Mail Archives: djgpp-workers/2006/06/12/16:51:38
Hello.
decker AT dacafe DOT com wrote:
[snip]
>> I estimate option 2 will require a day a week of someone's time for two
>> to three months -- maybe more. 50% of this time will be "unproductive"
>> in the sense of dealing with mail, getting information on issues,
>> rebuilding things.
>
>> Any volunteers?
>
> Would it really be a day a week for months like you say?
That was based on my experience of being 2.04 release manager for the
best part of a year during 2002/2003.
> 1) The core has not changed much in last several months, so regular
> re-compilations of the toolchain unneccessary
Looking at the list of issues, I see a couple related to symlinks. Say
you need to update libc to fix those issues. You will need to rebuild
several components, e.g.: coreutils, perl, automake.
Why more than just coreutils? Integration testing. That's a time
consuming task. The automake test suite took about 1.5 hours on my
Athlon 850MHz last time I ran it.
> 2) Email? What Email? What if the RM 'has no email' :))
The release manager can't do their job in isolation.
A number of the issues will require the help and knowledge of
djgpp-workers to diagnose, test and resolve.
Any changes to DJGPP will need reviewing by the djgpp-workers, before
they can be committed.
> 3) At this point half a job is better than no job at all (i.e. current
> status). The emphasis should be on familiarity with the issues and the
> team that has volunteered to work on them. There *is* obviously *some*
> efforts made on individual basis here. What is needed is some project
> management and visibility.
[snip]
And someone driving it forward. If it feels like someone is driving 2.04
forward, I'm sure more will happen.
Bye, Rich =]
--
Richard Dawe [ http://homepages.nildram.co.uk/~phekda/richdawe/ ]
"You can't evaluate a man by logic alone."
-- McCoy, "I, Mudd", Star Trek
- Raw text -