I think I can get rid of xf86Optrec.h in my code. But, data passing between wacom driver and xserwacom, a user space app,is required to enable the user with live function changes.
Do you have any good examples that fulfilled this kind of data passing mechanism which I can read/use? Thanks, Ping -----Original Message----- From: David Dawes To: [EMAIL PROTECTED] Sent: 11/21/03 9:48 PM Subject: Re: FW: RE: [Bug 537] wacom usb tablet uses wrong screen borders On Fri, Nov 21, 2003 at 08:17:45PM -0800, Ping Cheng wrote: >I am wondering if Egbert is taking vacation or something.Is there anyone >else can take care of my patch? I think he is. One problem with the patch is that modules are not permitted to include "xf86Optrec.h". The structures there are opaque to everything except the parser layer, and could change with a different parser layer. strcasecmp() isn't a reliable way to compare option names because option naming is more flexible than just case-insenstive. Use xf86NameCmp() instead. It also isn't permissible for drivers to be writing to files. If xsetwacom needs to know the device options, maybe it could link to the parser library itself and read the XF86Config file directly. I'm not clear on what else data is written out to files for. If data needs to be passed between a driver and an application, then a client-server communication mechanism should be used, either via an existing mechanism or via a new extension. David -- David Dawes developer/release engineer The XFree86 Project www.XFree86.org/~dawes _______________________________________________ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel _______________________________________________ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
