From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1D6R6t-0005c5-OF for qemu-devel@nongnu.org; Wed, 02 Mar 2005 05:29:44 -0500 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1D6R6c-0005bY-Mo for qemu-devel@nongnu.org; Wed, 02 Mar 2005 05:29:30 -0500 Received: from [199.232.76.173] (helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1D6R6E-0005Lz-4d for qemu-devel@nongnu.org; Wed, 02 Mar 2005 05:29:02 -0500 Received: from [193.96.240.2] (helo=br2uxs04.hb.orthogon.com) by monty-python.gnu.org with esmtp (Exim 4.34) id 1D6QYO-0005Lv-T0 for qemu-devel@nongnu.org; Wed, 02 Mar 2005 04:54:05 -0500 Received: from br2uxs03.hb.orthogon.com (br2uxs03.hb.orthogon.com [150.158.146.28]) by br2uxs04.hb.orthogon.com (8.11.7+Sun/8.8.8/Ortho-4.10) with ESMTP id j229s0718918 for ; Wed, 2 Mar 2005 10:54:01 +0100 (MET) Received: from [150.158.146.126] (br2wsl15.hb.orthogon.com [150.158.146.126]) by br2uxs03.hb.orthogon.com (8.8.8+Sun/8.8.8/Ortho-4.10) with ESMTP id KAA27727 for ; Wed, 2 Mar 2005 10:53:56 +0100 (MET) Message-ID: <42258D33.9080402@gmx.de> Date: Wed, 02 Mar 2005 10:53:55 +0100 From: Oliver Gerlich MIME-Version: 1.0 Subject: Re: [Qemu-devel] [Patch] Release mouse at window edges References: <6879.1109430585@www57.gmx.net> <4224DE76.4060803@bellard.org> <42254590.4070905@wasp.net.au> In-Reply-To: <42254590.4070905@wasp.net.au> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Brad Campbell wrote: > Fabrice Bellard wrote: > >> The idea is interesting. Does it cause regressions in the normal mode >> ? If so, an option to enable it may be good. I think it is a problem >> that your patch is dependent on the cursor size. >> > > This is a good point. I have had to revert it for WinXP as every time > the cursor changes shape (and it does all the time as you move over > stuff in Office2003) it releases the grab. > > In addition I have had all manner of painting and rendering issues with > the IDG display driver and have had to revert to the built in driver to > get some software to paint properly. > > Nice idea, might need a little work still. > > Brad An extra switch would be good, indeed. Currently, grabbing/ungrabbing is disabled if no hw cursor is used, but it's not reliable enough (for example, Linux initializes hw cursor at framebuffer initialization, although no cursor is visible). The problem with cursor size appears differently on different systems. On Win98, hw_cursor_x goes from 1 to 1023 (in 1024x768), while on Win2k, it goes from 0 to 1013... The patch circumvents this problem with an ugly hack, but I have no idea how to get around this cleanly. Brad: your problems with Office might be related to another bug I discovered (Internet Explorer 5 "disables" hw cursor by moving the cursor to 2047/2047 which ungrabs the cursor as soon as it enters an IE window). I'll post an improved patch tonight, with this and other bugs fixed. BTW. is there a "make realclean" (or similar) which also deletes qemu-doc.html, manpages, kqemu temp files etc, so that I can easily diff my patched snapshot to an unpatched one? Oliver Gerlich