* [okhttp-gson] Make initializer block for GSON builder static #16860
Make initializer block for GSON builder static, see #16860
* update sapmles
---------
Co-authored-by: Andre Wachsmuth <awa@xima.de>
* feat(client): support ability to disable boxed models
Add new additional property `avoidBoxedModels` which can be configured to avoid `Box<..>` of models.
It's very nice to have such config, because it doesn't make a lot of sense for really simple models.
* test: rollback config of existed test
* Fixed invalid extraction of response body in kotlin-client jvm-spring-*
* Generated echo-api for kotlin-jvm-spring-3-restclient
* Specific echo-api for Kotlin without allOf/anyOf
* Specific echo-api for Kotlin without allOf/anyOf
* Generated all samples
* Added kotlin-jvm-spring-3-restclient sample to workflow
* Fixed syntax problem
* [kotlin] Fixed multiple problems with default values
* [kotlin] Removed old commented code
* [kotlin] Generated new samples after merge with master
* [kotlin] ensure-up-to-date
* fix(rust): discriminator mapping to serde rename
Discriminator mapping has been ignored in some cases.
Even existing samples had wrong definition in some cases
This PR addresses this
* fix(rust): `oneOf` generation for client
Solves #17869 and #17896 and also includes unmerged $17898
Unfortunately it affects quite a lot of code, but we can see that only client-side models were affected by re-generation.
I tried to split this PR to several, but they're really coupled and hard to create a chain of PRs.
* fix: indentation in `impl Default`
* missing fixes
* fix: correct typeDeclaration with unaliased schema
* style: improve indentation for models
* fix: user toModelName for aliases of oneOf
* refactor: unify `getTypeDeclaration` for rust
* cover the case when `mapping` has the same `ref` for different mapping names
* test: add test for previous change
* style: remove extra qualified path to models
* add some comments
* fix(build): use method of `List` instead of specific for `LinkedList`
* add feature to generate only interface files
* generate sample
* add workflow file foe go gin service
* add workflow file foe go gin service
* add workflow file foe go gin service
* update samples
* fix(rust): simplify None logic
Used to be a code path (where isFreeFormObject and something else) where we could end up with two `: None` `: None`'s
This does add the None for the !required && !isArray && !isMap && !isNullable case (where nothing would have been printed before.
Above for the type on the individual vars inside the struct definition the logic is simply if not required then it's optional.
If it's optional we want a None when creating a new struct.
* ci: add tests for failure
* Use `exists()` to check for existing properties.
* Generate Petstore code.
* Fix use of `null` instead of `undefined`.
* Enhance function code.
* Regenerate Petstore code.
* Use `exists()` in generated API clients.
* Use `exists()` in generated API clients.
* Refer to properties differently.
Even though `title` is technically for the title of the API itself,
it's possible to set it for properties as well.
Right now, the `title` property is available in templates for regular properties,
but for ref-properties it gets lost right here.
This seemed rather surprising to me, so I think it's better to fix.
* fix(ruby): Detect content-type in form-data file upload using faraday (#17851)
* Update samples
* Add needed dependency libmagic-dev into node 1 setup of Circle Parallel CI
* After merge/pull last master + regeneration of samples
* [rust] basic oneOf support
Suport oneOf as a rust struct enum.
Details:
Enum without a discriminator is untagged being
"untagged" simply means serde won't attempt to
store the name of the enum inside the serialized
object. See
https://serde.rs/enum-representations.html#untagged
for more
Also check to make sure the mapping values
are not an empty object (aka null).
Co-authored-by: Nikita Puzankov <git@bshn.rs>
* refactor: fix clippy lints
No longer needed as of reqwest 0.10, it now takes the response as owned instead of mut ref.
Is not empty is more clear
* fix: discriminator and oneof case
Will show as a struct enum when there are additional fields, otherwise will be a tuple enum.
not sure the purpose of x-mapped-models, perhaps legacy code? mappedModels appears to do the same thing.
Also add default implementation for quality of life
* chore: update samples
---------
Co-authored-by: Nikita Puzankov <git@bshn.rs>
