Move the JsonSchemas to be generated at runtime
Generating all of these index files at compile time is kind of wasteful,
considering that it's really not that much code to generate them. This
commit modifies the update_schemas script to generate a C++ version of
the schemas that can then be used to generate a route. This allows
deleting a LOT of fixed files, for a very small incremental binary size
increase.
This change will cause two impacts that a user could possibly notice.
Expand will now work properly on JsonSchemas tree.
Keys on the JsonSchema schemas will now be sorted alphabetically, as we
do elsewhere.
All other things should remain the same.
Tested:
Redfish service validator passes.
curl --insecure --user root:0penBmc https://192.168.7.2/redfish/v1/JsonSchemas/Triggers
Returns the same object as before with changes above.
This adds 2512 bytes to the bmcweb binary size, and interestingly, adds
3764 bytes to the overall rootfs size, despite the level of file
deletion seen in the patch. While this is debatably "worse" than
what we had before in this regard, making JsonSchemas work similar
to how the rest of the Redfish tree operates, is worth the minor
increase in code size.
Signed-off-by: Ed Tanous <edtanous@google.com>
Change-Id: Ic0d568112a3784821aa6867f7bcf50705dc586db
diff --git a/scripts/update_schemas.py b/scripts/update_schemas.py
index af62962..8cd37e6 100755
--- a/scripts/update_schemas.py
+++ b/scripts/update_schemas.py
@@ -11,6 +11,17 @@
VERSION = "DSP8010_2021.4"
+WARNING = '''/****************************************************************
+ * READ THIS WARNING FIRST
+ * This is an auto-generated header which contains definitions
+ * for Redfish DMTF defined schemas.
+ * DO NOT modify this registry outside of running the
+ * update_schemas.py script. The definitions contained within
+ * this file are owned by DMTF. Any modifications to these files
+ * should be first pushed to the relevant registry in the DMTF
+ * github organization.
+ ***************************************************************/'''
+
# To use a new schema, add to list and rerun tool
include_list = [
'AccountService',
@@ -127,9 +138,15 @@
r.raise_for_status()
+
static_path = os.path.realpath(os.path.join(SCRIPT_DIR, "..", "static",
"redfish", "v1"))
+
+cpp_path = os.path.realpath(os.path.join(SCRIPT_DIR, "..", "redfish-core",
+ "include"))
+
+
schema_path = os.path.join(static_path, "schema")
json_schema_path = os.path.join(static_path, "JsonSchemas")
metadata_index_path = os.path.join(static_path, "$metadata", "index.xml")
@@ -291,51 +308,27 @@
zip_filepath = os.path.join("json-schema", basename)
schemadir = os.path.join(json_schema_path, schema)
os.makedirs(schemadir)
- location_json = OrderedDict()
- location_json["Language"] = "en"
- location_json["PublicationUri"] = (
- "http://redfish.dmtf.org/schemas/v1/" + schema + ".json")
- location_json["Uri"] = (
- "/redfish/v1/JsonSchemas/" + schema + "/" + schema + ".json")
- index_json = OrderedDict()
- index_json["@odata.context"] = \
- "/redfish/v1/$metadata#JsonSchemaFile.JsonSchemaFile"
- index_json["@odata.id"] = "/redfish/v1/JsonSchemas/" + schema
- index_json["@odata.type"] = "#JsonSchemaFile.v1_0_2.JsonSchemaFile"
- index_json["Name"] = schema + " Schema File"
- index_json["Schema"] = "#" + schema + "." + schema
- index_json["Description"] = schema + " Schema File Location"
- index_json["Id"] = schema
- index_json["Languages"] = ["en"]
- index_json["Languages@odata.count"] = 1
- index_json["Location"] = [location_json]
- index_json["Location@odata.count"] = 1
-
- with open(os.path.join(schemadir, "index.json"), 'w') as schema_file:
- json.dump(index_json, schema_file, indent=4)
with open(os.path.join(schemadir, schema + ".json"), 'wb') as schema_file:
schema_file.write(zip_ref.read(zip_filepath).replace(b'\r\n', b'\n'))
-with open(os.path.join(json_schema_path, "index.json"), 'w') as index_file:
- members = [{"@odata.id": "/redfish/v1/JsonSchemas/" + schema}
- for schema in schema_files]
+with open(os.path.join(cpp_path, "schemas.hpp"), 'w') as hpp_file:
+ hpp_file.write(
+ "#pragma once\n"
+ "{WARNING}\n"
+ "// clang-format off\n"
+ "\n"
+ "namespace redfish\n"
+ "{{\n"
+ " constexpr std::array schemas {{\n"
+ .format(
+ WARNING=WARNING))
+ for schema_file in schema_files:
+ hpp_file.write(" \"{}\",\n".format(schema_file))
- members.sort(key=lambda x: x["@odata.id"])
-
- indexData = OrderedDict()
-
- indexData["@odata.id"] = "/redfish/v1/JsonSchemas"
- indexData["@odata.context"] = ("/redfish/v1/$metadata"
- "#JsonSchemaFileCollection."
- "JsonSchemaFileCollection")
- indexData["@odata.type"] = ("#JsonSchemaFileCollection."
- "JsonSchemaFileCollection")
- indexData["Name"] = "JsonSchemaFile Collection"
- indexData["Description"] = "Collection of JsonSchemaFiles"
- indexData["Members@odata.count"] = len(schema_files)
- indexData["Members"] = members
-
- json.dump(indexData, index_file, indent=2)
+ hpp_file.write(
+ " };\n"
+ "}\n"
+ )
zip_ref.close()