xml: fix DataClassification unmarshal over-reading past its element

[arpitjain099]

DataClassification.UnmarshalXML terminates its token loop on </dataflow> (a hardcoded literal), but the same type is also bound to the <classification> element via EvidenceData.Classification (xml:"data>classification"), not only <dataflow> (xml:"data>dataflow"). When decoding a <classification> element the loop never sees </dataflow>, so it keeps reading past its own end tag and consumes following tokens until io.EOF, which surfaces as a decode error. Any BOM that carries an evidence-data classification then fails to decode, and content after it is silently dropped.

The fix compares against the element the unmarshaler was actually called for (start.Name) instead of a literal, which is the usual depth-0 termination pattern for a custom UnmarshalXML. Child elements are still consumed by the inner loops, so the first top-level end element is always this element's own close.

I added a round-trip test with a 1.6 BOM whose declarations carry an evidence-data classification followed by a components block: it fails to decode with EOF before the change and decodes cleanly (classification parsed, components preserved) after.

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}