I finally tracked down a bug that was causing an error when trying to
use rubygems and rake. When a .gemspec was being read and evaluated to
create a Gem::Specification instance, it was not creating a Gem::Version
object properly. Turns out that the Version.create() method was checking
to see if the parameter responds_to? :version. Somewhere along the line,
the String class added a version method - which caused the error.
Regards,
Jay Turpin
“Always do right. This will gratify some people and astonish the rest.”
- Mark Twain
From: Turpin, Jay
Sent: Tuesday, September 23, 2008 10:36 AM
To: ‘[email protected]’
Subject: RE: [Ironruby-core] question on RubyGems
Tomas - The problem seems to be coming from rubygems/custom_required.rb.
It redefines Kernel.required() and is not re-throwing the error
properly:
This works:
module Kernel
alias gem_original_require require # :nodoc:
def require(path) # :nodoc:
gem_original_require path
end
end # module Kernel
but this doesn’t:
module Kernel
alias gem_original_require require # :nodoc:
def require(path) # :nodoc:
gem_original_require path
rescue LoadError => load_error
raise load_error
end
end # module Kernel
Regards,
Jay Turpin
“It is always the best policy to tell the truth, unless, of course, you
are an exceptionally good liar.” - Jerome K. Jerome
From: [email protected]
[mailto:[email protected]] On Behalf Of Tomas M.
Sent: Tuesday, September 23, 2008 10:12 AM
To: [email protected]
Subject: Re: [Ironruby-core] question on RubyGems
I see. Well it seems more broken than I thought it is 
Tomas
From: [email protected]
[mailto:[email protected]] On Behalf Of Turpin, Jay
Sent: Tuesday, September 23, 2008 9:37 AM
To: ‘[email protected]’
Subject: Re: [Ironruby-core] question on RubyGems
Tomas - Thanks for the advice. However, it is still not working as
expected. When I run this ruby script:
require ‘rubygems’
require ‘rake’
puts $LOAD_PATH
It give me this output:
C:\temp>ir -D t.rb
C:/Projects/IronRuby/lib/ruby/site_ruby/1.8\rubygems/custom_require.rb:31:in
require': Bogus Error (RuntimeError) from C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\MatchCaller.Generated.cs:35:in Call3’
from
C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\CallSite.cs:275:in
UpdateAndExecute' from C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\UpdateDelegates.Generated.cs:45:in Update3’
from t.rb:3
from
C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\MatchCaller.Generated.cs:35:in
Call3' from C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\CallSite.cs:275:in UpdateAndExecute’
from
C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\UpdateDelegates.Generated.cs:45:in
`Update3’
from t.rb:3
But the error is actually coming from deep inside the rubygems library
Thanks!
Regards,
Jay Turpin
“In the end, everything is a gag.” - Charlie Chaplin
From: [email protected]
[mailto:[email protected]] On Behalf Of Tomas M.
Sent: Tuesday, September 23, 2008 9:30 AM
To: [email protected]
Subject: Re: [Ironruby-core] question on RubyGems
-D is actually what you need:
C:\Temp>rbd x.rb
C:\M1\ndp\fx\src\Core\Microsoft\Scripting\Actions\CallSite.cs:275:in
UpdateAndExecute' C:\M1\ndp\fx\src\Core\Microsoft\Scripting\Actions\UpdateDelegates.Generated.cs:38:in Update2’
x.rb:7:in bar' C:\M1\ndp\fx\src\Core\Microsoft\Scripting\Actions\MatchCaller.Generated.cs:30:in Call2’
C:\M1\ndp\fx\src\Core\Microsoft\Scripting\Actions\CallSite.cs:275:in
UpdateAndExecute' C:\M1\ndp\fx\src\Core\Microsoft\Scripting\Actions\UpdateDelegates.Generated.cs:38:in Update2’
x.rb:3:in foo' C:\M1\ndp\fx\src\Core\Microsoft\Scripting\Actions\MatchCaller.Generated.cs:30:in Call2’
C:\M1\ndp\fx\src\Core\Microsoft\Scripting\Actions\CallSite.cs:275:in
UpdateAndExecute' C:\M1\ndp\fx\src\Core\Microsoft\Scripting\Actions\UpdateDelegates.Generated.cs:38:in Update2’
x.rb:10
As I said, some internal frames are included as well, which is a bug
that I’m going to fix soon.
Tomas
From: [email protected]
[mailto:[email protected]] On Behalf Of Turpin, Jay
Sent: Tuesday, September 23, 2008 9:15 AM
To: ‘[email protected]’
Subject: Re: [Ironruby-core] question on RubyGems
Michael - Thanks for the quick response, but that’s not quite what I
meant. When I run Ruby and get an error, I get the complete stack trace
of the Ruby code:
C:\temp>ruby t.rb
c:/ruby/lib/ruby/site_ruby/1.8/rubygems/version.rb:283:in satisfy?': Bogus Error (RuntimeError) from c:/ruby/lib/ruby/site_ruby/1.8/rubygems/version.rb:273:in satisfied_by?’
from
c:/ruby/lib/ruby/site_ruby/1.8/rubygems/source_index.rb:86:in all?' from c:/ruby/lib/ruby/site_ruby/1.8/rubygems/version.rb:273:in each’
from c:/ruby/lib/ruby/site_ruby/1.8/rubygems/version.rb:273:in
all?' from c:/ruby/lib/ruby/site_ruby/1.8/rubygems/version.rb:273:in satisfied_by?’
from
c:/ruby/lib/ruby/site_ruby/1.8/rubygems/source_index.rb:187:in search' from c:/ruby/lib/ruby/site_ruby/1.8/rubygems/source_index.rb:185:in each’
from
c:/ruby/lib/ruby/site_ruby/1.8/rubygems/source_index.rb:185:in search' from c:/ruby/lib/ruby/site_ruby/1.8/rubygems/source_index.rb:165:in find_name’
from c:/ruby/lib/ruby/site_ruby/1.8/rubygems.rb:187:in
activate' from c:/ruby/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:32:in require’
from t.rb:2
However, when I run IronRuby and get an error, I only get one line of
ruby code and the rest is the C# stack trace:
C:\temp>ir t.rb
custom_require.rb:26:in require': Bogus Error (RuntimeError) from C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\MatchCaller.Generated.cs:35:in Call3’
from
C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\CallSite.cs:275:in
UpdateAndExecute' from C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\UpdateDelegates.Generated.cs:45:in Update3’
from :0
from
C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\MatchCaller.Generated.cs:35:in
Call3' from C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\CallSite.cs:275:in UpdateAndExecute’
from
C:\Projects\IronRuby\src\Microsoft.Scripting.Core\Actions\UpdateDelegates.Generated.cs:45:in
`Update3’
from :0
The error really happens deep inside the Ruby code. Any ideas?
Thanks
Regards,
Jay Turpin
“Over 48.7% of all statistics are useless.” - Anonymous