scuffed-code/icu4c/source/tools/makeconv/gb18030/gbstates.txt

# ICU state information for the GB 18030 MBCS codepage
# Note that the entire block for the supplementary Unicode planes is
# marked unassigned because they are handled algorithmically.
# Similarly, some of the BMP mappings are marked as unassigned for the same reason.

# Mostly assigned sequences, with branches in the lead bytes
<icu:state> 0-7f, 81:7, 82:8, 83:9, 84:a, 85-fe:4
<icu:state> 30-39:2, 40-7e, 80-fe
<icu:state> 81-fe:3
<icu:state> 30-39

# All-unassigned 4-byte sequences
<icu:state> 30-39:5, 40-7e, 80-fe
<icu:state> 81-fe:6
<icu:state> 30-39.u

# Some unassigned 4-byte sequences, one state for each of the lead bytes 81-84
# Each of these states branch on the second of four bytes; for the third and fourth bytes,
# unassigned sequences continue with state 5, assigned ones with state 2
<icu:state> 30:2, 31-35:5, 36-39:2, 40-7e, 80-fe
<icu:state> 30-35:2, 36-39:5, 40-7e, 80-fe
<icu:state> 30-35:5, 36:2, 37-39:5, 40-7e, 80-fe
<icu:state> 30-31:2, 32-39:5, 40-7e, 80-fe