eemeli
parseDocument fails to parse arrays correctly
Describe the bug
parseDocument not parsing some arrays correctly, for instance:
A: 1
B: [ {
"foo": "bar"
} ]
C: 3
In this example, C will not appear in the parsed document contents
Additionally, multiple errors will appear:
YAMLParseError: Flow map in block collection must be sufficiently indented and end with a } atYAMLParseError: Flow sequence in block collection must be sufficiently indented and end with a ]YAMLParseError: Unexpected flow-map-end token in YAML stream: "}"YAMLParseError: Unexpected flow-seq-end token in YAML stream: "]"YAMLParseError: Unexpected scalar token in YAML stream: "C"YAMLParseError: Unexpected map-value-ind token in YAML stream: ":"YAMLParseError: Unexpected scalar token in YAML stream:
To Reproduce
import util from 'node:util';
import { parseDocument } from 'yaml';
const raw = `A: 1\nB: [ {\n "foo": "bar"\n} ]\nC: 3`;
const doc = parseDocument(raw);
console.log(doc.errors) // many errors
console.log(util.inspect(doc.contents, { depth: null })); // C:3 is missing
Expected behaviour
The YAML from the example, which is a valid YAML document, should be parsed correctly and without errors.
Versions (please complete the following information):
- Environment: NodeJS 22.13.1
yaml: 2.6.0, 2.8.1
The example is not a valid yaml document. As the first error indicates, the } is insufficiently indented.
Values in block mappings must be more indented than their key. There is a specific relaxation for the closing indicator of a single flow collection being allowed on the same indent level as the key, but this is not extended to multiple closing indicators, as you have here.
@eemeli hmm I do see now that there's inconsistency between different tools regarding this behavior.
Just FYI my use-case involves AWS's CloudFormation templates, which accept this pattern.