On Monday 12 June 2006 03:16, matt ingalls wrote:

> how are you changing the API?
> 
> if you are going to change the API anyway,
> i would say leave sensekey as it was and instead add
> a 'sensekeyup' opcode and related API function.

The changed sensekey opcode (with the optinal output argument) is
already in the 5.02.1 sources. However, I did not decide yet how
the API should be changed if using that is preferred over just
replacing the opcode from the host. The old csoundKeyPress()
function could be changed as you suggested earlier to interpret
key codes > 128 as key release, at the expense of limiting the
number of key codes that can be used to 127. Alternatively, a
new function could be added, either something like an extended
version of csoundKeyPress(), or a "sensekey callback". If a new
function is added, I would probably prefer the callback approach,
e.g. csoundSetSenseKeyCallback(CSOUND *, int (*func)(CSOUND *)),
as it leaves implementation details like buffering, thread
synchronization (if needed), etc. up to the host. Of course, one
could think of the idea of overriding the opcode as some sort of
"sensekey callback", too, a new API function is only more elegant
and avoids your specific problem of not being able to use csdl.h
(even if the latter would better be fixed anyway).

However, currently I am waiting for the new Gogins 5.02 Win32
installers before making any CVS changes; I already have some
uncommitted stuff.


_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net