Jim Weirich wrote:
> 
> On Jul 8, 2008, at 4:45 PM, Sam Ruby wrote:
> 
>> It is not just prodding them with a stick - we have gem maintainers
>> who haven't integrated patches, or even responded to email.
>>
> 

This might be far-fetched, but ...

If the licensing allows it, why wait for a gem developer/administrator to
approve and merge patches/fixes?

1. Create 1.9 git repos from latest versions of gems (or some other
distributed version control system).
2. Create a central registry of git repos for each gem.
3. Change rubygems to be able pull from git repos.
4. Change rubygems to use a SHA1 hash as a version string.
5. People who use and fix the gems can share their efforts, without waiting
for someone by posting a new git repo location and version string for the gem.
6. For the security conscious, gem installer can verify a gem source's SHA1
hash that has been signed against a gem's known public key.  There could be a
--verify mode that would only pull and install gems that can be verified
against the gem' trusted public key.

Once the majority of important gems are working under 1.9 we can all go back
to dragging our feet with SVN. :)

Nothing is more annoying than abandon-ware,
Kurt Stephens