Method _create_pax_generic_header

Lib/tarfile.py:1228–1276

Return a POSIX.1-2008 extended or global header sequence that contains a list of keyword, value pairs. The values must be strings.

(cls, pax_headers, type, encoding)

    @classmethod
    def _create_pax_generic_header(cls, pax_headers, type, encoding):
        """Return a POSIX.1-2008 extended or global header sequence
           that contains a list of keyword, value pairs. The values
           must be strings.
        """
        # Check if one of the fields contains surrogate characters and thereby
        # forces hdrcharset=BINARY, see _proc_pax() for more information.
        binary = False
        for keyword, value in pax_headers.items():
            try:
                value.encode("utf-8", "strict")
            except UnicodeEncodeError:
                binary = True
                break

        records = b""
        if binary:
            # Put the hdrcharset field at the beginning of the header.
            records += b"21 hdrcharset=BINARY\n"

        for keyword, value in pax_headers.items():
            keyword = keyword.encode("utf-8")
            if binary:
                # Try to restore the original byte representation of 'value'.
                # Needless to say, that the encoding must match the string.
                value = value.encode(encoding, "surrogateescape")
            else:
                value = value.encode("utf-8")

            l = len(keyword) + len(value) + 3   # ' ' + '=' + '\n'
            n = p = 0
            while True:
                n = l + len(str(p))
                if n == p:
                    break
                p = n
            records += bytes(str(p), "ascii") + b" " + keyword + b"=" + value + b"\n"

        # We use a hardcoded "././@PaxHeader" name like star does
        # instead of the one that POSIX recommends.
        info = {}
        info["name"] = "././@PaxHeader"
        info["type"] = type
        info["size"] = len(records)
        info["magic"] = POSIX_MAGIC

        # Create pax header + record blocks.
        return cls._create_header(info, USTAR_FORMAT, "ascii", "replace") + \
                cls._create_payload(records)

    @classmethod
    def frombuf(cls, buf, encoding, errors):