Byte-level decoding for the DOM layer. @xmldom takes a JavaScript string,
so a part stored as UTF-16 must be decoded from its ORIGINAL BYTES first —
reading it as UTF-8 (fs.readFile(file, "utf-8")) yields NUL-interleaved
garbage that no parser can rescue.
Encoding selection follows XML 1.0 §4.3.3: the byte order mark wins, then
the encoding pseudo-attribute of the XML declaration. Only the three
encodings OOXML parts actually use are supported — UTF-8, UTF-16LE,
UTF-16BE — and anything else falls back to UTF-8 rather than guessing.
The result carries no BOM and no leading whitespace, so callers can hand it
straight to parseXml.
A caller that writes the part back must use decodeXmlPart, which
reports the encoding it found and refuses an encoding it cannot decode.
Byte-level decoding for the DOM layer.
@xmldomtakes a JavaScript string, so a part stored as UTF-16 must be decoded from its ORIGINAL BYTES first — reading it as UTF-8 (fs.readFile(file, "utf-8")) yields NUL-interleaved garbage that no parser can rescue.Encoding selection follows XML 1.0 §4.3.3: the byte order mark wins, then the
encodingpseudo-attribute of the XML declaration. Only the three encodings OOXML parts actually use are supported — UTF-8, UTF-16LE, UTF-16BE — and anything else falls back to UTF-8 rather than guessing. The result carries no BOM and no leading whitespace, so callers can hand it straight to parseXml.A caller that writes the part back must use decodeXmlPart, which reports the encoding it found and refuses an encoding it cannot decode.