| Age | Commit message (Collapse) | Author |
|
|
|
This means that it can be overriden, and is no longer static. With great
power comes great responsibility...
This also allows the user to insert view derivations over the
rendered_view. This would allow the same sort of functionality as
the BeforeRenderer event, by allowing the derivation to modify the
Response/dictionary that the view returned.
|
|
Specifically we want to make sure that rendered_view always comes first,
unless a user overrides it by name.
|
|
So we adopt over/under from tweens. Now it becomes simple to
think/reason about stuff again =)
|
|
|
|
Instead of passing in a default to the view derivation, we attempt to
pull it out of the options kwarg, if that doesn't exist, then we pass in
the default.
This should make it simpler to program view derivations because now you
don't need to check to see if the kwarg is different from the default.
|
|
|
|
|
|
Specifically predicated_view and attr_wrapped_view exist for MultiView
to function, without them the view machinery doesn't work. They aren't
technically view derivations because they are just containers for some
data that the view machinery requires.
|
|
|
|
|
|
|
|
This way we can use that in tests.
|
|
This is an integration test in the sense that it adds both a view
derivation and then adds a view. That view is then tested to verify that
it correctly returns and that the derivations haven't broken anything.
|
|
We also use the previously defined names() method on PredicateList to
fetch the list of predicates and pull out all predicate values into it's
own dictionary.
Then the rest of the values are the view options for the view
derivations. At the moment we do not check if the options matches a
valid view derivation. No seatbelts for view options.
|
|
This is required for future changes, where we need the list of predicate
names.
|
|
|
|
|
|
|
|
feature/configurable-view-deriver
|
|
Fix python 3.5 support
|
|
populate pshell's env_help with docstring of values
|
|
Mention Circus as a pserve --daemon replacement. Refs #1554, #1641
|
|
|
|
Python 3.5 inserts an extra stackframe when calling extract_traceback,
this adds in a check to see if the last frame is this extra frame, and
if so removes it. This doesn't affect any other version of Python.
|
|
|
|
|
|
|
|
Use entry points for pshell
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This way they are valid according to the RFC, and newer versions of
WebOb won't complain.
|
|
Due to changes in WebOb we may no longer create a Response() object with
an invalid status (i.e. 'None None'). For the top-level HTTPException if
it is not correctly overriden in child classes then we now use '520
Unknown Error'.
While not an official RFC error, this has been used by Microsoft's Azure
as well as CloudFlare to signify an error that doesn't have a more
appropriate error message.
|
|
- Use pyramid.compat.input_ instead of manually checking
for python version and chosing the appropriate
input/raw_input function
- Renamed--force-conflicting-name option to --ignore-conflicting-name
- Display invalid scaffold names in a sorted manner
|
|
- Made sure str.format() calls are py2.6 compatible
- Removed test_scaffolds/test_init.py#test_pre_site
since the check is handled in pcreate script
|
|
- Use input() in python3 and raw_input in python 2.x
- Remoced unused local variables
|
|
|
|
- Check for importable packages/modules with the same
name as the package to be created
- Added --force-conflicting-name option to enforce the
use of a conflicting package
- Added appropriate tests
|
|
Return concrete classes from 'pyramid.httpexceptions.exception_response'
|
|
The base classes are not appropriate for 400 and 500 status codes.
See: #1832
|
|
request_iface to _call_view; we don't want _call_view to use
request.request_iface, because render_view_to_response and friends are
pretty much limited to finding views that are not views associated with
routes, and the only thing request.request_iface is used for is to find
route-based views. The render_view_to_response API is (and always has
been) a stepchild API reserved for use of those who actually use
traversal. Doing this fixes an infinite recursion bug introduced in
Pyramid 1.6a1, and causes the render_view* APIs to behave as they did in
1.5 and previous. We should probably provide some sort of different API
that would allow people to find views for routes.
Fixes #1643.
|
|
|
|
|
|
|