enlightenment/BUGS

107 lines
3.7 KiB
Plaintext

From: Kim Woelders <kim@woelders.dk>
To: enlightenment-devel@lists.sourceforge.net
Subject: Re: [E-devel] Re: bugs with iconification/deiconification in e17.
Date: Fri, 02 Dec 2005 19:14:46 +0100
Sender: enlightenment-devel-admin@lists.sourceforge.net
Carsten Haitzler (The Rasterman) wrote:
> On Fri, 02 Dec 2005 09:19:20 +0200 <vkojouharov@gmail.com>
> babbled:
>
>
>>On Thu, 2005-12-01 at 18:26 +0100, Kim Woelders wrote:
>>
>>>Carsten Haitzler (The Rasterman) wrote:
>>>
>>>>On Wed, 30 Nov 2005 19:44:37 +0200
>>>><vkojouharov@gmail.com> babbled:
>>>
>>>...
>>>
>>>>>The other app is with deiconifying a window. Some programs (actually,
>>>>>only alltray comes to mind right now) use xlib to do the whole
>>>>>iconification thing. For the alltray instance, it seems to use
>>>>>XSetWMHints, set the state to NormalState, and basically that's it. And
>>>>>it seems to work for a lot of window managers too, so that must be a
>>>>>proper way to do it. But that doesn't work for e17, and the window stays
>>>>>iconified.
>>>>
>>>>
>>>>e waits for a map request. it doesn't respond to a change in hints for a
>>>>map. we can make it do so though :) i will write these down in the TODO.
>>>>
>>>
>>>I don't think a client can deiconify by changing a hint. e16 doesn't but
>>>does work with alltray. The normal way is to map the client window. In
>>>some cases clients send a _NET_ACTIVE_WINDOW message, but IIRC always to
>>>deiconify some other client as in tasklist and pager type
>>
>>I'm just curious here, what does alltray use to deiconify a window? cuz
>>that's the only relevant thing I could find in the code
>
>
> it prbably SHOULD use XMapWindow() or XMapRaised()
>
After having taken a peek at what is does in e16.8 on "alltray xterm",
the short version is that it uses XMapWindow() to map the client (which
actually is an alltray window containing the reparented real client)
first time. After that, when having been iconified, it sends a
_NET_ACTIVE_WINDOW client message to deiconify.
/Kim
dj2 raster, heres an interesting bug for you
dj2 i have twinview setup on my box
dj2 i can move the mouse between the 2 heads and e17 sees tehm as 2 heads
dj2 (2 pages etc)
dj2 if i try to drag a window between the 2 heads (say from left to right)
the mouse will lmove as expected
dj2 but when the window hits the right edge of the left monitor it will
appear again off the left edge of the left monitor
dj2 tho the mouse is now on the right monitor
raster xdpyinfo
raster see how many screens u have
raster screen #0:...
raster is there a screen #1 ?
dj2 number of screens: 2
dj2 yea screen #0 and screen #1
NB: in multihead if the mouse exits a screen during move or resize - either
disallow it (warp back to previous position ) or limit resize/move
From: Daniel Kasak <dkasak@nusconsulting.com.au>
To: enlightenment-devel@lists.sourceforge.net
Subject: [E-devel] Crasher ... switch to an app that's closing
Date: Mon, 20 Mar 2006 11:07:12 +1100
Sender: enlightenment-devel-admin@lists.sourceforge.net
Hi all.
I have a sort-of reproducible bug.
If you try to switch to an app which is in the process of shutting down,
and you do it at *just* the right moment, Enlightenment-0.17 will crash.
I've done this only about 3 times over probably more than a year of usage.
This particular time, I hit the 'close' button on a vmware-player
window. When this app gets the close signal, it actually minimises
itself, and then proceeds to shutdown ( which includes saving the
current VM state, which takes a while, hence the minimising first ).
After hitting close, I went to switch to another app by middle-clicking
on the desktop, but I missed the other app, and hit vmware-player
accidentally. Then everything came down in a heap :)
Sorry I don't have any debugging info.