Help with require in 1.9.2 p0

[Michel Demazure]

I have this strange bug with a chain of requires, after a certain number
of hops. It is independant of the file to require. Even in the following
case with the same file twice

<processor.rb>
6 require 'slot.rb'
7 puts "yes"
8 require 'slot.rb'

the first require works, "yes" is printed, and the second require to the
*same* file bugs :

<internal:lib/rubygems/custom_require>:29:in `require': can't convert
nil into String (TypeError)
yes
from <internal:lib/rubygems/custom_require>:29:in `require'
from F:/dev/ruby/Capucine/lib/processor.rb:8:in `<top
(required)>'
from <internal:lib/rubygems/custom_require>:29:in `require'
from <internal:lib/rubygems/custom_require>:29:in `require'
from F:/dev/ruby/Capucine/lib/manager.rb:6:in `<top (required)>'
from <internal:lib/rubygems/custom_require>:29:in `require'
from <internal:lib/rubygems/custom_require>:29:in `require'
from F:/dev/ruby/Capucine/lib/main.rb:10:in `<main>'

This bug appears also with another chain (using rspec), this time
refusing files which are correctly loaded in the previous case.

Maybe this is for ruby-core, but I have no access to it.

_md

 

[Michel Demazure]

Michel Demazure wrote in post #949797:

I have this strange bug with a chain of requires, after a certain number
of hops. It is independant of the file to require.

Actually, it is due to some method_missing hack in a previously loaded
file, which worked under 1.9.1, and no more in 1.9.2.

My bad !

_md