Re: wms offsetting XMoveWindow() coords

On Thu, 23 Sep 1999, Tim Janik wrote:

> hello everyone,
> i was just trying to fix up desk-guide (the GNOME pager) movement in a
> window manager independant fashion and encountered two different window
> manager behaviours (probably due to window manager families):
> for a window MyApp being managed by the window manager, and its window
> manager decoration window Frame, with MyApp being offsetted within
> Frame by (fx, fy),
> XMoveWindow (D, MyApp, x, y);
> 1) will move MyApp to (x, y) and thus Frame to (x-fx, y-fy) for
>    Enlightenment,
>    IceWM,
>    AfterStep,
>    fvwm
> 2) will move Frame to (x, y) and thus MyApp to (x+fx, y+fy) for
>    Sawmill
>    twm, ctwm, vtwm,
>    olvwm
>    wm2
>    WindowMaker

Interesting... which does BB follow? =) Had to ask.

> since the application and pagers can hardly tell where exactly the window
> will be moved to, i'd like to propose the addition of another property to
> the new window manager spec:
> _NET_WIN_MOVE_OFFSET xoffset,yoffset CARDINAL[2]/32

Will add it to the spec on tonight in the DB
(I should have the rest of the spec migrated by the end of Saturday).
However, I do need to know what section of the spec you want it located

>     An array of two cardinals to be set on all window managed by the window
>     manager. The two cardinals xoffset and yoffset specify the offset in pixels
>     that will be added to the window position by the window manager upon
>     XMoveWindow() requests for mapped windows.
>     Note: This is required by applications (pagers) that need to determine
>           where exactly the application window will end up after movement
>           requests, if the window manager will actually move its decorative
>           frame window to the specified coordinates, instead of the real
>           application window.
> so for case 2) window managers, these properties will contain the application
> windows' offsets within their frame windows, and for case 1) window managers,
> they will simply contain (0, 0).

For those using the site, would it be
helpful to create a mechanism to add an actual text item to the DB as
well, or is it preferred to comment there and add items after notification
on the mailing list?

And for those that don't have a login, please let me know with a crypt()'d
password as soon as possible. I really want to migrate towards it for
developing the spec further.


Nathan P. Clemons                    "Peace favor your code."              ICQ: 2810688

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]