[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: AMD64 for sarge [Re: <rant> Package: ftpmasters, Severity: serious, ...]



Matthias Urlichs <smurf@smurf.noris.de> writes:

> Hi, Martin Michlmayr - Debian Project Leader wrote:
>
>>   - A general port inclusion policy: there are a number of pending ports
>>     (s390x, powerp64 and various BSD ports), and therefore it is
>>     important to have a clear policy saying which criteria a new port
>>     has to fulfil.
>> 
> Just use the policy we used for the last new architectures. I wasn't there
> at that time, but IMHO any sane policy should boil down to "if it works,
> it's in". (OK, so there also should be enough users -- let's say, "more
> than the second-least-popular architecture already in Debian" should
> work.)

The size of a sensible distribution should also factor into this (but
its hard to make rules for that).

What I mean is the following: Once multiarch is added after sarge is
out ports for s390x, powercp64, sparc64, mips64 and mipsel64 can be
done as partial ports. Only a small portion of packages (+depends)
would get ported, those that might require the larger address space
like postgresql.

Any of those ports could consist of maybe 300 packages only. Its
unreasonable to port more packages since they would be slower and
waste resources (ram, disk and mirror space). Forcing an s390x port to
have >90% ported before inclusion would be insane.

>>   - The mirror situation: [...] The tools for this exist now; what is
>>     left is creating a policy about this and switching the mirrors to
>>     the new system.
>> 
> OK, that's not an AMD64 problem.

Link please. Or is this realy a closed door developement?

>>   - Some technical AMD64 questions: ftpmaster had some specific
>>     questions about the AMD64 port they want to see answered.
>
> Which questions? Link please.
>  
>>     Also, an
>>     LSB person recently expressed some technical concerns (see [1]).
>
> Those have been answered. Some people may not like the answer, but IMHO
> it's good enough and, frankly, somewhat more sensible than the alternate
> solution.
>
>> [timeframe]
>
> Frankly, the complaints about the responsiveness of ftpmaster lately
> suggest that at this rate, the archive structure conversion and thus the
> amd64 inclusion won't happen for sarge. IMHO, the job needs more manpower.
>
> I'd think that you'd find some volunteers among the amd64 people...

Many aren't DDs but that still leaves some.

Also I'm sure the mirror restructuring will need changes to the DAK
which even non-DDs could help write and esspecially test on private
archives. Or will that be done by partially mirroring a complete
ftp-master archive?

MfG
        Goswin



Reply to: