gen-os2-unicode-ranges.py (1239B)
1 #!/usr/bin/env python3 2 3 """Generates the code for a sorted unicode range array as used in hb-ot-os2-unicode-ranges.hh 4 Input is a tab separated list of unicode ranges from the otspec 5 (https://docs.microsoft.com/en-us/typography/opentype/spec/os2#ur). 6 """ 7 8 import re 9 import sys 10 11 12 print ("""static OS2Range _hb_os2_unicode_ranges[] = 13 {""") 14 15 args = sys.argv[1:] 16 input_file = args[0] 17 18 with open (input_file, mode="r", encoding="utf-8") as f: 19 20 all_ranges = [] 21 current_bit = 0 22 while True: 23 line = f.readline().strip() 24 if not line: 25 break 26 fields = re.split(r'\t+', line) 27 if len(fields) == 3: 28 current_bit = fields[0] 29 fields = fields[1:] 30 elif len(fields) > 3: 31 raise Exception("bad input :(.") 32 33 name = fields[0] 34 ranges = re.split("-", fields[1]) 35 if len(ranges) != 2: 36 raise Exception("bad input :(.") 37 38 v = tuple((int(ranges[0], 16), int(ranges[1], 16), int(current_bit), name)) 39 all_ranges.append(v) 40 41 all_ranges = sorted(all_ranges, key=lambda t: t[0]) 42 43 for ranges in all_ranges: 44 start = ("0x%X" % ranges[0]).rjust(8) 45 end = ("0x%X" % ranges[1]).rjust(8) 46 bit = ("%s" % ranges[2]).rjust(3) 47 48 print (" {%s, %s, %s}, // %s" % (start, end, bit, ranges[3])) 49 50 print ("""};""")