Wed Jan 17 16:11:13 PST 2007 Kevin Turner tagged release-2.0.0-b0 Wed Jan 17 16:10:43 PST 2007 Kevin Turner * Set version number to 2.0.0-b0 Wed Jan 17 16:09:43 PST 2007 Kevin Turner * LICENSE: really added this time Wed Jan 17 16:00:38 PST 2007 Kevin Turner * consumer.discover: add module-level docstring Wed Jan 17 16:00:03 PST 2007 Kevin Turner * consumer.consumer: minor doc tweaks Wed Jan 17 15:58:23 PST 2007 Kevin Turner * setup.py: remove reference to COPYING, replace with Apache license boilerplate. Wed Jan 17 15:47:35 PST 2007 Kevin Turner * admin/combo-setup.py: goodbye (other packages are now openid subpackages) Wed Jan 17 13:05:41 PST 2007 Kevin Turner * examples/consumer: unicode, fixes #2284 ask for UTF-8 in the form element, and decode as UTF-8 when building self.query. Mon Jan 15 16:11:51 PST 2007 Kevin Turner * openid.server: documentation updates Mon Jan 15 15:42:54 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 15:31:58 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 15:30:13 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 15:29:59 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 15:28:15 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 15:28:09 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 15:27:51 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 15:26:41 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 15:26:29 PST 2007 Kevin Turner * epydoc syntax Mon Jan 15 14:51:44 PST 2007 Kevin Turner * Apache License Mon Jan 15 13:54:24 PST 2007 Kevin Turner * examples/server: server advertises itself as v2, so replace openid:Delegate with LocalID Mon Jan 15 13:41:18 PST 2007 Kevin Turner * consumer.GenericConsumer._verifyDiscoverySingle: use endpoint.getLocalID, closes #2258. Fri Jan 12 18:47:43 PST 2007 Kevin Turner * examples/server: reorganize to pass endpoint url to server.Server. a chicken-and-egg situation here, httpserver wanted oidserver to be constructed first, and oidserver wanted httpserver to be constructed first. Is that going to be a common problem? Fri Jan 12 18:43:24 PST 2007 Kevin Turner * examples/server.main: remove unused code Fri Jan 12 18:42:00 PST 2007 Kevin Turner * consumer.discover.OpenIDServiceEndpoint.__str__: added Fri Jan 12 18:35:35 PST 2007 Kevin Turner * x is y is z, wtf? Fri Jan 12 17:35:19 PST 2007 Kevin Turner * [API] server.CheckIDRequest.answer: include op_endpoint in response [#2221] server.Server: take op_endpoint in constructor server.CheckIDRequest.fromMessage: also take Server The story: There were several things about the URL argument when it first appeared in the signature to CheckIDRequest.answer. It was only required in this weird little corner case, and we weren't really sure how it was going to be used, and user_setup_url might have been different every time. So I added it to answer() and made it optional except in the case where it actually needed to be there, and if people thought it was weird, well, maybe they'd chime in and help remove it from the protocol. Then, based on feedback from library users, it changed from being the user_setup_url to just the server URL, which we would use to construct a plausable-looking user_setup_url from, as nobody was going to do anything else with it anyway. But it was still optional but for that one corner case. Now we need the URL for *every* id_res response, and it's the server's endpoint URL, not some weird poorly-understood thing that might vary with each request. Given that it's the server's endpoint URL, I'm putting it on the Server object, and doing some internal reference-passing where needed. (I hope I didn't create any reference cycles, but given that Servers don't hold references to their Requests, I'm pretty sure we're okay.) The result: Server() takes a second parameter, op_endpoint, which is optional for backwards compatibility but is actually required if you ever want to reply to a version 2.0 message. It uses warnings.warn to throw a warning if you leave it out. OpenIDRequest.fromMessage takes a second parameter, server, which most of the concrete implementations of OpenIDRequest don't use, but is not optional for CheckIDRequest. This does break compatibility, but I'll be pretty surprised if anyone is using fromMessage() directly instead of going through Decoder.decode(). Fri Jan 12 15:02:58 PST 2007 Kevin Turner * server.server.CheckIDRequest.answer: immediate mode negative response is now mode=setup_needed Fri Jan 12 14:30:04 PST 2007 Kevin Turner * [API] server.server.CheckIDRequest.answer: add a claimed_id parameter. This is an API change; please review on #2222. Fri Jan 12 11:58:40 PST 2007 Kevin Turner * server.server.CheckIDRequest.answer: include claimed_id in response Fri Jan 12 11:49:16 PST 2007 Kevin Turner * server.server.CheckIDRequest.fromMessage: add a claimed_id attribute Fri Jan 12 11:32:29 PST 2007 Kevin Turner * server.server.CheckIDRequest: mention that trust_root AKA realm in the docstring. Fri Jan 12 11:23:54 PST 2007 Kevin Turner * flag some undocumented exceptions with FIXME Fri Jan 12 11:20:38 PST 2007 Kevin Turner * test.test_server: add claimed_id to where it was missing from some OpenID 2 messages Thu Jan 11 17:11:22 PST 2007 Kevin Turner * test_server.TestCheckID: add some FIXMEs Wed Jan 10 16:02:08 PST 2007 Josh Hoyt * Added test for return_to missing when parsing the nonce from an OpenID 1 response Wed Jan 10 15:52:37 PST 2007 Josh Hoyt * Change consumer.ServerError API to be a little cleaner Wed Jan 10 15:35:30 PST 2007 Josh Hoyt * Removed code in consumer that was trying to handle None returns from fetchers The fetcher API stopped supporting None return values a while ago Wed Jan 10 15:32:07 PST 2007 Josh Hoyt * Added test for complete's OpenID 1 setup needed handling Wed Jan 10 15:26:31 PST 2007 Josh Hoyt * Added test for discovery returning no services Wed Jan 10 15:18:43 PST 2007 Josh Hoyt * Add a command that finds all of the Python source that constitutes the library, excluding tests. Wed Jan 10 15:13:25 PST 2007 Josh Hoyt * Added test for HTTP errors when fetching in the high-level consumer, and refined the implementation Wed Jan 10 15:12:13 PST 2007 Josh Hoyt * Added more discovery verification tests Wed Jan 10 15:11:05 PST 2007 Josh Hoyt * Clean up verification test to make it easier to understand Tue Jan 9 16:40:14 PST 2007 Josh Hoyt * organization of private methods in consumer.consumer Tue Jan 9 16:24:20 PST 2007 Josh Hoyt * Add test for using pre-discovered information in discovery for OpenID2 Tue Jan 9 16:16:20 PST 2007 Josh Hoyt * Add discovery verification tests and break up the logic in _verifyDiscoveryResults into a function for OpenID1 and a function for OpenID2 Tue Jan 9 15:20:44 PST 2007 Josh Hoyt * Implement OpenID 2.0 Auth spec svn rev 277, openid.mode=setup_needed Replaces bare "id_res" response for cancellation of immediate mode Tue Jan 9 14:49:16 PST 2007 Josh Hoyt * Add support for OpenID 2.0's op_endpoint URL, bare responses, and discovery verification when there is a mismatch between discovered information Tue Jan 9 14:46:51 PST 2007 Josh Hoyt * Change the discover function in the consumer's discovery module to work for either kind of identifier Tue Jan 9 14:46:15 PST 2007 Josh Hoyt * Added a supportsType method with smarts for OP identifiers Tue Jan 9 13:26:41 PST 2007 cygnus@janrain.com * API CHANGE: Reinstate SuccessResponse.extensionResponse() and implement support for requiring args to be signed Mon Jan 8 16:51:31 PST 2007 Josh Hoyt * Fixed (and tested) behaviour of OP identifier service endpoints and compatibility mode Mon Jan 8 16:53:05 PST 2007 Josh Hoyt * Added fromOPEndpointURL to OpenIDServiceEndpoint Thu Jan 4 17:22:17 PST 2007 Josh Hoyt * whitespace Thu Jan 4 17:17:39 PST 2007 Josh Hoyt * Switched a lot of the internal logic in Consumer to raise ProtocolErrors instead of returning FailureResponse and/or logging This change is not as complete as it could be. It's possible that users will want to catch ProtocolErrors, so maybe we should not be converting it to a FailureResponse Thu Jan 4 16:37:56 PST 2007 Josh Hoyt * raise ValueError -> raise ProtocolError in consumer internals Mon Jan 8 14:58:45 PST 2007 cygnus@janrain.com * Fix delArg test Fri Jan 5 21:40:16 PST 2007 Kevin Turner tagged heraldry-493420