* Support for multipart/form-data; charset="utf-8"
Previously, it was only checked for an exact equal string, which failed when there were additional options such as charset.
* Update samples
* [haskell][haskell-yesod] remove fixOperatorChars()
fixOperatorChars() does not change input strings since special
characters have already been replaced in DefaultCodegen.fromModel().
* [haskell][haskell-yesod] do not prefix with quote ("'") when generating removeFieldLabelPrefix table
We switched from the conversion done by fixOperatorChars() to the
conversion done by DefaultCodegen.fromModel() and the latter does not
insert quote characters. So We modify the removeFieldLabelPrefix table
to conform the new mapping.
* [haskell][haskell-yesod] remove forParsing parameter from removeFieldLabelPrefix function
Aeson's fieldLabelModifier always convert Haskell field names to JSON
field names, whether at parse time or not. (Note that stripPrefix and
uncapitalize do not take such parameter)
* [haskell][haskell-yesod] perform replaceSpecialChars after stripping prefix
Because replaceSpecialChars can corrupt prefix if the prefix contains
a replacement string of a specfial character as a substring.
* [haskell][haskell-yesod] regenerate samples
I spotted a mention of `github.com/antihax/optional` in
AbstractGoCodegen. I was curious and investigated. However, it turns out
that this code is an obsolete remnant of the go-deprecated generator
removed before 6.0.0, so is now just a false lead. This PR cleans it up
so no-one else has to go down the same investigative pathway.
* [Ruby] Fix Operation Servers
* user `server_index` instead of `index`
* use nil as default
* add tests
* revert tests
* add tests to custom file
* add test file to ruby-faraday
This fixes#15043
The issue is that browsers like "text/html,...,*/*;q=0.8" (see for
instance https://developer.mozilla.org/en-US/docs/Web/HTTP/Content_negotiation/List_of_default_Accept_values
)
Without this commit we end up with an array of accepted type like
`("text/html", "*/*;q=0.8)` so when we then check if the array contains
`*/*` the check fails, and we return a 406 even though the client is
able to get the response.
This commit fixes it by removing the `;q=0.8` part.
(Ideally we should not just discard that part, we should extract that
value, and order by it. See
https://developer.mozilla.org/en-US/docs/Glossary/Quality_values for
more info about that. However this could be done in a subsequent PR:
this already fixes the 406 error, which is pretty blocking)
* [Julia] update samples test and a minor fix
- updating samples test for Julia generated code to use the latest OpenAPI.jl Julia package, which includes tests for file uploads
- also including a fix where generated code duplicated the `mandatory` flag in one of the method calls
* add blank line, trigger build with julia workflow
* more updates to julia samples tests
* add samples changes
* [haskell] fix exampe code in generated README.md
Since create{{title}}Client is not a monadic action, we need to use
"let" instead of "<-".
* regenerate samples
- detect multipart and urlencoded post data formats correctly on the server side
- generate data types correctly for binary file uploads
- re-generated samples
* Switch Poison to Jason
* generate-samples.sh
* Finalize Poison -> Jason switch
* parse date-time values to Elixir DateTime
* improve formatting in various places, so there's less changes by `mix
format` later
* fix Java version in flake.nix
* Use List.delete/2 instead of Enum.reject/2 for performance reasons
* mix format test/*
* Install dialyxir and fix reported issues
* Fix RequestBuilder.decode/2 hardcoded module name
* Update docs
* Revert changes to API spec (HTTP -> HTTPS)
* Revert uneeded change to Elixir code generator
* Use HTTP in Elixir tests
HTTPS doesn't work for folks who setup petstore.swagger.io as described
in docs/faq-contributing.md.
---------
Co-authored-by: Wojciech Piekutowski <wojciech@piekutowski.net>
