On 30 September 2015 at 13:32, Antti Kantee <[email protected]> wrote:
> Oh you still planned to keep rumpbake?

Well that was what I understood the proposal to be. Getting rid of it
would be better.

> I think Krishna's comparison to pkg-config is appropriate, and also exposes
> the problem.
>
> pkg-config produces output that cc(1) doesn't know about, nor is the output
> required for every execution of cc(1).
>
> Your proposed rumplinkconfig produces output that rumpbake requires for
> every single successful link.
>
> As always, we'll follow the "separate normal case and worst case" principle.

Not sure I understand. pkg-config is designed to produce output for cc(1)?

Justin

Reply via email to