summaryrefslogtreecommitdiff
path: root/paste
Commit message (Collapse)AuthorAgeFilesLines
* assert send_http_response was triggered by an HTTPException, incase it's being0.9.7pjenvey2006-08-111-0/+2
| | | | used incorrectly
* o catch_errors_app and _wrap_app_iter_app were not in sync with thepjenvey2006-08-111-3/+6
| | | | | | | exception catch_errors_app was specified to catch, causing the wrapper to consume all exceptions o avoid allowing _wrap_app_iter_app to close() its wrapped iter twice in certain situations
* RegistryManager was close()ing the app_iter, when it needs to simply ↵pjenvey2006-08-111-2/+0
| | | | propagate it up the chain
* the catching_iter method is no longer used; removing itpjenvey2006-08-111-24/+0
|
* Fixed issue with attempting to clean-up the Registry environment twice under ↵bbangert2006-08-101-8/+7
| | | | certain exception tossing conditions.
* Fixed issue with Registry still referring to the wrong place if an app down ↵bbangert2006-08-091-11/+7
| | | | the chain uses Registry and returns an iter. Apps returning iter's should hold their own ref to any objects that used a StackedObjectProxy.
* doc format fix; better error message from Jamie Wilkinsonianb2006-08-031-0/+2
|
* having StackedObjectProxy.__repr__ avoid raising an exception when itpjenvey2006-08-031-1/+6
| | | | does not contain an internal object
* only displaying EvalException's full traceback button when the full tracebackpjenvey2006-07-311-7/+19
| | | | differs from the regular (includes hidden frames)
* evalexception was using the wrong contet-type, probably a typopjenvey2006-07-291-1/+1
|
* o no longer exhausting the app_iterpjenvey2006-07-271-5/+0
| | | | | o as much as I like the new StatusBasedForward forward debug message, it can be chatty with a lack of favicon. removing it
* o use response.HeaderDict instead of wsgilib.ResponseHeaderDictpjenvey2006-07-272-2/+3
| | | | o fixed the misleading ResponseHeaderDict deprecation warning
* cleaner gathering of the debug config setting via converters.asboolpjenvey2006-07-261-4/+2
|
* StatusBasedForward will consume the enclosed app's response when the Forwardpjenvey2006-07-261-0/+4
| | | | condition is met, but wasn't close()ing it when it's iterable
* renamed the multidict class to MultiDictpjenvey2006-07-253-16/+16
|
* commenting out make_repost_button's unused block of codepjenvey2006-07-241-0/+3
|
* o request.parse_formvars changes:pjenvey2006-07-243-56/+35
| | | | | | | | | | | | - now returns a multidict instead of dict (like parse_dict_querystring) - removed all_as_list keyword arg: use multidict.mixed() instead to produce an all_as_list dict - preventing FieldStorage from parsing out QUERY_STRING during GET/HEAD requests - the fake_out_cgi check wasn't careful enough about multipart/form-data content-types. Was also missing a StringIO import (r5471 broke file uploads) o evalexception's wsgiapp decorator assumed parse_formvars would return GET vars during a GET (buggy behavior), now it explicitly asks for them
* added copy methodpjenvey2006-07-231-1/+4
|
* Updating HTTPNotModified so it works with ETag. This still needs to be fixed ↵bbangert2006-07-221-1/+2
| | | | to require either an ETag or a Date header.
* Some more changes to paste.errordocuments and an update to paste.recursivethejimmyg2006-07-202-150/+325
|
* Fixed a bug in paste.errordocuments.forward and updated unit teststhejimmyg2006-07-201-0/+1
|
* Removed errordocument.empty_error, since it never actually worked in a ↵0.9.5ianb2006-07-201-47/+0
| | | | useful way; changed the news to 0.9.5 (since that's what the next release will be)
* Fixed problem reported by Charlie Groves with TestApp file uploads with ↵ianb2006-07-201-1/+1
| | | | length-two tuples
* Keep wsgi.input from being eaten when the cgi module is looking for POST ↵ianb2006-07-201-1/+15
| | | | variables
* having WSGIResponse.params return a multidictpjenvey2006-07-201-2/+2
|
* Changed usage of MultiDict to util.multidictianb2006-07-192-32/+6
|
* added WSGIThreadPoolServer. Provides much better performance thanpjenvey2006-07-191-9/+146
| | | | | | | | WSGIServer. paste.httpserver.serve now uses the thread pooled server by default. Users can set the number of worker threads (defaults to 10) or switch back to the normal WSGIServer with config file toggles. submitted by: james@jamestaylor.org (Thanks!) resolves: #112
* reverting r5459, as the issue has already been fixedpjenvey2006-07-171-5/+2
|
* Added tests for file-related conditional requests. Added support for ↵ianb2006-07-172-1/+22
| | | | If-None-Match to DataApp. Added parsing support for multi-value headers.
* Added a 'no' argument to mustcontainianb2006-07-151-1/+17
|
* Added a middleware to clear out error bodies, making them more accessible to ↵ianb2006-07-142-1/+96
| | | | Apache; added an app_iter wrapper for chaining app_iters from multiple sources (needed for peeking at status)
* Added a Paste Deploy entry point for paste.errordocument (not tested, bad me)ianb2006-07-141-1/+23
|
* More permissive parsing for dates in headers. IE in particular will add '; ↵ianb2006-07-131-1/+4
| | | | length=200' to If-Modified-Since, and that was producing BadRequest errors. Instead parameters on dates are now simply ignored
* Avoid a weird upstream traceback in certain failure situationsianb2006-07-111-1/+1
|
* Added a setvars method, to complement setvar/varianb2006-07-111-0/+9
|
* Patch from Brad Clements to add Paste Deploy support for paste.auth methodsianb2006-06-303-0/+55
|
* o revised the r5420 patch to be specific to the _IfModifiedSince headerpjenvey2006-06-261-4/+10
| | | | o added a test to reproduce ticket #114
* Clean date headers of extra attributes sent by IEpjenvey2006-06-261-2/+4
| | | | for: ticket #114
* Fix for paste.urlmap and vhostsianb2006-06-221-2/+2
|
* Added a method to get a cgi-style dictionary out of this dictionaryianb2006-06-221-2/+27
|
* Make multidict parse args+kw more like dict() doesianb2006-06-221-6/+3
|
* Made it possible to create the session after start_response is called but ↵ianb2006-06-211-4/+15
| | | | before the first content is sent to the server
* Added another app_iter wrapper similar to add_close, except that also calls ↵ianb2006-06-211-0/+39
| | | | a function just before the first content is returned
* Put in a test to keep paste.testing_variables from clobbering an attribute ↵ianb2006-06-211-0/+5
| | | | of the response object
* made MultiDict iterablepjenvey2006-06-201-2/+2
|
* Preventing circumvention of StaticURLParser and PkgResourcesParser's root ↵pjenvey2006-06-191-6/+24
| | | | | | directories. Ensuring fix with tests
* Fixed urlparser classes to handle quoted (e.g. %20) URL paths.pjenvey2006-06-181-3/+7
| | | | | Added tests for this (Thanks Jose Galvez)
* fixed the logged HTTP protocol, now grabbing it from HTTP_VERSION ↵pjenvey2006-06-181-1/+1
| | | | environment var
* fixed TransLogger to log to the Apache combined log format as advertisedpjenvey2006-06-181-1/+9
| | | | | submitted by: http://nate.myopenid.com/ resolves: #118
* added optional host and max_children args to scgi's serve_application.pjenvey2006-06-171-2/+28
| | | | | | | | | | | | | | | Ideally the function prototype would match SCGIServers, i.e.: def serve_application(application, prefix, host=, port=None, max_children=None): or: def serve_application(application, prefix, **kwargs): (host, port and max_children could then default to SCGIServer's defaults), but that could break compatibility with those not explicitly using keyword args suggested by: greg@electricrain.com resolves: #55
View Lorry Controller Status