8000 Merge pull request #8 from clarcharr/master · unicode-rs/rust-caseless@67eb850 · GitHub
[go: up one dir, main page]

Skip to content
Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Appearance settings

Commit 67eb850

Browse files
authored
Merge pull request #8 from clarcharr/master
Update CaseFolding.txt
2 parents 281535c + 237fb8c commit 67eb850

File tree

3 files changed

+238
-12
lines changed

3 files changed

+238
-12
lines changed

Cargo.toml

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,6 @@
11
[package]
2-
32
name = "caseless"
4-
version = "0.1.3"
3+
version = "0.2.0"
54
authors = ["Simon Sapin <simon.sapin@exyr.org>"]
65
description = "Unicode caseless matching"
76
repository = "https://github.com/SimonSapin/rust-caseless"

CaseFolding.txt

Lines changed: 228 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -1,10 +1,11 @@
1-
# CaseFolding-7.0.0.txt
2-
# Date: 2014-04-09, 20:00:56 GMT [MD]
1+
# CaseFolding-10.0.0.txt
2+
# Date: 2017-04-14, 05:40:18 GMT
3+
# © 2017 Unicode®, Inc.
4+
# Unicode and the Unicode Logo are registered trademarks of Unicode, Inc. in the U.S. and other countries.
5+
# For terms of use, see http://www.unicode.org/terms_of_use.html
36
#
47
# Unicode Character Database
5-
# Copyright (c) 1991-2014 Unicode, Inc.
6-
# For terms of use, see http://www.unicode.org/terms_of_use.html
7-
# For documentation, see http://www.unicode.org/reports/tr44/
8+
# For documentation, see http://www.unicode.org/reports/tr44/
89
#
910
# Case Folding Properties
1011
#
@@ -23,7 +24,7 @@
2324
#
2425
# NOTE: case folding does not preserve normalization formats!
2526
#
26-
# For information on case folding, including how to have case folding
27+
# For information on case folding, including how to have case folding
2728
# preserve normalization formats, see Section 3.13 Default Case Algorithms in
2829
# The Unicode Standard.
2930
#
@@ -587,6 +588,21 @@
587588
10C5; C; 2D25; # GEORGIAN CAPITAL LETTER HOE
588589
10C7; C; 2D27; # GEORGIAN CAPITAL LETTER YN
589590
10CD; C; 2D2D; # GEORGIAN CAPITAL LETTER AEN
591+
13F8; C; 13F0; # CHEROKEE SMALL LETTER YE
592+
13F9; C; 13F1; # CHEROKEE SMALL LETTER YI
593+
13FA; C; 13F2; # CHEROKEE SMALL LETTER YO
594+
13FB; C; 13F3; # CHEROKEE SMALL LETTER YU
595+
13FC; C; 13F4; # CHEROKEE SMALL LETTER YV
596+
13FD; C; 13F5; # CHEROKEE SMALL LETTER MV
597+
1C80; C; 0432; # CYRILLIC SMALL LETTER ROUNDED VE
598+
1C81; C; 0434; # CYRILLIC SMALL LETTER LONG-LEGGED DE
599+
1C82; C; 043E; # CYRILLIC SMALL LETTER NARROW O
600+
1C83; C; 0441; # CYRILLIC SMALL LETTER WIDE ES
601+
1C84; C; 0442; # CYRILLIC SMALL LETTER TALL TE
602+
1C85; C; 0442; # CYRILLIC SMALL LETTER THREE-LEGGED TE
603+
1C86; C; 044A; # CYRILLIC SMALL LETTER TALL HARD SIGN
604+
1C87; C; 0463; # CYRILLIC SMALL LETTER TALL YAT
605+
1C88; C; A64B; # CYRILLIC SMALL LETTER UNBLENDED UK
590606
1E00; C; 1E01; # LATIN CAPITAL LETTER A WITH RING BELOW
591607
1E02; C; 1E03; # LATIN CAPITAL LETTER B WITH DOT ABOVE
592608
1E04; C; 1E05; # LATIN CAPITAL LETTER B WITH DOT BELOW
@@ -1157,8 +1173,93 @@ A7AA; C; 0266; # LATIN CAPITAL LETTER H WITH HOOK
11571173
A7AB; C; 025C; # LATIN CAPITAL LETTER REVERSED OPEN E
11581174
A7AC; C; 0261; # LATIN CAPITAL LETTER SCRIPT G
11591175
A7AD; C; 026C; # LATIN CAPITAL LETTER L WITH BELT
1176+
A7AE; C; 026A; # LATIN CAPITAL LETTER SMALL CAPITAL I
11601177
A7B0; C; 029E; # LATIN CAPITAL LETTER TURNED K
11611178
A7B1; C; 0287; # LATIN CAPITAL LETTER TURNED T
1179+
A7B2; C; 029D; # LATIN CAPITAL LETTER J WITH CROSSED-TAIL
1180+
A7B3; C; AB53; # LATIN CAPITAL LETTER CHI
1181+
A7B4; C; A7B5; # LATIN CAPITAL LETTER BETA
1182+
A7B6; C; A7B7; # LATIN CAPITAL LETTER OMEGA
1183+
AB70; C; 13A0; # CHEROKEE SMALL LETTER A
1184+
AB71; C; 13A1; # CHEROKEE SMALL LETTER E
1185+
AB72; C; 13A2; # CHEROKEE SMALL LETTER I
1186+
AB73; C; 13A3; # CHEROKEE SMALL LETTER O
1187+
AB74; C; 13A4; # CHEROKEE SMALL LETTER U
1188+
AB75; C; 13A5; # CHEROKEE SMALL LETTER V
1189+
AB76; C; 13A6; # CHEROKEE SMALL LETTER GA
1190+
AB77; C; 13A7; # CHEROKEE SMALL LETTER KA
1191+
AB78; C; 13A8; # CHEROKEE SMALL LETTER GE
1192+
AB79; C; 13A9; # CHEROKEE SMALL LETTER GI
1193+
AB7A; C; 13AA; # CHEROKEE SMALL LETTER GO
1194+
AB7B; C; 13AB; # CHEROKEE SMALL LETTER GU
1195+
AB7C; C; 13AC; # CHEROKEE SMALL LETTER GV
1196+
AB7D; C; 13AD; # CHEROKEE SMALL LETTER HA
1197+
AB7E; C; 13AE; # CHEROKEE SMALL LETTER HE
1198+
AB7F; C; 13AF; # CHEROKEE SMALL LETTER HI
1199+
AB80; C; 13B0; # CHEROKEE SMALL LETTER HO
1200+
AB81; C; 13B1; # CHEROKEE SMALL LETTER HU
1201+
AB82; C; 13B2; # CHEROKEE SMALL LETTER HV
1202+
AB83; C; 13B3; # CHEROKEE SMALL LETTER LA
1203+
AB84; C; 13B4; # CHEROKEE SMALL LETTER LE
1204+
AB85; C; 13B5; # CHEROKEE SMALL LETTER LI
1205+
AB86; C; 13B6; # CHEROKEE SMALL LETTER LO
1206+
AB87; C; 13B7; # CHEROKEE SMALL LETTER LU
1207+
AB88; C; 13B8; # CHEROKEE SMALL LETTER LV
1208+
AB89; C; 13B9; # CHEROKEE SMALL LETTER MA
1209+
AB8A; C; 13BA; # CHEROKEE SMALL LETTER ME
1210+
AB8B; C; 13BB; # CHEROKEE SMALL LETTER MI
1211+
AB8C; C; 13BC; # CHEROKEE SMALL LETTER MO
1212+
AB8D; C; 13BD; # CHEROKEE SMALL LETTER MU
1213+
AB8E; C; 13BE; # CHEROKEE SMALL LETTER NA
1214+
AB8F; C; 13BF; # CHEROKEE SMALL LETTER HNA
1215+
AB90; C; 13C0; # CHEROKEE SMALL LETTER NAH
1216+
AB91; C; 13C1; # CHEROKEE SMALL LETTER NE
1217+
AB92; C; 13C2; # CHEROKEE SMALL LETTER NI
1218+
AB93; C; 13C3; # CHEROKEE SMALL LETTER NO
1219+
AB94; C; 13C4; # CHEROKEE SMALL LETTER NU
1220+
AB95; C; 13C5; # CHEROKEE SMALL LETTER NV
1221+
AB96; C; 13C6; # CHEROKEE SMALL LETTER QUA
1222+
AB97; C; 13C7; # CHEROKEE SMALL LETTER QUE
1223+
AB98; C; 13C8; # CHEROKEE SMALL LETTER QUI
1224+
AB99; C; 13C9; # CHEROKEE SMALL LETTER QUO
1225+
AB9A; C; 13CA; # CHEROKEE SMALL LETTER QUU
1226+
AB9B; C; 13CB; # CHEROKEE SMALL LETTER QUV
1227+
AB9C; C; 13CC; # CHEROKEE SMALL LETTER SA
1228+
AB9D; C; 13CD; # CHEROKEE SMALL LETTER S
1229+
AB9E; C; 13CE; # CHEROKEE SMALL LETTER SE
1230+
AB9F; C; 13CF; # CHEROKEE SMALL LETTER SI
1231+
ABA0; C; 13D0; # CHEROKEE SMALL LETTER SO
1232+
ABA1; C; 13D1; # CHEROKEE SMALL LETTER SU
1233+
ABA2; C; 13D2; # CHEROKEE SMALL LETTER SV
1234+
ABA3; C; 13D3; # CHEROKEE SMALL LETTER DA
1235+
ABA4; C; 13D4; # CHEROKEE SMALL LETTER TA
1236+
ABA5; C; 13D5; # CHEROKEE SMALL LETTER DE
1237+
ABA6; C; 13D6; # CHEROKEE SMALL LETTER TE
1238+
ABA7; C; 13D7; # CHEROKEE SMALL LETTER DI
1239+
ABA8; C; 13D8; # CHEROKEE SMALL LETTER TI
1240+
ABA9; C; 13D9; # CHEROKEE SMALL LETTER DO
1241+
ABAA; C; 13DA; # CHEROKEE SMALL LETTER DU
1242+
ABAB; C; 13DB; # CHEROKEE SMALL LETTER DV
1243+
ABAC; C; 13DC; # CHEROKEE SMALL LETTER DLA
1244+
ABAD; C; 13DD; # CHEROKEE SMALL LETTER TLA
1245+
ABAE; C; 13DE; # CHEROKEE SMALL LETTER TLE
1246+
ABAF; C; 13DF; # CHEROKEE SMALL LETTER TLI
1247+
ABB0; C; 13E0; # CHEROKEE SMALL LETTER TLO
1248+
ABB1; C; 13E1; # CHEROKEE SMALL LETTER TLU
1249+
ABB2; C; 13E2; # CHEROKEE SMALL LETTER TLV
1250+
ABB3; C; 13E3; # CHEROKEE SMALL LETTER TSA
1251+
ABB4; C; 13E4; # CHEROKEE SMALL LETTER TSE
1252+
ABB5; C; 13E5; # CHEROKEE SMALL LETTER TSI
1253+
ABB6; C; 13E6; # CHEROKEE SMALL LETTER TSO
1254+
ABB7; C; 13E7; # CHEROKEE SMALL LETTER TSU
1255+
ABB8; C; 13E8; # CHEROKEE SMALL LETTER TSV
1256+
ABB9; C; 13E9; # CHEROKEE SMALL LETTER WA
1257+
ABBA; C; 13EA; # CHEROKEE SMALL LETTER WE
1258+
ABBB; C; 13EB; # CHEROKEE SMALL LETTER WI
1259+
ABBC; C; 13EC; # CHEROKEE SMALL LETTER WO
1260+
ABBD; C; 13ED; # CHEROKEE SMALL LETTER WU
1261+
ABBE; C; 13EE; # CHEROKEE SMALL LETTER WV
1262+
ABBF; C; 13EF; # CHEROKEE SMALL LETTER YA
11621263
FB00; F; 0066 0066; # LATIN SMALL LIGATURE FF
11631264
FB01; F; 0066 0069; # LATIN SMALL LIGATURE FI
11641265
FB02; F; 0066 006C; # LATIN SMALL LIGATURE FL
@@ -1237,6 +1338,93 @@ FF3A; C; FF5A; # FULLWIDTH LATIN CAPITAL LETTER Z
12371338
10425; C; 1044D; # DESERET CAPITAL LETTER ENG
12381339
10426; C; 1044E; # DESERET CAPITAL LETTER OI
12391340
10427; C; 1044F; # DESERET CAPITAL LETTER EW
1341+
104B0; C; 104D8; # OSAGE CAPITAL LETTER A
1342+
104B1; C; 104D9; # OSAGE CAPITAL LETTER AI
1343+
104B2; C; 104DA; # OSAGE CAPITAL LETTER AIN
1344+
104B3; C; 104DB; # OSAGE CAPITAL LETTER AH
1345+
104B4; C; 104DC; # OSAGE CAPITAL LETTER BRA
1346+
104B5; C; 104DD; # OSAGE CAPITAL LETTER CHA
1347+
104B6; C; 104DE; # OSAGE CAPITAL LETTER EHCHA
1348+
104B7; C; 104DF; # OSAGE CAPITAL LETTER E
1349+
104B8; C; 104E0; # OSAGE CAPITAL LETTER EIN
1350+
104B9; C; 104E1; # OSAGE CAPITAL LETTER HA
1351+
104BA; C; 104E2; # OSAGE CAPITAL LETTER HYA
1352+
104BB; C; 104E3; # OSAGE CAPITAL LETTER I
1353+
104BC; C; 104E4; # OSAGE CAPITAL LETTER KA
1354+
104BD; C; 104E5; # OSAGE CAPITAL LETTER EHKA
1355+
104BE; C; 104E6; # OSAGE CAPITAL LETTER KYA
1356+
104BF; C; 104E7; # OSAGE CAPITAL LETTER LA
1357+
104C0; C; 104E8; # OSAGE CAPITAL LETTER MA
1358+
104C1; C; 104E9; # OSAGE CAPITAL LETTER NA
1359+
104C2; C; 104EA; # OSAGE CAPITAL LETTER O
1360+
104C3; C; 104EB; # OSAGE CAPITAL LETTER OIN
1361+
104C4; C; 104EC; # OSAGE CAPITAL LETTER PA
1362+
104C5; C; 104ED; # OSAGE CAPITAL LETTER EHPA
1363+
104C6; C; 104EE; # OSAGE CAPITAL LETTER SA
1364+
104C7; C; 104EF; # OSAGE CAPITAL LETTER SHA
1365+
104C8; C; 104F0; # OSAGE CAPITAL LETTER TA
1366+
104C9; C; 104F1; # OSAGE CAPITAL LETTER EHTA
1367+
104CA; C; 104F2; # OSAGE CAPITAL LETTER TSA
1368+
104CB; C; 104F3; # OSAGE CAPITAL LETTER EHTSA
1369+
104CC; C; 104F4; # OSAGE CAPITAL LETTER TSHA
1370+
104CD; C; 104F5; # OSAGE CAPITAL LETTER DHA
1371+
104CE; C; 104F6; # OSAGE CAPITAL LETTER U
1372+
104CF; C; 104F7; # OSAGE CAPITAL LETTER WA
1373+
104D0; C; 104F8; # OSAGE CAPITAL LETTER KHA
1374+
104D1; C; 104F9; # OSAGE CAPITAL LETTER GHA
1375+
104D2; C; 104FA; # OSAGE CAPITAL LETTER ZA
1376+
104D3; C; 104FB; # OSAGE CAPITAL LETTER ZHA
1377+
10C80; C; 10CC0; # OLD HUNGARIAN CAPITAL LETTER A
1378+
10C81; C; 10CC1; # OLD HUNGARIAN CAPITAL LETTER AA
1379+
10C82; C; 10CC2; # OLD HUNGARIAN CAPITAL LETTER EB
1380+
10C83; C; 10CC3; # OLD HUNGARIAN CAPITAL LETTER AMB
1381+
10C84; C; 10CC4; # OLD HUNGARIAN CAPITAL LETTER EC
1382+
10C85; C; 10CC5; # OLD HUNGARIAN CAPITAL LETTER ENC
1383+
10C86; C; 10CC6; # OLD HUNGARIAN CAPITAL LETTER ECS
1384+
10C87; C; 10CC7; # OLD HUNGARIAN CAPITAL LETTER ED
1385+
10C88; C; 10CC8; # OLD HUNGARIAN CAPITAL LETTER AND
1386+
10C89; C; 10CC9; # OLD HUNGARIAN CAPITAL LETTER E
1387+
10C8A; C; 10CCA; # OLD HUNGARIAN CAPITAL LETTER CLOSE E
1388+
10C8B; C; 10CCB; # OLD HUNGARIAN CAPITAL LETTER EE
1389+
10C8C; C; 10CCC; # OLD HUNGARIAN CAPITAL LETTER EF
1390+
10C8D; C; 10CCD; # OLD HUNGARIAN CAPITAL LETTER EG
1391+
10C8E; C; 10CCE; # OLD HUNGARIAN CAPITAL LETTER EGY
1392+
10C8F; C; 10CCF; # OLD HUNGARIAN CAPITAL LETTER EH
1393+
10C90; C; 10CD0; # OLD HUNGARIAN CAPITAL LETTER I
1394+
10C91; C; 10CD1; # OLD HUNGARIAN CAPITAL LETTER II
1395+
10C92; C; 10CD2; # OLD HUNGARIAN CAPITAL LETTER EJ
1396+
10C93; C; 10CD3; # OLD HUNGARIAN CAPITAL LETTER EK
1397+
10C94; C; 10CD4; # OLD HUNGARIAN CAPITAL LETTER AK
1398+
10C95; C; 10CD5; # OLD HUNGARIAN CAPITAL LETTER UNK
1399+
10C96; C; 10CD6; # OLD HUNGARIAN CAPITAL LETTER EL
1400+
10C97; C; 10CD7; # OLD HUNGARIAN CAPITAL LETTER ELY
1401+
10C98; C; 10CD8; # OLD HUNGARIAN CAPITAL LETTER EM
1402+
10C99; C; 10CD9; # OLD HUNGARIAN CAPITAL LETTER EN
1403+
10C9A; C; 10CDA; # OLD HUNGARIAN CAPITAL LETTER ENY
1404+
10C9B; C; 10CDB; # OLD HUNGARIAN CAPITAL LETTER O
1405+
10C9C; C; 10CDC; # OLD HUNGARIAN CAPITAL LETTER OO
1406+
10C9D; C; 10CDD; # OLD HUNGARIAN CAPITAL LETTER NIKOLSBURG OE
1407+
10C9E; C; 10CDE; # OLD HUNGARIAN CAPITAL LETTER RUDIMENTA OE
1408+
10C9F; C; 10CDF; # OLD HUNGARIAN CAPITAL LETTER OEE
1409+
10CA0; C; 10CE0; # OLD HUNGARIAN CAPITAL LETTER EP
1410+
10CA1; C; 10CE1; # OLD HUNGARIAN CAPITAL LETTER EMP
1411+
10CA2; C; 10CE2; # OLD HUNGARIAN CAPITAL LETTER ER
1412+
10CA3; C; 10CE3; # OLD HUNGARIAN CAPITAL LETTER SHORT ER
1413+
10CA4; C; 10CE4; # OLD HUNGARIAN CAPITAL LETTER ES
1414+
10CA5; C; 10CE5; # OLD HUNGARIAN CAPITAL LETTER ESZ
1415+
10CA6; C; 10CE6; # OLD HUNGARIAN CAPITAL LETTER ET
1416+
10CA7; C; 10CE7; # OLD HUNGARIAN CAPITAL LETTER ENT
1417+
10CA8; C; 10CE8; # OLD HUNGARIAN CAPITAL LETTER ETY
1418+
10CA9; C; 10CE9; # OLD HUNGARIAN CAPITAL LETTER ECH
1419+
10CAA; C; 10CEA; # OLD HUNGARIAN CAPITAL LETTER U
1420+
10CAB; C; 10CEB; # OLD HUNGARIAN CAPITAL LETTER UU
1421+
10CAC; C; 10CEC; # OLD HUNGARIAN CAPITAL LETTER NIKOLSBURG UE
1422+
10CAD; C; 10CED; # OLD HUNGARIAN CAPITAL LETTER RUDIMENTA UE
1423+
10CAE; C; 10CEE; # OLD HUNGARIAN CAPITAL LETTER EV
1424+
10CAF; C; 10CEF; # OLD HUNGARIAN CAPITAL LETTER EZ
1425+
10CB0; C; 10CF0; # OLD HUNGARIAN CAPITAL LETTER EZS
1426+
10CB1; C; 10CF1; # OLD HUNGARIAN CAPITAL LETTER ENT-SHAPED SIGN
1427+
10CB2; C; 10CF2; # OLD HUNGARIAN CAPITAL LETTER US
12401428
118A0; C; 118C0; # WARANG CITI CAPITAL LETTER NGAA
12411429
118A1; C; 118C1; # WARANG CITI CAPITAL LETTER A
12421430
118A2; C; 118C2; # WARANG CITI CAPITAL LETTER WI
@@ -1269,5 +1457,39 @@ FF3A; C; FF5A; # FULLWIDTH LATIN CAPITAL LETTER Z
12691457
118BD; C; 118DD; # WARANG CITI CAPITAL LETTER SSUU
12701458
118BE; C; 118DE; # WARANG CITI CAPITAL LETTER SII
12711459
118BF; C; 118DF; # WARANG CITI CAPITAL LETTER VIYO
1460+
1E900; C; 1E922; # ADLAM CAPITAL LETTER ALIF
1461+
1E901; C; 1E923; # ADLAM CAPITAL LETTER DAALI
1462+
1E902; C; 1E924; # ADLAM CAPITAL LETTER LAAM
1463+
1E903; C; 1E925; # ADLAM CAPITAL LETTER MIIM
1464+
1E904; C; 1E926; # ADLAM CAPITAL LETTER BA
1465+
1E905; C; 1E927; # ADLAM CAPITAL LETTER SINNYIIYHE
1466+
1E906; C; 1E928; # ADLAM CAPITAL LETTER PE
1467+
1E907; C; 1E929; # ADLAM CAPITAL LETTER BHE
1468+
1E908; C; 1E92A; # ADLAM CAPITAL LETTER RA
1469+
1E909; C; 1E92B; # ADLAM CAPITAL LETTER E
1470+
1E90A; C; 1E92C; # ADLAM CAPITAL LETTER FA
1471+
1E90B; C; 1E92D; # ADLAM CAPITAL LETTER I
1472+
1E90C; C; 1E92E; # ADLAM CAPITAL LETTER O
1473+
1E90D; C; 1E92F; # ADLAM CAPITAL LETTER DHA
1474+
1E90E; C; 1E930; # ADLAM CAPITAL LETTER YHE
1475+
1E90F; C; 1E931; # ADLAM CAPITAL LETTER WAW
1476+
1E910; C; 1E932; # ADLAM CAPITAL LETTER NUN
1477+
1E911; C; 1E933; # ADLAM CAPITAL LETTER KAF
1478+
1E912; C; 1E934; # ADLAM CAPITAL LETTER YA
1479+
1E913; C; 1E935; # ADLAM CAPITAL LETTER U
1480+
1E914; C; 1E936; # ADLAM CAPITAL LETTER JIIM
1481+
1E915; C; 1E937; # ADLAM CAPITAL LETTER CHI
1482+
1E916; C; 1E938; # ADLAM CAPITAL LETTER HA
1483+
1E917; C; 1E939; # ADLAM CAPITAL LETTER QAAF
1484+
1E918; C; 1E93A; # ADLAM CAPITAL LETTER GA
1485+
1E919; C; 1E93B; # ADLAM CAPITAL LETTER NYA
1486+
1E91A; C; 1E93C; # ADLAM CAPITAL LETTER TU
1487+
1E91B; C; 1E93D; # ADLAM CAPITAL LETTER NHA
1488+
1E91C; C; 1E93E; # ADLAM CAPITAL LETTER VA
1489+
1E91D; C; 1E93F; # ADLAM CAPITAL LETTER KHA
1490+
1E91E; C; 1E940; # ADLAM CAPITAL LETTER GBE
1491+
1E91F; C; 1E941; # ADLAM CAPITAL LETTER ZAL
1492+
1E920; C; 1E942; # ADLAM CAPITAL LETTER KPO
1493+
1E921; C; 1E943; # ADLAM CAPITAL LETTER SHA
12721494
#
12731495
# EOF

src/build.rs

Lines changed: 9 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -13,17 +13,22 @@ const MAX_FOLDED_CODE_POINTS: usize = 3;
1313
fn main() {
1414
let mut lines = include_str!("../CaseFolding.txt").lines();
1515
let first_line = lines.next().unwrap();
16-
let version_regex = Regex::new(r"^# CaseFolding-(\d.\d.\d).txt$").unwrap();
17-
let unicode_version = &version_regex.captures(first_line).unwrap()[1];
16+
let version_regex = Regex::new(r"^# CaseFolding-(\d+)\.(\d+)\.(\d+).txt$").unwrap();
17+
let unicode_version = &version_regex.captures(first_line).unwrap();
18+
let (major, minor, patch): (u64, u64, u64) = (
19+
unicode_version[1].parse().unwrap(),
20+
unicode_version[2].parse().unwrap(),
21+
unicode_version[3].parse().unwrap(),
22+
);
1823

1924
let dst = Path::new(&env::var("OUT_DIR").unwrap()).join("case_folding_data.rs");
20-
let mut f = &mut File::create(&dst).unwrap();
25+
let f = &mut File::create(&dst).unwrap();
2126

2227
macro_rules! w {
2328
($($args: tt)+) => { (write!(f, $($args)+)).unwrap(); }
2429
};
2530

26-
w!("pub const UNICODE_VERSION: &'static str = \"{}\";\n", unicode_version);
31+
w!("pub const UNICODE_VERSION: (u64, u64, u64) = ({}, {}, {});\n", major, minor, patch);
2732
w!("const CASE_FOLDING_TABLE: &'static [(char, [char; 3])] = &[\n");
2833

2934
// Entry with C (common case folding) or F (full case folding) status

0 commit comments

Comments
 (0)
0