This is not really fakeweb's fault, but perhaps there's another way to
handle Basic Auth... If the username is an e-mail address (which is
used more and more frequently, then the Standard Library URI parse
croaks on the URI passed in.

Here's how I reproduce this:

$ gem list -l fakeweb
fakeweb (1.2.4)
r$ irb -rubygems -r fakeweb
irb(main):001:0> FakeWeb.register_uri(:get, "http://
some...@example.org:test@localhost:3000/foobar.html", :body => "")
URI::InvalidURIError: the scheme http does not accept registry part:
some...@example.org:test@localhost:3000 (or bad hostname?)
        from /opt/local/lib/ruby/1.8/uri/generic.rb:195:in `initialize'
        from /opt/local/lib/ruby/1.8/uri/http.rb:78:in `initialize'
        from /opt/local/lib/ruby/1.8/uri/common.rb:488:in `new'
        from /opt/local/lib/ruby/1.8/uri/common.rb:488:in `parse'
        from /opt/local/lib/ruby/gems/1.8/gems/fakeweb-1.2.4/lib/fake_web/
registry.rb:95:in `normalize_uri'
        from /opt/local/lib/ruby/gems/1.8/gems/fakeweb-1.2.4/lib/fake_web/
registry.rb:21:in `register_uri'
        from /opt/local/lib/ruby/gems/1.8/gems/fakeweb-1.2.4/lib/fake_web.rb:
113:in `register_uri'
        from (irb):1

On Jul 2, 10:04 pm, Ken Mayer <li...@bitwrangler.com> wrote:

Hey Ken, thanks for posting. I think what you were trying to do is not
quite valid, but I wrote a test for the actual valid approach and it
doesn't work either :) oops. So I just fixed that:
http://github.com/chrisk/fakeweb/commit/1ad638500be9ddd570c6a888e5b34...

Here's what you should do once that's released (in 1.2.5):

# percent-encode unsafe characters in the userinfo part, per RFC 3986
FakeWeb.register_uri(:get, "http://someone
%40example.org:test@localhost:3000/foobar.html", :body => "")

http = Net::HTTP.new("localhost", 3000)
request = Net::HTTP::Get.new("/foobar.html")
request.basic_auth("some...@example.org", "test")  # but no need to
escape here, since it's not a URI
http.request(request)

That should do the trick. I'll try to release 1.2.5 soon -- I need to
refactor the ugly implementation in that commit a bit, and write some
more tests.

Thanks again,
Chris

Chris,

Thanks for the update!

Aloha a hui hou,
Ken
--
Ken Mayer / kma...@bitwrangler.com / 808-722-6142 / http://www.bitwrangler.com/

On Jul 2, 2009, at 8:49 PM, Chris Kampmeier wrote: