Puccini TOSCA Quirks

These are activated via the --quirk/-x switch for puccini-tosca:

• imports.implicit.disable: In TOSCA 1.0-1.3 the Simple Profile is implicitly imported by default. This quirk will disable implicit imports.

• imports.version.permissive: By default Puccini will report an error if a unit imports another unit with an incompatible grammar. This quirk will disable the check.

• imports.topology_template.ignore: Allows imported units to contain a topology_template section, which is ignored.

• imports.sequencedlist: Allows the “import” syntax to be a sequenced list, in which the name is ignored.

• data_types.string.permissive: By default Puccini is strict about “string”-typed values and will consider integers, floats, and boolean values to be problems. This quirk will accept such values and convert them as sensibly as possible to strings. This includes accepting floats and integers for the TOSCA “version” primitive type. Note that string conversions may very well not be identical to the literal YAML. For example, 1.0000 in YAML (a float) would become the string 1 in TOSCA.

• data_types.timestamp.permissive: By default Puccini requires all “timestamp” values to be specified as strings in the ISO 8601 format. However, some YAML environments may support the optional !!timestamp type. This quirk will allow such values. Note that such values will not have the “$originalString” key, because the literal YAML is not preserved by the YAML parser.

• namespace.normative.ignore: This will ignore any type that is has the “tosca.normative: true” metadata.

• namespace.normative.shortcuts.disable: In TOSCA 1.0-1.3 all the normative types have long names, such as “tosca.nodes.Compute”, prefixed names (“tosca:Compute”), and also short names (“Compute”). Those short names are annoying because it means you can’t use those names for your own types. This quirk disables the short names (the prefixed names remain).

• substitution_mappings.requirements.list: According to the examples in the TOSCA 1.0-1.3 specs, the requirements key under substitution_mappings is syntactically a map. However, this syntax is inconsistent because it doesn’t match the syntax in node templates, which is a sequenced list. (In node types, too, it is a sequenced list, although grammatically it works like a map.) This quirk changes the expected syntax to be a sequenced list.

• substitution_mappings.requirements.permissive: Normally the requirements under substitution_mappings must be mapped to an assigned requirement in a node template. This quirk allows unassigned requirements to be mapped.

• annotations.ignore: Ignores the “annotation_types” keyword in service templates and the “annotations” keyword in parameter definitions.