In order to use the new assert mechanism that returns explicit assertion failure message, all the default matchers does not return a boolean, but only do an assert statement with a basic assertion message (value_1 != value_2).
The requests_match function has been refactored to use the 'get_matchers_results' function in order to have explicit failures that are logged if any.
Many unit tests have been changed as the matchers does not return a boolean value anymore.
Note: Only the matchers "body" and "raw_body" does not have an assertion message, the body values might be big and not useful to be display to spot the differences.
The function returns two list:
- the first one is the list of matchers names that have succeeded.
- the second is a list of tuples with the failed matchers names and the related assertion message like this ("matcher_name", "assertion_message").
If the second list is empty, it means that all the matchers have passed.
A matcher can now return other results than a boolean :
- An AssertionError exception meaning that the matcher failed, with the exception we get the assertion failure message.
- None, in case we do an assert in the matcher, meaning that the assertion has passed, the matcher is considered as a success then.
- Boolean that indicates if a matcher failed or not. If there is no match, a boolean does not give any clue what it is the differences compared to the assertion.
HeadersDict is a subclass of CaseInsensitiveDict with two new features:
1. Preserve the case of the header key from the first time it was set.
This means that later munging won't modify the key case. (You can
force picking up the new case with `del` followed by setting.)
2. If the value is a list or tuple, unpack it and store the first
element. This is the same as how `Request.add_header()` used to work.
For backward compatibility this commit preserves `Request.add_header()` but
marks it deprecated.
This is an actual bugfix. If before_record_response is passed into VCR as
an iterable then it won't be included in filter_functions. This commit
repairs the logic to separate the tests (just as it's already done for
before_record_request).
Also use .extend() rather than looping on .append()
It shouldn't matter whether the request body comes from a file or a
string, or whether it is passed to the Request constructor or assigned
later. It should always be stored internally as bytes.
