Note that specification extensions beginning x-oas- are reserved

Allow specification extensions in discriminator object

security; add new securityScheme type of mutualTLS

* Fix: #832. Add info.summary.

* Fix: summary is shord, description is verbose.

Be consistent with other definitions of summary and description.

Signed-off-by: Axel Nennker <axel.nennker@telekom.de>

fix OIDC url and OAuth2 requirements

#2066)

* Extend style, explode, allowReserved in encoding to multipart-formdata (#2018)

* Update versions/3.1.0.md

Co-Authored-By: Ron <ron@swagger.io>

* Replace details of multipart/form-data format with referce to RFC 7578

* Update versions/3.1.0.md

Co-Authored-By: Darrel <darrmi@microsoft.com>

Update Schema Objects to JSON Schema Draft 2019-09

…on Style Examples, they support objects.

* Add webhooks as a top-level element to the spec

* Add the changes from #2048 and signpost webhooks

* Add an example of webhooks

* Relocate and expand on webhooks section following feedback

* Better wording to describe expectations on API consumers

* Clearer wording for why the paths element is here

* Update language to make callbacks clearer

This adapts the language from PR #2046, with minimal wording tweaks
to account for type now being able to have multiple values (type arrays).

…OAS 3.1 (#2107)

* Checkpoint of draft

* Fix typo.

Co-Authored-By: Darrel <darrmi@microsoft.com>

* Fix plural anchor

Co-Authored-By: Mike Ralphson <mike.ralphson@gmail.com>

* Remove superfluous specification

Co-Authored-By: Phil Sturgeon <me@philsturgeon.uk>

Co-authored-by: Darrel <darrmi@microsoft.com>
Co-authored-by: Mike Ralphson <mike.ralphson@gmail.com>
Co-authored-by: Phil Sturgeon <me@philsturgeon.uk>

* Make paths object optional

* Adding reusable Path Item Objects

Under `components`

* Adopt DM's suggested change to OpenAPI doc definition

* Cleanup use of specification and definition where we mean document

* multipartite>composite, define ACL

* Add ' | Reference Object' to callbacks/webhooks

Co-authored-by: Ron <ron@swagger.io>

* fix typo in Callback Object

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* retain typo in v3.0.2; fix for v3.0.3 (#1899)

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Clarify empty Security Requirement Object usage and validity (#1886)

* Clarify empty Security Requirement Object usage and validity

* Reorder sentences to make clearer.

* Remove wrong text.

* Removed unneeded text.

Co-authored-by: Ron <ron@swagger.io>
Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Ron's wording for Darrels feedback

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* ted updates

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Replace 'application' by 'API' within the 'Info Object' definition. (#2004)

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Path Templating Clarification - proposed fix for #1830. (#1831)

* Proposed fix for #1830. Each variable expression in a path must have a corresponding path parameter.

* #1830 - Removed 'at least once' to defer the question about repeated references to a single path parameter.

* Update #1830 fix with suggestion from Darrel

@darrelmiller suggestions we use "template expression" instead of "variable expression" to align with RFC6570. Good idea.

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* yaml.org supports https, but www.yaml.org is misconfigured

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Updated text for OperationRef

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* fix a typo in the Security Filtering section (#1837)

* fix a typo in the Security Filtering section

* Security filtering slight reword

Co-authored-by: Mike Ralphson <mike.ralphson@gmail.com>
Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Make ABNF for runtime expressions complete

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Explain unclear semantics of property `$ref` in Path Item Object (#1964)

* Explain unclear semantics of property `$ref` in Path Item Object

Currently, as explained in #1038 (comment) the description of `$ref` in [Path Item Object](https://github.com/OAI/OpenAPI-Specification/blob/3.0.2/versions/3.0.2.md#pathItemObject) is unclear about the semantics behing it. I took the explaination from issue #1038 to make it more clear.

* Update versions/3.1.0.md

Co-authored-by: Ron <ron@swagger.io>
Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Clarify constraints on Security Scheme Object Scheme Property (#1880)

* Wording around scheme extensions

* Clarified that securitySchemeScheme is only a SHOULD be registered scheme

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* fix difference between yaml and json in Response Object Examples

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Server Variable Object clarifications (#1809)

* Server Variable Object clarifications

* Toned language down for proper semver versioning

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Fix formatting errors in example (#2132)

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Update 3.0.3 for release (#2149)

* Update README.md for release

* Update release date for 3.0.3

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Update versions/3.1.0.md

Co-Authored-By: Darrel <darrmi@microsoft.com>
Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Fixed typo

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* explicit 'forward slash'

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* fix #2053: `style` keyword is not supported inside Schema object

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* OpenAPI not Open API

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* backticks

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* minor clarification for operationId usage in link objects (#1733)

* minor clarification

it's a bit confusing that both the id and the reference are called "operationId", so this tweak makes the text a bit more explicit.

* use right terminology

Co-Authored-By: Mike Ralphson <mike.ralphson@gmail.com>

Co-authored-by: Ron <ron@swagger.io>
Co-authored-by: Mike Ralphson <mike.ralphson@gmail.com>
Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Update 3.1.0.md

fixed typo

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Removed confusing comment

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Clarify the spec to allow optional or unspecified OAuth scopes (#1888)

* Referencing issue #513. Clarify the spec to accommodate OAuth schemes where scope may be unspecified (optional scope) or where scope is not used at all.

* Removed the provision for default scope represented as empty string. This introduces some ambiguities in the Security Requirement Object that would need to be addressed.

* For #513, adjusting language and removing examples

For #513, adjusting language and removing examples as suggested by @webron.

* removed unnecessary example header

Co-authored-by: Ron <ron@swagger.io>
Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* The examples keyword is not supported inside schema (#2042)

* examples not supported inside schema

* figured it out

* a tiny little edit

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Fix 'Security Scheme Object' definition with OAuth 2.0 grant types. (#2006)

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Fix formatting errors in example (#2132)

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

Co-authored-by: seiya <r108338@yahoo.co.jp>
Co-authored-by: Adam Leventhal <ahl@transposit.com>
Co-authored-by: Sebastián Ramírez <tiangolo@gmail.com>
Co-authored-by: Ron <ron@swagger.io>
Co-authored-by: Phil Sturgeon <me@philsturgeon.uk>
Co-authored-by: Patrice Krakow <patrice.krakow@gmail.com>
Co-authored-by: Ted Epstein <ted.epstein@reprezen.com>
Co-authored-by: Darrel Miller <darrmi@microsoft.com>
Co-authored-by: Carsten Brandt <mail@cebe.cc>
Co-authored-by: Henry Andrews <andrews_henry@yahoo.com>
Co-authored-by: Sergej <sergej2705@users.noreply.github.com>
Co-authored-by: nasa9084 <nasa.9084.bassclarinet@gmail.com>
Co-authored-by: Erik Wilde <dret@users.noreply.github.com>

)

Co-authored-by: Ron <ron@swagger.io>

Correct Styles Values for spaceDelimited and pipeDelimited, as based …

Co-authored-by: Phil Sturgeon <me@philsturgeon.uk>

While going over the changes for the release notes, found two issues:
- The TOC entry for `Relative references in URIs` was not modified to match the change in the spec.
- The `Paths Object` had an extra sentence that should have not been there (referencing sub-documents and overlays).

Missing updates

This brings exclusiveMinimum, exclusiveMaximum, minimum, and
maximum, into full modern JSON Schema compatibility.

There are no edits directly mentioning minimum and maximum,
but removing the boolean form simplifies their processing
by making it context-independent.

* Update "format" and "content*" for new JSON Schema

This removes OAS formats and examples that are now superfluous
as they are part of the 2019-09 JSON Schema draft.

Similarly it deprecates the "byte" and "binary" formats in favor
of JSON Schema's "contentEncoding" and "contentMediaType" keywords,
and updates various related exapmles and other guidance.

It also removes confusingly blank rows in the OAS format table.

* "format" is an annotation

* Fix broken table, type, in Encoding Object

Broke some things while updating for "content*"

* Fix format of `format`

Backticks, not double quotes.

* Remove unneeded detail on "format"

This was just duplicating info from the JSON Schema spec.

Co-authored-by: Darrel <darrmi@microsoft.com>

* Remove "byte" and "binary" formats altogether.

Instead of just deprecating.  The "content*" keywords now
cover these use cases.

* Harmonize JSON Schema content* + Media Type Object

Includes harmonizing with the Encoding Object.  In general,
OpenAPI objects set the media type, although there is a case
for `contentMediaType` with multipart/form-data.  Otherwise,
`contentEncoding` replaces the now-removed custom formats.

A possibly controversial change is to indicate unencoded binary
data by omitting `type` (or omitting the schema altogether), as
binary data does not conform to JSON string requirements.

This could still be done with `type: string` if that is preferred.
It's going to be a bit weird either way.

I can add wording in the next JSON Schema draft to clarify
whichever approach makes more sense.

* Fix typos from review

* Remove stray {}

* Fix inconsistencies contentMediaType and Encoding Object

Co-authored-by: Darrel <darrmi@microsoft.com>

* drop OAS semver requirement

* Update versions/3.1.0.md

Co-authored-by: Darrel <darrmi@microsoft.com>

* chore: explain how $schema might work

* reordered and made it specifically only schema resources

* Update versions/3.1.0.md

Co-authored-by: Karen Etheridge <ether@cpan.org>

* Update versions/3.1.0.md

Co-authored-by: Ben Hutton <relequestual@gmail.com>

* new approach

Co-authored-by: Phil Sturgeon <me@philsturgeon.uk>
Co-authored-by: Karen Etheridge <ether@cpan.org>
Co-authored-by: Ben Hutton <relequestual@gmail.com>

…ed for any data type. (#2302)

* The JSON schema specification states the format keyword can be used for any data type, not just primitive types

* The JSON schema specification states the format keyword can be used for any data type, not just primitive types

Co-authored-by: Darrel Miller <darrel.miller@microsoft.com>

Followup to #1809, now that we allow breaking changes.

Co-authored-by: Darrel Miller <darrel.miller@microsoft.com>

Fix indentation of a YAML comment

* Update 3.1.0.md

* Merge branch 'master' into v3.1.0-dev

* Added words relating to adopting semantics of JSON Schema

* Update versions/3.1.0.md

Co-authored-by: Mike Ralphson <mike.ralphson@gmail.com>

* Update versions/3.1.0.md

Co-authored-by: Mike Ralphson <mike.ralphson@gmail.com>

Co-authored-by: Darrel Miller <darrel.miller@microsoft.com>
Co-authored-by: Mike Ralphson <mike.ralphson@gmail.com>

fix typo in release history table

fix link to style values in serialization table

[v3.1.0-dev] reserve both x-oas- and x-oai- prefixes

)

…ncluded in JSON Schema (#2394)

* Added link to JSON Schema Validation docs explain which formats are included in JSON Schema

* Update verbiage to be more accurate

Co-authored-by: Mike Ralphson <mike.ralphson@gmail.com>

Co-authored-by: Mike Ralphson <mike.ralphson@gmail.com>

Improve wording about 'summary' and 'description' in Reference Object

Co-authored-by: Phil Sturgeon <me@philsturgeon.uk>

* oas 3.0 doesn't mention slashes not allowed

* none of those either

Co-authored-by: Phil Sturgeon <me@philsturgeon.uk>

…2362)

* OAS schema dialect clarifications

* OAS schema dialect clarifications

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* $schema is allowed in subschemas when bundling

Co-authored-by: Ben Hutton <relequestual@gmail.com>

* Schema dialect clarifications from Ben

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Use top-level jsonSchemaDialect field

Co-authored-by: Ben Hutton <relequestual@gmail.com>

…plicit (#2437)

* fix http link to json-schema.org

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* fix http link to spec.commonmark.org

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Specify rules for $ref resolution

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Specify relative resolution rules for pathItem $ref and example externalValue

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Update JSON Schema draft links to 2020-12 IETF pages

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Make language about 'MUST be in the form of a ...' consistent

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Make it clear pathItem $refs don't need to be external now

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Make RFC links consistent with regard to spacing

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Allow a URI for example.externalValue fields

This makes it fall under the rules for relative references.

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Explicitly call out $ref as a Relative Reference

* Remove wording about what implementations SHOULD/MAY do with a $ref

* Prefer 'referenced document' to 'referrant document' for clarity

* Fix JSON Schema $ref resolution fallback rule

* Add links back to #relativeReferences definition

* Split #relativeReferences definition into URL and URI sections

Signed-off-by: Mike Ralphson <mike.ralphson@gmail.com>

* Clean-up wording about $refs in responsesObjects, fixes #1679

* Agreed to remove explicit verbiage around $refs in responseObjects, fixes #1679

* Fix, clarify, and simplify content type schemas

This fixes #2349, which caught that an encoded PNG image
is encoded into a text media type.

In the process I realized some other errors, and simplified things.

* HTTP `Content-Type` is always handled by OAS
    * Media Type Object key in most cases
    * Encoding object (possibly inferred from schema) in `multipart/form-data`
* HTTP-level `Content-Encoding` is always handled by the OAS Header Object
* JSON Schema "content*" is used for embedding one media type into another
    * the encoded resource is of media type `text/plain`
    * `"contentMediaType"` is the embedded media type after decoding
    * `"contentEncoding"` is how to encode/decode binary to/from text

This removes any chance of `"contentMediaType"` conflicting with
the Media Type Object key or with `contentType` in the Encoding Object,
as they now always do different things.

Likewise, the HTTP `Content-Encoding` header (with values like
gzip, deflate, etc.) does different things than `"contentEncoding"`
(which has values like base64, base64url, quoted-printable, etc.).

The deprecated part header `Content-Transfer-Encoding` is likewise
handled in the Encoding Object, but is probably never used.

* Fix Content-Type to indicate semantics

...rather than literal content format on the wire.

* Update 3.1.0.md

Fixed a typo and changed a SHOULD to MAY.

* Update versions/3.1.0.md

* clarify default encoding content type value.

* Describe interaction between JSON Schema contentEncoding and HTTP Content-Encoding header

Co-authored-by: Mike Kistler <mkistler@us.ibm.com>

Co-authored-by: Darrel <darrmi@microsoft.com>
Co-authored-by: Mike Kistler <mkistler@us.ibm.com>

* 3.1.0 release prep

* Update README.md

Per comment in review, authors determine whether a spec is a single or multipart document. Those who consume the spec care more about the information itself and less (or not at all directly) about how it was assembled.