Index: [thread] [date] [subject] [author]
  From: Andrew Apted <ajapted@netspace.net.au>
  To  : ggi-develop@eskimo.com
  Date: Tue, 25 Aug 1998 00:27:40 +1000

Re: Sigh...

Marcus writes:

>  Andrew Apted wrote:
>  > Those static vars are there because the code was going to be shared with
>  > the SUIDKGI target, so vis->private wouldn't do. What I'm going to do
>  > is move them into display/L_common/vtswitch.inc (ETC), and get fbdev and
>  > suidkgi to #include them.  (put linux2ggikeysym.inc in there as well).
>  > (the 'L' means Linux BTW).
>  
>  I'm not following you there...
>  Does the fbdev target really need static variables?
>  And if so, which and why?

The vtswitch.c, kbd.c, and mouse.c code was intended to be shared by
both fbdev target and suidkgi target.  That means I couldn't use
LIBGGI_PRIVATE(vis) since these two targets have different data
structures there.

Those files will become L_common/vtswitch.inc, L_common/kbd.inc and
L_common/mouse.inc (and in the not-too-distant future, will add
L_common/joystick.inc) and access to the needed vars will be macrotized
(like mansync before the librarytization).  [gee, _two_ new words :))]

Also the signal handling code in vtswitch.c *must* use a static/global
variable since, under UNIX, you can't pass private data to a signal
handler.  (But I guess _that_ is obvious :).

Cheers,
_____________________________________________  ____
                                               \  /
  Andrew Apted   <andrew@ggi-project.org>       \/
  

Index: [thread] [date] [subject] [author]