's,./,,'.gmo \ && rm -f $file && /usr/bin/msgfmt -o $file ar.po file=echo as | sed 's,./,,'.gmo \ && rm -f $file && /usr/bin/msgfmt -o $file as.po file=echo ast | sed 's,./,,'.gmo \ && rm -f $file && /usr/bin/msgfmt -o $file ast.po file=echo az | sed 's,./,,'.gmo \ && rm -f $file && /usr/bin/msgfmt -o $file az.po file=echo be | sed 's,./,,'.gmo \ && rm -f $file && /usr/bin/msgfmt -o $file be.po file=echo be@latin | sed 's,._/,,'.gmo \ && rm -f $file && /usr/bin/msgfmt -o $file [email protected] as.po:120:msgid' and msgstr' entries do not both end with '\n' /usr/bin/msgfmt: found 1 fatal error make[2]: *_\* [as.gmo] Error 1 make[2]: **\* Waiting for unfinished jobs.... make[2]: Leaving directory /home/rave/rpmbuild/BUILD/marco-1.4.0/po'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory`/home/rave/rpmbuild/BUILD/marco-1.4.0'
make: *** [all] Error 2

I've disable all fedora patches and other modifications for this shot.

It was a lot more efficient when I could use this compiz affect on Ubuntu 10.04 (in gnome 2).

Here is the description: When you move the mouse away from window X to
window Y, window X becomes transparent so that window Y can be seen if it
is behind window X.

This is the most important feature I miss from gnome 2.

Are there any chances to see this window-pacement patch applied?
http://chad.glendenin.com/metacity/patch.html

Connecting to a mate-session via VNC;

While holding Alt;

• After first tab stroke the popup-window appears and the "next" window is selected.
• After second tab stoke the popup-window vanishes and you're back in the original window.

Thus -- Alt-tab only work toggling between 2 windows.

Only in 1.6.

• Switching window manager to something like metacity gives back the desired behaviour
• Also tried a new user to only default settings -- and the same problem is there

Thus it's unlikely to be a config/setting error?

MATE is easily best Linux desktop for me.
Only this think missing for me is feature some people call Windows Aero Snap.

In compiz you can maximize window when you drag it to top or make it half width by dragging it to left or right side of screen. Some people at this point would say to install compiz but I don't want to because I don't need the effects and MATE's window manager is much more stable.

Mutter/muffin's snapping similar to compiz.

XFCE also has something similar if you disable setting that allows you to change workspace by moving mouse to left or right side of the screen.
What I don't like about XFCE:s implementation that dragging to top doesn't maximize window but makes it's height half of screen height.
But woudn't hurt to add setting that would make this feature also work like in xfwm4:

Please add this as setting for MATE window manager so that I can get my best Linux desktop experience ever :)

Hi everybody,
I am not sure if this is really a bug of marco, but my mate desktop sometimes lost response of mouse clicks, as well as some keyboard controls such as switching workspaces using 'Ctrl + Alt + arrow', or switching between windows using 'alt + tab'. Beside these, all other keyboard controls are okay. This faulty state arises frequently, sometimes it restore automatically after a couple of minutes, sometimes it won't restore, and I found killing and restating MARCO brings me back to normal.

I've tried this with both opensource and Catalyst drivers, and the problem remains the same.
It's simple: after enabling compositing in mate-conf-editor moving windows becomes very slow and laggy. I've tried to disable/enable VSync, but it didn't seem to help.
Please tell me if you need more information, and sorry for my English.

I've experienced this issue in all of the MATE installations I've done so far (all of them with Linux Mint), on different architectures, hardware profiles, and MATE versions (from 1.2 onwards).

Simply put, when I hover the pointer over the network connection icon at the bottom right of the screen, a small pop-up appears with some details about the current active connection. This pop-up is standard behaviour on all of the plugins in the desktop bar, as far as I can tell. But the network connection plugin has an unexpected behaviour, in that the shadow that accompanies the little pop-up remains in the screen after the pop-up is gone. See the following screencaps for an example:

pop-up message with outline present: http://s10.postimage.org/9c8z6f8uh/Screenshot4.png
pop-up message gone, the outline remains: http://postimage.org/image/aryzv3gff/

I'm pretty ure it's a Marco thing, because I briefly tried Compiz and this didn't happen.

The option, "Select windows when the mouse moves over them", regularly stops working for me. Unfortunately, I haven't been able to determine any cause or pattern to describe how this might be reproduced.

Unchecking and rechecking the option in Window Preferences has no effect. However, running marco --replace is effective in restoring this functionality.

When mate-window-manager's (1.6.0) compositing is enabled and Audacious (3.3.4) is using the Winamp Classic Interface, double-clicking its title bar will shade it (leaving artifacts, as in this image:
),
but double-clicking its title bar again will not unshade it. Rather, a black square the size the Audacious window should be is shown and the program's display seems to freeze (see:
)
If I change nothing except for turning off mate-window-manager's compositing, this no longer happens.

Theme used in Audacious does not affect this. I'm using Arch Linux 32-bit, up to date as of Apr 8, 2013. My video card is nVidia GeForce 8800 GT, using proprietary drivers version 313.30.

Patch from metacity:
http://bugzilla-attachments.gnome.org/attachment.cgi?id=141446
Without this patch you'll see a mysterious window after login if using mdm.
A detail description of this issue you'll find at mate forum.
http://forums.mate-desktop.org/viewtopic.php?f=8&t=14

If a user activates composite mode, the windows get a bit of shadowing; I find the default shadow print a bit exagerated; would it be possible we could improve it to a smoother thing more pleasant to the eye ?

There are currently 2 options how new windows get focus and in my opinion it lacks an option which is more workflow friendly (and therefore less strict).

Consider the following example:
I open a gui application (normally I would expect focus on that window) but now I manually switch focus to another (older) window and doing work while the application has finished loading. What happens now is that this new window steals focus from my manually selected window.

Solution:
Add a new focus mode which does not focus new windows when focus has changed after starting an application.

http://openbox.org/wiki/Help:WindowsInteractionWithWindowsKey

I used OpenBox before Mate, and I loved the openbox's windows interactions.
I think this is very useful things. I am missing from MATE.

Title is almost self-explanatory. It would help a lot when using multiple monitors and having many windows with that option.

Window doesn't maximize, when drop over the top screen edge. Option side-by-side-tiling is checked.

I had been using 25 in GNOME 2, and then 25 in xfce.

In mate 1.6 it allowed me to set preferences to use 25 (5_5), but caused marco to error out when I logged back in.
I suggest the max should be 9_9 or 81 - though I doubt if anyone sane would more than 6*6 - but I have been known to be wrong!

Hey,
the new "open new windows in the middle of your screen" option is not working on the secound screen. The new windows spawnes again in the top left corner on the secound screen.

Greetings
Leo

All desktop files i've to validate with 'desktop-file-validate' for fedora.

/builddir/build/BUILDROOT/mate-window-manager-1.4.0-5.fc16.x86_64//usr/share/mate/wm-properties/marco-wm.desktop: error: file contains group "Window Manager", but groups extending the format should start with "X-"

This patch fix the error.
https://dl.dropbox.com/u/49862637/Mate-desktop/patches/mate-window-manager-1.4.0_fix_group_in_desktop_file.patch

I know this it a old gnome remnant.
If you get confirm with me i can do a pull request.

User on Mate IRC channel likes to have KDE like behaviour starting application on specific workspace.

TCW suggested this could be done on the command-line like this:
$ marco --wp=3 application
or
$ marco --start-app --wp=3 application

The user can then use this feature with .desktop files to get desired result. The marco on command-line needs to communicate this data over DBus to the running marco window manager before exiting.

I have set-up following configuration with TWO X-screens and THREE physical monitors:

DISPLAY=:0.0
Monitor 1: 1920x1200 at 0,0
Monitor 2: 1600x1200 at 1920,0

DISPLAY=:0.1
Monitor 3: 1920x1200 at 0,0

All window on screen 0 maximized across both monitors, i.e. becomes 3520x1200

But without second X-screen window maximization is correct, to 1920x1200 at 0,0 or 1600x1200 at 1920,0 depending on current monitor.

I write small test program, that calls gdk_screen_get_monitor_at_point() / get_monitor_geometry() and I can see correct monitor layout:

export DISPLAY=:0.0

./test 10 10
Position: X=10, Y=10.
Monitor: 0
Geometry: 1920x1200+0+0
Plug: DFP6

./test 2010 10
Position: X=2010, Y=10.
Monitor: 1
Geometry: 1600x1200+1920+0
Plug: CRT1

export DISPLAY=:0.1

./test 10 10
Position: X=10, Y=10.
Monitor: 0
Geometry: 1920x1200+0+0
Plug: DFP1

Fedora 16, marco-1.4.1-0102.fc16.x86_64 installed from repository.

Hi,
I'm not sure if I reported this here already, or if we just spoke about this in #mate. I wasn't able to find this report in the tracker anywhere.

I set up keyboard shortcuts in mateconf-editor under the marco app. They are for "move window to east edge" and "move window to west edge".

Those keyboard combos work for many applications, but don't for firefox. Initially the window assumes the new position, but as soon as I let go of the keyboard combination the window jumps back to where it was before.

I believe there were more programs that did that but I can't remember what they were off the top of my head.

Since Firefox is a fairly popular program and moving windows around is daily routine in two-monitor environments I would ask that this gets looked into. Currently Firefox makes up more than 50% of my work day and it would be a huge improvement in productivity for me if this started working properly. As it is I can only use FF on one monitor.

I am on Ubuntu 12.04.

$ marco --version
marco 1.4.1

$ firefox --version
Mozilla Firefox 18.0

Thanks

When window compositing is enabled, a 1px white bar shows on the left edge of the screen, and sometimes also on the top edge. This only happens when an application is in full screen or a window making use of compositing is on screen (like a transparent background terminal window).

This seems to only happen on LMDE and debian with the Mate repository, not the Ubuntu-based Linux Mint.

I have tested and successfully reproduced this behaviour with an intel graphics chipset and an AMD radeon card (open-source drivers in both cases), on debian and on LMDE.

Note that on those machines, Compiz-fusion handles compositing properly and works as expected.

For some reason when I start calibre (ebook manager) sometimes on startup it makes all panels transparent. The panels are still there and can be interacted.

When I turn compositing off and then on it is back to normal. Or restarting marco has the same affect.

The programs I had running at the time are:
virtualbox running a vm
firefox
hexchat
thunderbird
a couple of mate-terminals

The only thing I see in .xsession-error is the following:
Windowmanager waarschuwing:Ongeldig WM_TRANSIENT_FOR venster 0x4200004 opgegeven voor 0x4200002 (calibre-gu).

The error happens always but it does not make the panels transparent every time.

The hacking file is out of date, and contains incorrect instructions :

$ svn checkout http://svn.gnome.org/svn/marco/trunk marco
$ cd marco
$ ./autogen.sh --prefix /usr

I am also quite puzzled by the reference to fibonnaci sequence, the mention of CVS, etc

If an application is made to display a tooltip, and the virtual desktop (workspace) is switched using a keyboard shortcut that doesn't have a side-effect of turning off that tooltip, the tooltip will be inappropriately displayed on all workspaces, and can only be removed by returning to the original workspace.

For me this most commonly happens with tooltips displayed by the Chrome browser when workspaces are switched by CTRL-LEFT and CTRL-RIGHT key combos.

Workspace switcher 1.6.0

    $ mate-about --version
    MATE Desktop Environment 1.6.0
    $ uname -a
    Linux rotfisch 3.5.0-32-generic #53~precise1-Ubuntu SMP Wed May 29 20:33:37 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

When you rightclick the workspace switcher and name your workspaces, they don't show up in context menu of windows ("Move to other workspace").

I hope I'm in the correct place. My package lsb-base_4.0-0ubuntu27.1_all updated to *.27.3_all, and so did s/base/release/ on 13 Sept via update manager.

Now the symptoms are: if I start uxterm or xchat, the monitors start blinking (mate display manager I presume). Most things seem to work ok. gnome-terminal works ok, for example. I can run this browser, and so on.

I started xchat from a terminal but saw no errors thrown. I'm not certain where the mdm error log is.

They were apparently fixing some python hooks or something in lsb (I don't know much about it, obviously).

http://changelogs.ubuntu.com/changelogs/pool/main/l/lsb/lsb_4.0-0ubuntu27.3/changelog

I'm on asus amd_64 with nvidia drivers geforce G210M cuda 512 MB
linux mint 15 olivia mate

Thank you

libmarco-private0.i586: W: shared-lib-calls-exit /usr/lib/libmarco-private.so.0.0.0 exit@GLIBC_2.0
This library package calls exit() or _exit(), probably in a non-fork()
context. Doing so from a library is strongly discouraged - when a library
function calls exit(), it prevents the calling program from handling the
error, reporting it to the user, closing files properly, and cleaning up any
state that the program has. It is preferred for the library to return an
actual error code and let the calling program decide how to handle the
situation.

Hi there!

I'm really loving Mate. Switching back to a Gnome2-style desktop after using XFCE4 for over a year feels really good. :)

I wanted to suggest a feature for Marco's Alt-Tab window -- vertical layout:
http://www.ntwind.com/software/vistaswitcher/screenshots.html

I'm a big fan of this layout -- it lets you see the full window titles, and there's something more intuitive about the recently used windows being on the top of a stack, instead of on the left of a row of icons.

Would this be reasonably simple to implement, given Marco's current codebase?

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

Examples:

1. If one window is maximized in a workspace, when you go to other workspace, that maximized windows is carried together to other workspace
2. If two or more windows maximized in a workspace, when you go to other workspace, that maximized and with focus windows is carried together to other workspace.
3. If you have two maximized windows, when you switch to another workspace, one of the is carried as I said and the other is minimized.

• Failed Running gtkdocize !
• Include in your bugreport the contents of:
• /var/tmp/portage/mate-base/mate-corba-1.4.0/temp/gtkdocize.out
• ERROR: mate-base/mate-corba-1.4.0 failed (prepare phase):
• Failed Running gtkdocize !
• Call stack:

• ebuild.sh, line   85:  Called src_prepare
  

• environment, line 3357: Called mate_src_prepare
• environment, line 2915: Called autotools_run_tool 'gtkdocize' '--copy'
• environment, line 576: Called die
• The specific snippet of code:

•       die "Failed Running $1 !";
  

• If you need support, post the output of emerge --info '=mate-base/mate-corba-1.4.0',
• the complete build log and the output of emerge -pqv '=mate-base/mate-corba-1.4.0'.
• This ebuild used the following eclasses from overlays:
• /var/lib/layman/mate/eclass/mate.eclass
• /var/lib/layman/mate/eclass/mate-desktop.org.eclass
• /var/lib/layman/mate/eclass/mate-utils.eclass
• This ebuild is from an overlay named 'mate-overlay': '/var/lib/layman/mate/'
• The complete build log is located at '/var/tmp/portage/mate-base/mate-corba-1.4.0/temp/build.log'.
• The ebuild environment file is located at '/var/tmp/portage/mate-base/mate-corba-1.4.0/temp/environment'.
• Working directory: '/var/tmp/portage/mate-base/mate-corba-1.4.0/work/mate-corba-1.4.0'
• S: '/var/tmp/portage/mate-base/mate-corba-1.4.0/work/mate-corba-1.4.0'

***** gtkdocize *****
***** PWD: /var/tmp/portage/mate-base/mate-corba-1.4.0/work/mate-corba-1.4.0
***** gtkdocize --copy

/var/tmp/portage/mate-base/mate-corba-1.4.0/temp/environment: line 567: gtkdocize: command not found

Applications with Java GUIs (such as Netbeans or Arduino IDE) won't work when their windows are resized or maximized.

A workaround is to place the window at the top left corner before maximizing it.

According to this discussion linuxmint/muffin#11, the problem has to do with JDK hardcoding the names of the different window managers and not having some into account (like muffin), so I guess it may be the same with Marco.

The issue may be related to the one described in #30

The problem persists in openjdk-6, openjdk-7 and sun-java-6.

I'm using Debian Wheezy amd64, MATE 1.6 and Nvidia propietary drivers.

With compositing enabled, shadowed tooltips sometimes leave their shadows behind when they disappear. See the attached image. I hovered over one of the links, which brought a tooltip showing the link's destination. When I moved my mouse, the tooltip disappeared, but the shadow did not.

Arch Linux x86
mate-window-manager 1.6.1 from the Arch Maté repo
the shown program is evince 3.6.1