* Support for CloudFormation stack resource deletion via backend resource method
delete_from_cloudformation_json() via parse_and_delete_resource().
* Correction to the inappropriate inclusion of EndingSequenceNumber in open shards.
This attribute should only appear in closed shards. This regretfully prevents
confirmation of consistent record counts after split/merge in unit tests.
* Added parameters/decorator to CloudFormationModel method declarations to calm-down Pycharm.
Co-authored-by: Joseph Weitekamp <jweite@amazon.com>
* First cut of S3 Cloudformation Update support: encryption property.
* Update type support for S3. Abstract base class for CloudFormation-aware models, as designed by @bblommers, introduced to decentralize CloudFormation resource and name property values to model objects.
* Blackened...
* Un-renamed param in s3.models.update_from_cloudformation_json() and its call to stay compatible with other modules.
Co-authored-by: Bert Blommers <bblommers@users.noreply.github.com>
Co-authored-by: Joseph Weitekamp <jweite@amazon.com>
Co-authored-by: Bert Blommers <info@bertblommers.nl>
* fix sqs tag list from cloudformation resource creation
the method `create_from_cloudformation_json` of the Sqs resource
does not handle the difference of format of the Tags field in the
resource template and the format expected in Sqs resource class.
In cfn resource template Tags is specified as a list of dicts. But
the Sqs resource expects that the tags field be a single dict.
This behaviour causes a crash when a queue is created with tags
from `create_from_cloudformation_json` and later the list_queue_tags
is called because it tries to call `items` from `queue.tags` but
tags is actually a list of dicts.
* fix comment
* fix linter
* minor
Co-authored-by: Hudo Assenco <hudo.assenco@gmail.com>
* Keep order in request body to ensure auth signing works.
* Lint.
* More OrderedDict to ensure data parameter order.
* Lint.
* Improve CF test assertions.
* Fix syntax error.
* Cleanup CF test.
The "request" object in CallbackResponse is the PreparedRequest send by
whatever client is used to contact the mocked moto service. This can end
up with unparsed bodies, as we added for processing presigned post
requests in #2155. This will make sure that all of the requests comming
in from mocked functions also get processed by werkzeug as if it was
running a live server.
This is a second attempt at resolving the issues with producing an XML consistent with what is produced from AWS (i.e., no spaces/new lines between tags). Another attempt (https://github.com/spulec/moto/pull/2205) is currently failing in tests.
This attempt uses precompiled regex patterns as class attributes of the `_TemplateEnvironmentMixin` to remove trailing spaces and newlines after a `">"`, and preceding newlines/spaces before a `"<"`. This *explicitly* wasn't done with a single regex to ensure that even things like `"...<ID>\n 12345\n </ID>"` are properly collapsed.
