I'm extracting package descriptions and keywords for offline content analysis. Parse currently returns identity, dependencies, licenses, and scripts, but drops descriptive fields even when the input format is supported.
At library revision 04d388904f3439cbabf24ac986353a7f12266086, using Go 1.27.1:
result, err := manifests.Parse("package.json", []byte(`{
"name": "example",
"version": "1.0.0",
"description": "A currency conversion library",
"keywords": ["currency", "conversion"],
"homepage": "https://example.test"
}`))
fmt.Printf("result=%+v error=%v\n", result, err)
Parsing succeeds, but ParseResult has no fields for the description, keywords, or homepage. The nuspec parser similarly omits description, summary, tags, project URL, and a declared README path. Callers must parse the same file again to recover them. This is a library call; no git-pkgs CLI is involved.
Please expose descriptive metadata alongside the existing results, initially for package.json and nuspec. Preserve summary and long description separately where applicable, declared URLs, content type when supplied, and absent versus explicitly empty values. A declared README path should remain a reference without triggering filesystem access.
Public Parse tests should cover these fields while retaining the existing identity, dependency, license, and script results.
I'm extracting package descriptions and keywords for offline content analysis.
Parsecurrently returns identity, dependencies, licenses, and scripts, but drops descriptive fields even when the input format is supported.At library revision
04d388904f3439cbabf24ac986353a7f12266086, using Go 1.27.1:Parsing succeeds, but
ParseResulthas no fields for the description, keywords, or homepage. The nuspec parser similarly omits description, summary, tags, project URL, and a declared README path. Callers must parse the same file again to recover them. This is a library call; no git-pkgs CLI is involved.Please expose descriptive metadata alongside the existing results, initially for package.json and nuspec. Preserve summary and long description separately where applicable, declared URLs, content type when supplied, and absent versus explicitly empty values. A declared README path should remain a reference without triggering filesystem access.
Public
Parsetests should cover these fields while retaining the existing identity, dependency, license, and script results.