[Pdx-pm] svn4cpan thinkfest

[Eric Wilhelm]

Hi all,

Since we said after the last meeting that this has to happen within one 
or two weeks, I guess that leaves us with sometime between now and 
about 8pm on Tuesday.

How's Tuesday afternoon running into Tuesday night sound?  Urban Grind 
in the Pearl?

I say "thinkfest" and not hackfest because I haven't looked at this code 
for over a year.  Thus, there will be some code review necessary.  
There are also several things I haven't solved yet, such as what the 
mirroring and commit-bit management involves.  I'm sure hacking will 
happen, but expect to be prototyping rather than actually finishing.

Also, the backpan import being a 17 hour process means we need to come 
up with a faster way to find all of the edge cases (which may involve 
the 6 CPUs in these 4 not-racked-yet boxen I'm sitting on.)  Once we 
know the code will survive every tarball, zipfile, and mislabeled 
shovel, it's no problem to make a fresh svn and let it spin for a day.  
I think the linearity is worth the extra wait.

Of course, it's possible that the hard limit of 3 commits/second exists 
only in the silly svn client code.  Possibly an svk-based import would 
solve that.

There may also be discussion of OSCON proposals, unless somebody wants 
to host that (and possibly involve booze) on another day.

--Eric
-- 
Moving pianos is dangerous.
Moving pianos are dangerous.
Buffalo buffalo buffalo buffalo buffalo buffalo buffalo.
---------------------------------------------------
    http://scratchcomputing.com
---------------------------------------------------