On Thu, Oct 27, 2016 at 11:03 AM, <Adrian.Danis@data61.csiro.au> wrote:

Not being able to get error information after doing an invocation is unacceptable as a default. I suppose if you have an application where you *know* you won't need the error information then you might want to re-enable this.

Ideally we come up with some better stub generation that doesn't throw away the errors.

OK, clearly I have to read these functions way more closely to grok the difference. :-)

Oddly, I can't find any public code that uses the *WithMRs functions. In order to give them regular exercise, would you like me to add some of them to the syscalls tests? :-)

Thanks,

Jeff