* update parser to 2.0.29
* better handling of null in dereferencing
* update parser to 2.0.30
* update core to newer version
* add new files
* rollback to previous stable version
* remove files
* Fixes for python-experimental NullableShape component
Co-authored-by: Justin Black <justin.a.black@gmail.com>
* add handling of array of oneOfs
* handle res model-type lowercase name +shorten code
* remove unnecessary vendor-extension model enum
* handle openapi lowercase schema name for array res
* change xs:anytype to string for file responses
* update checkstyles
* remove not needed imports again
* update samples
* upper/lowercase use local getdefault
* update samples again
* Add backslash to load namespace from global namespace
* Update example
* Remove hardcoded slash
* Add backslash from java side
* Update samples
* Fix first sample
* Fix other snapshot
* Update version file
The FILES metadata is supplementary, but was logging a warning while
generating Ada client that FILES couldn't be written. An exception was
being thrown because Path was being reported as two different "types of
Path". One would be reported as mac file, while the other was reported
as a unix file. The fix here is to disconnect path details from the
underlying file system provider by creating a new Path based on the
file's absolute path. This change also fixes sorting so it works
alphabetically in ascending order.
* FILES is now path relative with no prefixes
some Java implementations don't honor .relativize documentation fully.
When outDir is /a/b and the input is /a/b/c/d, the result should be c/d.
Some implementations make the output ./c/d which seems to mix the logic
as documented for symlinks. So we need to trim any / or ./ from the start,
as nobody should be generating into system root and our expectation is no ./
This resolves regeneration issues for those on such Java
implementations, although we've not been able to track down the exact
vendor or configurations which might lead to these differences.
* Ensure windows outputs FILES paths in same format
* Normalize FILES paths to remove relativization mid-path
* Add samples composer.lock to root .gitignore
composer.lock may produce CI errors when you need to test build against
different PHP versions. However users most likely want to commit this
file, so I think it's better to exclude it only in root .gitignore.
* Commit composer.lock in default PHP templates
* Refresh samples
* renamed async methods to end with async
* update samples
* updated samples
* updated test to use new name
Co-authored-by: William Cheng <wing328hk@gmail.com>
* Ensure CLI-level generator settings are available in additional properties
* Initial conversion to yaml configs
Configs prefix with 1- need to be manually evaluated.
* Add "other" configs not in ensure-up-to-date
* Add other/openapi3 files
* Cleanup all generation scripts
* Clean up: add missed configs and fix some openapi 3 diffs
* Move generate-samples script, error on batch failures
* Temporarily disable elm which requires skip validation of spec
* CI updates (todo: run all generators through appveyor or move to github workflows)
* Add success count to batch generation command output
* [samples] Regenerate
* Remove bin/windows
* Generate swift5 samples in bitrise
* Expand user input glob pattern
* Regenerate samples
* Update PR template
* Support config based generateAliasAsModel
* [samples] Regenerate
* [rust] Generate as alias for all samples configs
* [csharp] Move test staged file changes to in-directory
* Include "live" test files and ignore test cache files from samples.ci
* Remove concept of samples.ci staging
Generators support .openapi-generator-ignore, allowing maintainers to
explicitly ignore the regeneration of files which have been modified.
Note that the tooling does not overwrite test files whenever those files
exist, and it's not entirely necessary to add test files to the ignore
file.
* Re-add meta generator scripts
* Modify ensure-up-to-date to log stdout, so script does not look like it hangs on slower machines
* [csharp] Regenerate sample
* Set generateAliasAsModel CodegenConfigurator#toContext
This ensures the property (which is effectively a thread-local) gets set
as expected if the caller resets GlobalSettings as is done in the
GenerateBatch command.
* [rust] Regenerate samples
* Re-enable elm
* [java] Support OffsetDateTime example, fail ensures script on any generation error
* [samples] Regenerate
* Fix jersey2-java8 ignore file
* Fix elm spec validation
* Force UTC in sample generation to avoid timezone conflicts during generation (user vs ci)
Co-authored-by: William Cheng <wing328hk@gmail.com>
