Questionable breakpoint stepping code

[Vladimir Prus]

> On Mon, 2007-11-26 at 22:16 +0300, Vladimir Prus wrote:
> 
> > > As for your last question, no, I don't believe we approve
> > > of ifdefs...
> > 
> > Well, the question, then is -- how do we make this code work
> > correctly even if the instruction at PC is not 4 bytes in size?
> > Calling disassembler seems a plausible approach.

On Monday 26 November 2007 22:16:16 Michael Snyder wrote:

> Good suggestion.

On Monday 26 November 2007 22:24:07 Daniel Jacobowitz wrote:

> Don't Do It.  Just call error, I suggest.  Skipping an instruction
> can lead to bizarre failures later.

Ok, I'm now not quite sure what to do. I personally prefer calling 'error' --
is this going to be acceptable?

- Volodya