From: Andrew Apted <ajapted@netspace.net.au>
  To  : ggi-develop@eskimo.com
  Date: Sun, 20 Sep 1998 12:17:25 +1000

Re: coupla questions

Neal Tucker writes:

>  1) Should frame-flipping work on all targets?  I've been using it on
>  the X target, and when I run my program on kgicon (configured for
>  S3-968/S3-968/TVP3026/TVP3026/multisync), it looks like the frame-
>  flipping just never happens (and it's going to send me into an
>  epileptic fit if I'm not careful :-). 

The fbdev target now does frame-flipping, but I only committed the code
a few days ago, so maybe you haven't got the New and Improved (tm)
version yet ?

>  Is there some way I'm supposed to query to see how many frames are
>  available to draw on?

Set the frames field of ggi_mode to a high number (4, 8, or as many as
you'd really need) and call ggiCheckMode(vis, &mode).  The target will
either accept it, or lower the value to something that it can do.  

That's the theory anyway... the fbdev target's checkmode isn't fully
implemented yet.  [and it won't be until kgicon supports
FB_ACTIVATE_TEST, so don't hold your breath ;)].

>  2) (not directly ggi-related) Is it bad manners to call dlopen or dlsym
>  and then not call dlerror if it fails?  I check dlerror after every
>  call to dlopen/dlsym, and it took me a while to figure out that dlerror
>  is reporting an error to me because the last person who called dlopen
>  (libggi/dl.c) got an error, but just decided to bail without checking
>  dlerror.  The dl* manpage says that all the functions return 0 on
>  failure, so I've changed my own code to only call dlerror if I get a 0
>  back, but I thought I'd mention it anyway.
  
dlopen() and dlsym() return NULL on error (according to my manpage
anyway), so you only need to bother with dlerror() if a the dlsym()
could *validly* return a NULL, but for LibGGI I don't think it can.

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