/** * Opt-in capture of every `word/document.xml` the engine emits, so an * external OOXML schema validator can check the whole corpus after a test * run. No-op (zero I/O) unless `SDX_SCHEMA_CORPUS_DIR` is set. * * Files are named by content hash, so repeated emissions of the same XML * deduplicate and the corpus stays bounded across a full test run. The * consumer is `scripts/check_emitted_document_schema.mjs`, which validates * the captured corpus against the vendored Transitional WML schema. * * Capture sites are the three emission choke points: comparison pipelines * (`DocxArchive.setDocumentXml`), primitives saves (`DocxDocument.toBuffer`), * and generation (`generateDocx`). Capture failures propagate: when the * corpus is requested, a partially written corpus must not validate cleanly. * * @see https://github.com/UseJunior/safe-docx/issues/214 */ export declare function maybeCaptureEmittedDocumentXml(xml: string): void; //# sourceMappingURL=schema-corpus-capture.d.ts.map