Vojtech Pavlik schrieb:
On Tue, Feb 05, 2008 at 10:59:09PM -0500, Christoph Hellwig wrote:On Wed, Feb 06, 2008 at 12:12:05AM +0100, Tilman Schmidt wrote: > And by the way, what about that > > kind of tool in >> one of the input packages to attach input line disciplines > you mentioned in your mail dated 26.01.2008 20:28? Can you point me > to a source so that I can make sure ldattach covers its functionality > too? Ok, did a little search in my archives, and it's called inputattach. It's a package of it's own in most distributions but I can't actually find an upstream source for it. I've Cc'ed Vojtech because he wrote it.I used to maintain it in the linuxconsole.sf.net CVS, now it's maintained by Dmitry Torokhov. Ccing Dmitry - Dmitry, is the development now moved or is still the sf.net repository the master?
Ok, I found a version at http://www.kernel.org/pub/linux/kernel/people/dtor/inputattach.c which looks reasonably recent (v1.24 2006/02/06) and had a quick look. Apart from the TIOCSETD ioctl which it has in common with ldattach, it also contains a large table of device specific data and methods for initializing the various input devices which can be driven by the N_MOUSE line discipline. In order to merge this into ldattach, I would probably extend the <disc> command line argument in order to select table-driven pre- and post-TIOCSETD actions and parameter presets. For example: current: inputattach --mouseman /dev/ttyS0 -> new: ldattach INPUT=mouseman /dev/ttyS0 (Defining INPUT as an alias for MOUSE in order to avoid ugly combinations like MOUSE=sunkbd. ;-) I am however somewhat unsure whether such a solution would be desirable, or if the requirements of N_MOUSE are so unique as to warrant a separate utility. Opinions? Thx T. -- Tilman Schmidt E-Mail: tilman@xxxxxxx Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
Attachment:
signature.asc
Description: OpenPGP digital signature
