Commits on Jul 16, 2009
  1. Implemented :body accepting Proc object that yields the match objects…

    … when passing a regexp as :uri
    committed Jul 16, 2009
  2. Ignore Rubymine meta-info

    committed Jul 16, 2009
Commits on Jul 13, 2009
  1. Add a test for some accidental functionality: you can register respon…

    …se headers using string option keys, too
    chrisk committed Jul 13, 2009
Commits on Jul 9, 2009
  1. Release 1.2.5

    chrisk committed Jul 9, 2009
  2. Update CHANGELOG for 1ad6385

    chrisk committed Jul 9, 2009
Commits on Jul 6, 2009
  1. More concise tests

    chrisk committed Jul 6, 2009
  2. URI registry: remove the auto-expanding hash idiom, since the hash ne…

    …eded here is just singly nested
    chrisk committed Jul 6, 2009
Commits on Jul 5, 2009
Commits on Jul 4, 2009
  1. Nix the part of 1ad6385 that attempted to handle userinfo strings wit…

    …h more than one colon in them, which doesn't seem to be legal anyway
    The RFC doesn't mention this, but either way, it's not possible to reliably determine which part is the username and which is the password when handling the request, so the only way would be to iterate through the possible permutations. Yuck.
    chrisk committed Jul 4, 2009
Commits on Jul 3, 2009
  1. Fix that you couldn't register and request URIs if the userinfo part …

    …included escaped unsafe characters
    chrisk committed Jul 3, 2009
Commits on Jun 29, 2009
  1. Update CHANGELOG for 2c75fa6 and f64af6a

    chrisk committed Jun 29, 2009
  2. Fix that exact string and URI matches should always take precedence o…

    …ver regex matches when determining which response to use
    chrisk committed Jun 29, 2009
  3. Add some test coverage for response precedence when you register iden…

    …tical strings and regexes
    chrisk committed Jun 29, 2009
  4. Refactor the URI registry's #responses_for to always return an Enumer…

    …able (instead of returning nil when there are no matches)
    chrisk committed Jun 29, 2009
Commits on Jun 28, 2009
  1. Refactor tests so assertions aren't made inside Net::HTTP blocks.

    That way, if a Net::HTTP method doesn't yield, the assertions won't be silently skipped. Hat tip to Roman2K for pointing this out.
    chrisk committed Jun 28, 2009
  2. Remove old, unneeded TODO comment

    chrisk committed Jun 28, 2009
  3. Test cleanup: define a common setup and teardown that prevents shared…

    … state for all tests
    chrisk committed Jun 28, 2009
  4. Change request handling so it raises if more than one registered rege…

    …x matches a request URI. Less confusing this way.
    chrisk committed Jun 27, 2009
Commits on Jun 27, 2009
  1. Add missing test coverage to ensure that blocks attached to Net::HTTP…

    … request methods are yielded with the registered response body
    chrisk committed Jun 27, 2009
Commits on Jun 26, 2009
  1. Start refactoring regex support: just use one URI registry hash, inst…

    …ead of having a duplicate just for regexes
    chrisk committed Jun 26, 2009