On Feb 23, 2011, at 00:47 , James Tucker wrote:

>> This I can repro:
>> 
>> 10020 % ruby -Ilib -rubygems -e 'Gem.activate("rails", "< 3.0.0.beta"); 
>> require "active_support/ordered_hash"'
>> ./lib/rubygems.rb:265:in `activate': Unable to activate activesupport-3.0.4, 
>> but rails-2.3.9.pre depends on activesupport (= 2.3.9.pre, runtime) 
>> (Gem::LoadError)
>>      from ./lib/rubygems.rb:198:in `try_activate'
>>      from ./lib/rubygems/custom_require.rb:31:in `require'
>>      from -e:1
>> 
>> and I'm looking into it now...
> 
> Verified against the real use case and you got my test passing.
> 
> Many thanks!

No. I never got your test passing. I verified that the test would never pass on 
1.5 or 1.6 and removed it. I think a majority of the fault lies with the test 
infrastructure. We apparently don't have a good/clean way of setting up this 
type of test very well... tho I don't see why we don't by this time.

We should revisit this deficiency for 1.7.

_______________________________________________
Rubygems-developers mailing list
http://rubyforge.org/projects/rubygems
Rubygems-developers@rubyforge.org
http://rubyforge.org/mailman/listinfo/rubygems-developers

Reply via email to