openstack/codegenerator - codegenerator - OpenDev: Free Software Needs Free Tools

Change-Id: I9d8b5ee8f860c6de6ca2006ec85f9bb8f7a02e47

It can happen (in Keystone it does for sure) that certain API method
decorators when applied together result in certain schemas being present
in multiple decorator processing iterations. This results in some
schemas being multiplied. It should be safe to just use `set` instead of
a `list` to deduplicate them.
While we were working on this Nova "again" changed some decorators
breaking us so fix that as well since generated code is broken.
Change-Id: I25b2506264d6027f9d605c74297e6f0cc6ab2767
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

In addition to that fix the new issue in the compute schemas with enum
without type.
Make identity job to non-voting to temporarily bypass a deadlock and
land the fix.
Change-Id: I7f5f967824e338477a6979e34c2100433458e7c1
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

Currently we comment the operation_name attribute in the metadata that
it is used as an action name. This only creates confusion especially if
we want to use something different as the operation_name (i.e.
operation_name or opertaion_type for neutron router results in
"action"). So in addition to the renaming of the metadata attribute
explicitly pass the metadata operation key as operation_name parameters
into the generator (when unset).
Change-Id: Ic04eafe5b6dea012ca18b9835cd5c86fefa87055
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

In the metadata operation_name may be empty. In the cli output processor
we should fallback to the operation_type when operation_name is empty.
Change-Id: Ib7aebca1a6e5982ea5fa32c7dfe62fd382689cc9
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

Using operation_type is not correct since it is simply "action" often.
Use the `operation_name` instead as we use in the metadata.
Change-Id: I8395df6816d9db25d9395b4372878abeab9c4110

Switch to the new cli OutputProcessor initialization interface with
resource_key and oepration_type parameters for enabling command specific
hints.
Change-Id: I67eaef1fd5f9f1ed72a5f11a23e8559739792f8a
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

Change-Id: I93d5520bc67135db9b0ac9da40dbfb587cb9acad
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

Change-Id: Ifcd98f8533fe4403c1ff8299bd08ce45bfdb60b9
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

User schema in keystone is now good enough. Drop the hardcode.
Change-Id: I718906b0c9c12fd4c7936bdb5047725fa3267b39
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

Fix typo "market" where "marker" should be.
Change-Id: Ic9f94727b6abb33619bde96a610fedad7cc12398

`prop: {const: True}` should be converted to bool and not to integer.
Change-Id: Ic3775a36c138d451dbad7615bc0e446528d8b0be
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

In swift we service return few integer attributes which can only be
positive. Ensure this is incorporated in the schema.
Change-Id: I23509ca116f607a2b26fa1c01387dbd4f4df705a
Signed-off-by: Artem Goncharov <artem.goncharov@gmail.com>

When attribute is being locally renamed we miss the serde macros causing
deserialization to fail.
Change-Id: Ia8d8298c938ea088d1db62bf97df6defc41d93bc

Sort openapi spec to make it possible to put them under the version
control not taking care of changed ordering.
Change-Id: Ifb999f85a6d77e751cba8587c516c7d2c5c23644

In the cli generation it was assumed that the first element in the
response schema is the object itself. That is used to check whether the
response is list or not. Now the role_assignment schema return `links`
object first breaking that assumption. Extend the check to first look at
the `response_key` that we identified previously.
Change-Id: Ib2cf6e394486ebd0a475ed3d9bba97feefcd5205

Sometimes (and now very often) there is some race in the deep nested
structures rename handling. Apparently during renaming the wrong object
was modified causing it not to find referred entity later causing some
confusions depending on the order in which entities are processed.
Additionally when enum with the same name occur we should also look
whether the schema hash equals, in which case it is safe to simply skip
the schema completely).
Change-Id: I1bbb5f1e4f702ab54f72cdb70418b04567047187

For primitive types (not arrays and not dicts or similar) when resource
attribute can be explicitly nullable (while also being optional) it is
necessary to allow CLI user to set it to null. The only practical way
found till now is to add additional argument: `device_id / no_device_id`
which can be used to set explicit null. Neither relying on specific
value ("", "null") nor using default value when no value is sustainable
looking at the variations of APIs.
Change-Id: I6d549ffcc5c51a9c3ed849c590777b5ba4f6d721

Some conversion schemes missed explicit null support.
Change-Id: I01ed16f2690c78492a0d8673ad86f67de92c80ae

image in the server response may be an empty string if the server was
provisioned from a volume.
Change-Id: I4ec1d5391d373ec1aaba40b0547b06ebc9629ec2

On the response admin_state_up can be only boolean
Change-Id: I2d991491dbc5a97e36446a6d29c52b653b5a1657

port_details is an object and not array of objects.
Change-Id: I4095e664d714222d1d1ef8fdac0422393e7c451b

Change-Id: I8f394639f3272d4385618dda6731880e02e85a4e

Change-Id: I652e8b85632789d5172593154ddd9fef7508f359

In order to do more tests of the sdk/cli/types it is necessary to have a
convenient access to the source openapi spec. Copy this into the
`openstack_types/data`.
Change-Id: Ica1849dc81e77d05f277a14bb00030485c704ba2

Next step of enabling CLI output configuration. This time we add
resource_key into the generated CLI code so that we can map config
entries.
Change-Id: I43460adb71c18516c65ae20cf29a0a42a201a01f

- keypair create response does not contain ID
- server create response has a response_key (server). This is not really
 true, since creation may return reservation_id, but this is not
 typical and we can try to find a solution for that later. For now make
 sure cli server creation works.
Change-Id: I3f52f0e6686767dcc8862543c4c3a155f42ead0e

`port_range_min` and `port_range_max` of the security-groups[].rule
response has integer/null type
Change-Id: I10cfa565bede4c5f77069b6ac47a3b7ca0bc0e25

Change-Id: I5912a18f1a7f60b6eff96aeb9580eb2d05c3ebb6

There is case (at least on one cloud) where user_id attribute of the
instance_action is null (explicit null and not unset).
Change-Id: Ic8ac6416a4d48d13a62b37b4b08c25b425ecc2b7

Change-Id: I664672f9aca455991f737047d8ca53c844d4058f

Not all of the fields declared as mandatory are being returned by real
clouds.
Change-Id: I59ed1b1014a6eb91f067b61e3b57a3bb8f4e2b8e

container_format and disk_format in reality need to be `anyOf[String,
Enum]` since it is possible to modify this list on the deployment level.
Unfortunately this also means that the CLI (create case) is becoming
very tricky. For now prepare, but not enable corresponding changes in
the create schema until a good way of handling this in cli has been
found.
Change-Id: I9162e238d98fd494131c01ff02990e2ddeda4887

There are lot of changes related to the deduplication caused mostly be
lot of data simplifactions done on the cli side for the responses which
are not done on openstack_types. This changes were applied manually to
the target repository since the changes require also bumping structable
version.
Some other minor enum sorting and docstring formatting fixes has been
applied.
Change-Id: Ifc508c54c01bd39af12e3023668e698085829a05

Rework StructTable imports to point directly to structable (with
additional Options import). On the way drop unnecessary
`#[serde(untagged)]` for StringEnum and `serialize` for `Option<String>`
Change-Id: Ic9b7b5b319e5a7ea881f7b5e3e9da986276d079b

- add services added recently
- add info about sdk/cli/tui that are generated using codegenerator.
Change-Id: Id41a3262be58deefd1f127a4c1a0df0b3b2e0aec

Change-Id: Ibd1272ff55ea36624de2947945376291c460dce3

Change-Id: I5e390dbb7ecb79eef0dacca90a6897433ab1851e

openstackdocstheme changes got merged and we need to adopt to it.
Change-Id: I4781e311e332d0837036eb4225a3fcca2116572c