Re: [Simulavr-devel] gdb interrupt patch.

[Theodore A. Roth]

On Wed, 2 Jan 2002, ken restivo wrote:

:)
:)my understanding is that this is os-specific. if i remember correctly,
:)it is required on bsd and solaris, but not linux. linux signal()
:)behaves more like sigaction(), and reinstalls the handler. the signal
:)manpage in linux advises using sigaction anyway.
:)

I was lazy and threw in the signal() stuff as a quickie with no intention
of actually keeping it.

I'm doing the rewrite now to use sigaction(). If anyone has any objections
to sigaction, please speak up now. I'm creating two new files: signal.c
and signal.h. I'll try to document the (public) signal interface in the
header to be compatible with doxygen and see how it works out.

If we do go with doxygen, I'm thinking the the header files are going to 
need some reworking. avr.h is already too large and adding more comments 
to it won't help. My choices in splitting up the modularity so far haven't 
been optimal (nor some of my name choices). If we can clean up this it 
should make things a bit more intuitive as far as where to look for 
certain implementations.

Ted