Re: [Freevo-users] IVTV-record bugs and problems

[Rob S. <ro...@in...>]

Robert Winder wrote:
> Driver: ivtvr=CE=E0=E2=D2=CD=F3=BC=C4
> Card: Vanilla iTVC15 card q=CE=E2[=C0

Yeah, that junk's causing problems.

> QUERYCAP_ST  =3D "16s32s32sLL16x"
> QUERYCAP_NO  =3D _IOR('V',  0, QUERYCAP_ST)
>=20
> Not sure what to make of this mumbo jumbo ;-) Rob can you explain wha=
t
> you are trying to do here with 16s32s32sLL16x ? Willing to help if yo=
u
> got some pointers.

Heh.  Ok, start by reading:

http://www.python.org/doc/2.3.4/lib/module-struct.html

We need to turn the results of the QUERYCAP_NO (really a hex number)=20
driver ioctl call into something python and understands.

     def querycap(self):
         val =3D struct.pack( QUERYCAP_ST, "", "", "", 0, 0 )
         r =3D fcntl.ioctl(self.device, long(QUERYCAP_NO), val)
         return struct.unpack( QUERYCAP_ST, r )

That function gives us a list returned by the ioctl.  Also note in the=20
struct pack line: "", "", "", 0, 0; a string, string, string, int, int.=
=20
    Well in C these tyles are more complec, and from reading the v4l=20
header file you can see the struct returned has members like 16 char=20
array, 2 - 32 char arrays, 2 - unsigned longs, and 16 bytes of padding.=
=20
  Ok, so this one probably doesn't ring a bell now because I just=20
remembered Thomas was the one who made this file, and I think that was=20
based on another project, so, perhaps the interface has changed.  Maybe=
=20
someone should review the v4l or ivtv header files and compare all the=20
struct packing/unpacking.  I will take a peek later perhaps but atm I'm=
=20
a bit busy.

-Rob