Re: armhf TODO (was Re: Fosdem 2011: Debian on ARM)
On Mon, Feb 7, 2011 at 2:30 PM, Konstantinos Margaritis
<markos@genesi-usa.com> wrote:
> (removing a few people from the list)
> You're underestimating the value an automated system has,
no... i don't believe i am. the automated system itself is not the
important bit. the record of what changes/differences required to
overcome any problems found now that _is_ valuable.
> and overestimating the value of those notes.
... noo, i don't believe i am. _anything_ which saves other people
time, in a free software context, is valuable.
> What took me days to find out regarding the cyclic
> dependencies and what needs to be solved, would be a matter of seconds for
> such a tool.
ok. yes. true. and - let me check - you're saying that the bits
you actually solved are actually quite small [excluding the tracking
down of cyclic dependencies etc.]
> Mathematically speaking, such a tool if it does not exist already
> for debian packages, it should be trivial to write and detect such
> dependencies, no matter how many levels deep. Then with the output of this
> tool, policy changes might take place and the relevant packages fixed. There
> half the problem solved. What took me months, with these changes should take
> just a few weeks for a new porter. But this will never happen without this
> tool and without the policy changes.
i'm _so_ glad you've mentioned this :)
l.
Reply to: