RFC: partially available registers

[Tom Tromey]

>>>>> "Daniel" == Daniel Jacobowitz <drow@false.org> writes:

Daniel> As for the patch itself:
Daniel> * Some documentation on the gdbarch method would be nice, in
Daniel> particular, the return value.  Does 0 mean "not a pseudo"?

Yeah, I left that out for the initial change, but I think I shouldn't
have.

I find gdbarch.sh very hard to read.  Does anybody else?

Daniel> * I am not happy about having to implement both
Daniel> gdbarch_pseudo_register_read and gdbarch_pseudo_register_read_value,
Daniel> depending on which regcache read function was called.  So for a final
Daniel> version, is it practical to push this down and only call the value
Daniel> version if it is registered?  That means implementing the existing
Daniel> regcache read in terms of the new one, instead of the other way
Daniel> around.

I will look at it.  Thanks for looking at this.

Tom