List of symbols that control ISO-2022 encoder/decoder.
The value of the `:flags' attribute in the argument of the function
`define-coding-system' must be one of them.
If `long-form' is specified, use a long designation sequence on
encoding for the charsets `japanese-jisx0208-1978', `chinese-gb2312',
and `japanese-jisx0208'. The long designation sequence doesn't
conform to ISO 2022, but is used by such coding systems as
`compound-text'.
If `ascii-at-eol' is specified, designate ASCII to g0 at end of line
on encoding.
If `ascii-at-cntl' is specified, designate ASCII to g0 before control
codes and SPC on encoding.
If `7-bit' is specified, use 7-bit code only on encoding.
If `locking-shift' is specified, decode locking-shift code correctly
on decoding, and use locking-shift to invoke a graphic element on
encoding.
If `single-shift' is specified, decode single-shift code correctly on
decoding, and use single-shift to invoke a graphic element on encoding.
If `designation' is specified, decode designation code correctly on
decoding, and use designation to designate a charset to a graphic
element on encoding.
If `revision' is specified, produce an escape sequence to specify
revision number of a charset on encoding. Such an escape sequence is
always correctly decoded on decoding.
If `direction' is specified, decode ISO6429's code for specifying
direction correctly, and produce the code on encoding.
If `init-at-bol' is specified, on encoding, it is assumed that
invocation and designation statuses are reset at each beginning of
line even if `ascii-at-eol' is not specified; thus no codes for
resetting them are produced.
If `safe' is specified, on encoding, characters not supported by a
coding are replaced with `?'.
If `latin-extra' is specified, the code-detection routine assumes that a
code specified in `latin-extra-code-table' (which see) is valid.
If `composition' is specified, an escape sequence to specify
composition sequence is correctly decoded on decoding, and is produced
on encoding.
If `euc-tw-shift' is specified, the EUC-TW specific shifting code is
correctly decoded on decoding, and is produced on encoding.
If `use-roman' is specified, JIS0201-1976-Roman is designated instead
of ASCII.
If `use-oldjis' is specified, JIS0208-1976 is designated instead of
JIS0208-1983.