hgs at dmu.ac.uk
Fri Feb 1 07:02:00 EST 2008
On Fri, 1 Feb 2008, Eric Hodel wrote:
> On Feb 1, 2008, at 02:23 AM, Hugh Sasse wrote:
> > On Thu, 31 Jan 2008, Mark Hubbart wrote:
> >>>> Plus one might not have a rubyforge project.
> > [...]
> >> The "real" project could be on sourceforge.org, or code.google.com.
> > Maybe there needs to a directive to say whwere the project is?
> There is also a #homepage= for Gem::Specification, but with the
> current gem repository, this is a bikeshed objection as under 1% of
> gems are published without a rubyforge project.
I'd agree the default is correct most of the time, but there are a
lot of gems now, so 1% is not insignificant. I don't see that
arguing for choice is a bikeshed argument. If you argue that too
much code will have to change to support this, that it is the thin
end of a long wedge (multiple mirror sites will come next!) that it
would make the gemspec more brittle as it depends on more
information, or that the suggestion is otherwise malformed, then
fair enough. It was a suggestion offered in the hope that someone
might say "I have a better idea:...". Is there a wishlist to which
this might be added? Maybe *I'm* taking your use of the word
"bikeshed" too seriously! I'm really only trying to help suggest a
"fix" for a known edge case, I'm not saying things are fundamentally
wrong, because clearly gems have been marvellous for ages.
More information about the Rubygems-developers