On 27/05/15 10:12, Martin Lucina wrote:
On Wednesday, 27.05.2015 at 09:45, Antti Kantee wrote:
Does that statement mean that you have working code or you played
around with it in your head? You read like an academic paper where
one can never be quite sure if they're describing something that
works or works on paper ;)
Working on the code now...
Thanks.
Looks generally good.
Why do you need i686 for the -T argument to rumpbake? Did you mean
e.g. hw-generic instead?
It should be -T i686-hw-generic. The CPU is in there just as a safeguard so
that the user doesn't try to rumpbake a binary built for x86_64 to an i686
target.
I'm not sure I follow that argumentation. If the only purpose of
including the cpu is to complain when it's wrong, it doesn't seem to
have much purpose except in the "ha ha you didn't know the magic word so
you're not cool enough to play with us" sense.
I assume there will be something like "rumpbake -T list". How does
one add targets? .rumpbakerc? -c rumpbakelist?
How are multiple programs handled? You probably give multiple
binaries to rumpbake, but then how do you decide in which order they
run? That's probably a rumprun thing, but how do you remember which
"executables" have been baked?
Where is the configuration file handled?
No idea yet. Will try and get something minimal working first.
That's fine, was more a loose/guiding requirement specification than
questions needing immediate answers.
