* Add code comments
* Change 'setup' method to setPrivateKey
* Add support for configuring digest algorithm
* run script in bin directory
* format generated code
* Revert "format generated code"
This reverts commit 3b527784375bcf86cacf75bd0142f63b91c96ff8.
* [scala][akka-http] fix non-default packages for api, model and invoker
* remove default invoker additionalProperty
* fix invoker package for scala-akka and sttp
* fix invoker package and settings for scala-akka
* fix tests compatible with windows paths
* test output
* test output
* fix base output path for test
* fix output path for test
* fix sttp invoker package
* fix runtime deserialisation
* add security support for dart dio
* regenerate dart dio sample
* update dio pubspec.mustache deps
* check response type before parsing
* add default dateTime serializer
* regenerate sample
* Propagate deprecated property through $ref's
As $ref is supposed to completely replace the definition of a property,
make sure we also include the 'deprecated' property when generating the
type of a $ref property.
This makes a property $ref'ing a deprecated schema also become deprecated.
* Clarify why we're messing around with $ref
* Fix return type in model setters.
Previously return type was same, as method arguments. It`s wrong, and cause errors like
"Return value of Foo::setSuccess() must be of the type bool, object returned"
We cant use self and current {{classname}} as return type, because that can break class inheritance. So, better remove type hint on setters, until PHP-devs dont make realization for return static
* Add return self type hint for setters
* Revert "Add return self type hint for setters"
This reverts commit 07dd9715
* better anytype support
* add tests for any type
* fix test with any_value
* fix tests
* fix case additionalProperties: {}
* test with CI
* remove check in map schema
* Revert "remove check in map schema"
This reverts commit e016c4155f1bd8753e894d87bad8e70eee29f3d5.
* fix tests, comment out map schema fix
* fix tests
* fix tests with correct codegen model
* fix tests
* fix tests for map of any type
* fix array of any type
* fix array of any type
* update samples, remove log
* add typemapping to go, python
* Fix problem with clients, that put charset in content type header.
With this fix header "Content-Type: application/json; charset=utf-8" working same as "Content-Type: application/json" for parse input data
* Fix code style, add $consumes length check.
* Add isContentTypeAllowed static method and tests
* Fix old tests
Right now serializer doesn't support anything beside json and xml.
Call tests with application/json instead of form data.
Co-authored-by: Yuriy Belenko <yura-bely@mail.ru>
* Fixed valuetype parameters and discriminator deserialization
- Made non-required valuetypes nullable, and flagged required valuetypes as "x-csharp-value-type"
- Made sure to camelCase discriminator property names in Subtype converter
* Ran pet store sample script
* Ensure that readWriteVars and readOnlyVars are also marked as isEnum and isPrimitiveType where appropriate
* Updated petstore sample with enum fix
* Fields that are required should emit default values (otherwise the consuming API might throw a fit
* Added missing sample updated
* Re-ran petstore example script to grab all changes
* Rebased and re-ran example update script
* update csharp samples
Co-authored-by: William Cheng <wing328hk@gmail.com>
Add a lambda that formats the documentation correctly even if it
contains newlines. The generated docs with typedoc looks a lot better
and also renders Markdown correctly.
Also remove the "no description" fallback for APIs because we don't have
it anywhere else, other generators don't generate a default fallback and
I'd rather have no documentation than a "no description" string.
There are cases where minimizing quotes results in invalid YAML. For
example, an input YAML with string "1234_1234" will be converted to YAML
value 1234_1234 which is an int in YAML 1.1
(https://yaml.org/type/int.html)
THe only option in these cases is to either:
* Revert the option completely to always quote values
* Provide a user-customization to disable quotes minimization
This applies the latter with the assumption that this is an edge case
and users who are unaffected will default to the "prettier" version.
An alternative would be to write a custom serializer for strings, and if
they are in the format of of any of the valid formats defined in YAML:
[-+]?0b[0-1_]+ # (base 2)
|[-+]?0[0-7_]+ # (base 8)
|[-+]?(0|[1-9][0-9_]*) # (base 10)
|[-+]?0x[0-9a-fA-F_]+ # (base 16)
|[-+]?[1-9][0-9_]*(:[0-5]?[0-9])+ # (base 60)
Then wrap the result in quotes. That approach was not taken because of
the potential for significant performance impact on very large specs,
which our users are often tasked with transforming.
deepObject query parameters need to be specially marshalled.
Unfortunately, they're quite tricky to distinguish in mustache because
there's no boolean property for deepObject, so add a vendorExtension
property to ease the mustache template.
* Use string form of filename parameter
This works for the form-data library and is also compatible with the
browser FormData object.
* Add new option to select platform node or browser
When no platform is selected, a default is chosen by the framework
option and likewise the file data type option is implied by the
platform.
* Remove redundant import of node dns module
* Only use form-data library for node platform
* Generate npm package from npmName option
* Use method convertPropertyToBooleanAndWriteBack
* Generate typescript samples with ensure-up-to-date
* Remove unused supportsES6 field from codegen
* Add a new switch for RXJS
* Remove redundant npm dependency on rxjs4 types
* Fix return type of PromiseMiddleware methods
* Install webpack dependency to run jquery tests
* Update form-data to 2.5 which includes typings
* Add missing dependency on node typings
* Fix test artifact name typo
* Stub rxjs when it is not explicitly enabled
