~mdw / disorder / blame
| Commit | Line | Data |
|---|---|---|
| 460b9539 | 1 | /* |
| 2 | * This file is part of DisOrder. | |
| 033fd4e3 | 3 | * Copyright (C) 2005, 2007 Richard Kettlewell |
| 460b9539 | 4 | * |
| 5 | * This program is free software; you can redistribute it and/or modify | |
| 6 | * it under the terms of the GNU General Public License as published by | |
| 7 | * the Free Software Foundation; either version 2 of the License, or | |
| 8 | * (at your option) any later version. | |
| 9 | * | |
| 10 | * This program is distributed in the hope that it will be useful, but | |
| 11 | * WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
| 13 | * General Public License for more details. | |
| 14 | * | |
| 15 | * You should have received a copy of the GNU General Public License | |
| 16 | * along with this program; if not, write to the Free Software | |
| 17 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 | |
| 18 | * USA | |
| 19 | */ | |
| 033fd4e3 | 20 | /** @file lib/test.c @brief Library tests */ |
| 460b9539 | 21 | |
| 22 | #include <config.h> | |
| 23 | #include "types.h" | |
| 24 | ||
| 25 | #include <stdio.h> | |
| 26 | #include <string.h> | |
| 27 | #include <stdlib.h> | |
| 28 | #include <errno.h> | |
| 29 | #include <ctype.h> | |
| 033fd4e3 | 30 | #include <assert.h> |
| e5a5a138 RK |
31 | #include <sys/types.h> |
| 32 | #include <sys/stat.h> | |
| 9f28e855 | 33 | #include <unistd.h> |
| 0cd21fc4 | 34 | #include <signal.h> |
| 460b9539 | 35 | |
| 36 | #include "utf8.h" | |
| 37 | #include "mem.h" | |
| 38 | #include "log.h" | |
| 39 | #include "vector.h" | |
| 40 | #include "charset.h" | |
| 41 | #include "mime.h" | |
| 42 | #include "hex.h" | |
| 033fd4e3 | 43 | #include "heap.h" |
| e5a5a138 RK |
44 | #include "unicode.h" |
| 45 | #include "inputline.h" | |
| e2452add | 46 | #include "wstat.h" |
| ea387d53 | 47 | #include "signame.h" |
| 9f28e855 | 48 | #include "cache.h" |
| 00e36cd0 | 49 | #include "filepart.h" |
| 65bb0fff RK |
50 | #include "hash.h" |
| 51 | #include "selection.h" | |
| 460b9539 | 52 | |
| 53 | static int tests, errors; | |
| bb48024f RK |
54 | static int fail_first; |
| 55 | ||
| 56 | static void count_error() { | |
| 57 | ++errors; | |
| 58 | if(fail_first) | |
| 59 | abort(); | |
| 60 | } | |
| 460b9539 | 61 | |
| 033fd4e3 | 62 | /** @brief Checks that @p expr is nonzero */ |
| 460b9539 | 63 | #define insist(expr) do { \ |
| 033fd4e3 | 64 | if(!(expr)) { \ |
| bb48024f | 65 | count_error(); \ |
| 460b9539 | 66 | fprintf(stderr, "%s:%d: error checking %s\n", \ |
| 67 | __FILE__, __LINE__, #expr); \ | |
| 68 | } \ | |
| 69 | ++tests; \ | |
| 70 | } while(0) | |
| 71 | ||
| 72 | static const char *format(const char *s) { | |
| 73 | struct dynstr d; | |
| 74 | int c; | |
| 75 | char buf[10]; | |
| 76 | ||
| 77 | dynstr_init(&d); | |
| 78 | while((c = (unsigned char)*s++)) { | |
| 79 | if(c >= ' ' && c <= '~') | |
| 80 | dynstr_append(&d, c); | |
| 81 | else { | |
| 82 | sprintf(buf, "\\x%02X", (unsigned)c); | |
| 83 | dynstr_append_string(&d, buf); | |
| 84 | } | |
| 85 | } | |
| 86 | dynstr_terminate(&d); | |
| 87 | return d.vec; | |
| 88 | } | |
| 89 | ||
| e5a5a138 RK |
90 | static const char *format_utf32(const uint32_t *s) { |
| 91 | struct dynstr d; | |
| 92 | uint32_t c; | |
| 93 | char buf[64]; | |
| 94 | ||
| 95 | dynstr_init(&d); | |
| 96 | while((c = *s++)) { | |
| 16506c9d RK |
97 | sprintf(buf, " %04lX", (long)c); |
| 98 | dynstr_append_string(&d, buf); | |
| e5a5a138 RK |
99 | } |
| 100 | dynstr_terminate(&d); | |
| 101 | return d.vec; | |
| 102 | } | |
| 103 | ||
| 460b9539 | 104 | #define check_string(GOT, WANT) do { \ |
| 105 | const char *g = GOT; \ | |
| 106 | const char *w = WANT; \ | |
| 107 | \ | |
| 108 | if(w == 0) { \ | |
| 109 | fprintf(stderr, "%s:%d: %s returned 0\n", \ | |
| 110 | __FILE__, __LINE__, #GOT); \ | |
| bb48024f | 111 | count_error(); \ |
| 460b9539 | 112 | } else if(strcmp(w, g)) { \ |
| 113 | fprintf(stderr, "%s:%d: %s returned:\n%s\nexpected:\n%s\n", \ | |
| 114 | __FILE__, __LINE__, #GOT, format(g), format(w)); \ | |
| bb48024f | 115 | count_error(); \ |
| 460b9539 | 116 | } \ |
| 117 | ++tests; \ | |
| 118 | } while(0) | |
| 119 | ||
| 120 | static uint32_t *ucs4parse(const char *s) { | |
| 121 | struct dynstr_ucs4 d; | |
| 122 | char *e; | |
| 123 | ||
| 124 | dynstr_ucs4_init(&d); | |
| 125 | while(*s) { | |
| 126 | errno = 0; | |
| 127 | dynstr_ucs4_append(&d, strtoul(s, &e, 0)); | |
| 128 | if(errno) fatal(errno, "strtoul (%s)", s); | |
| 129 | s = e; | |
| 130 | } | |
| 131 | dynstr_ucs4_terminate(&d); | |
| 132 | return d.vec; | |
| 133 | } | |
| 134 | ||
| 135 | static void test_utf8(void) { | |
| 136 | /* Test validutf8, convert to UCS-4, check the answer is right, | |
| 137 | * convert back to UTF-8, check we got to where we started */ | |
| 138 | #define U8(CHARS, WORDS) do { \ | |
| 139 | uint32_t *w = ucs4parse(WORDS); \ | |
| 140 | uint32_t *ucs; \ | |
| 141 | char *u8; \ | |
| 142 | \ | |
| 143 | insist(validutf8(CHARS)); \ | |
| caecd4f4 | 144 | ucs = utf8_to_utf32(CHARS, strlen(CHARS), 0); \ |
| 460b9539 | 145 | insist(ucs != 0); \ |
| caecd4f4 RK |
146 | insist(!utf32_cmp(w, ucs)); \ |
| 147 | u8 = utf32_to_utf8(ucs, utf32_len(ucs), 0); \ | |
| 460b9539 | 148 | insist(u8 != 0); \ |
| 149 | insist(!strcmp(u8, CHARS)); \ | |
| 150 | } while(0) | |
| 151 | ||
| 033fd4e3 | 152 | fprintf(stderr, "test_utf8\n"); |
| 18cda350 | 153 | #define validutf8(S) utf8_valid((S), strlen(S)) |
| 033fd4e3 | 154 | |
| 460b9539 | 155 | /* empty string */ |
| 156 | ||
| 157 | U8("", ""); | |
| 158 | ||
| 159 | /* ASCII characters */ | |
| 160 | ||
| 161 | U8(" !\"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~", | |
| 162 | "0x20 0x21 0x22 0x23 0x24 0x25 0x26 0x27 0x28 0x29 0x2a 0x2b 0x2c 0x2d " | |
| 163 | "0x2e 0x2f 0x30 0x31 0x32 0x33 0x34 0x35 0x36 0x37 0x38 0x39 0x3a " | |
| 164 | "0x3b 0x3c 0x3d 0x3e 0x3f 0x40 0x41 0x42 0x43 0x44 0x45 0x46 0x47 " | |
| 165 | "0x48 0x49 0x4a 0x4b 0x4c 0x4d 0x4e 0x4f 0x50 0x51 0x52 0x53 0x54 " | |
| 166 | "0x55 0x56 0x57 0x58 0x59 0x5a 0x5b 0x5c 0x5d 0x5e 0x5f 0x60 0x61 " | |
| 167 | "0x62 0x63 0x64 0x65 0x66 0x67 0x68 0x69 0x6a 0x6b 0x6c 0x6d 0x6e " | |
| 168 | "0x6f 0x70 0x71 0x72 0x73 0x74 0x75 0x76 0x77 0x78 0x79 0x7a 0x7b " | |
| 169 | "0x7c 0x7d 0x7e"); | |
| 170 | U8("\001\002\003\004\005\006\007\010\011\012\013\014\015\016\017\020\021\022\023\024\025\026\027\030\031\032\033\034\035\036\037\177", | |
| 171 | "0x1 0x2 0x3 0x4 0x5 0x6 0x7 0x8 0x9 0xa 0xb 0xc 0xd 0xe 0xf 0x10 " | |
| 172 | "0x11 0x12 0x13 0x14 0x15 0x16 0x17 0x18 0x19 0x1a 0x1b 0x1c 0x1d " | |
| 173 | "0x1e 0x1f 0x7f"); | |
| 174 | ||
| 175 | /* from RFC3629 */ | |
| 176 | ||
| 177 | /* UTF8-2 = %xC2-DF UTF8-tail */ | |
| 178 | insist(!validutf8("\xC0\x80")); | |
| 179 | insist(!validutf8("\xC1\x80")); | |
| 180 | insist(!validutf8("\xC2\x7F")); | |
| 181 | U8("\xC2\x80", "0x80"); | |
| 182 | U8("\xDF\xBF", "0x7FF"); | |
| 183 | insist(!validutf8("\xDF\xC0")); | |
| 184 | ||
| 185 | /* UTF8-3 = %xE0 %xA0-BF UTF8-tail / %xE1-EC 2( UTF8-tail ) / | |
| 186 | * %xED %x80-9F UTF8-tail / %xEE-EF 2( UTF8-tail ) | |
| 187 | */ | |
| 188 | insist(!validutf8("\xE0\x9F\x80")); | |
| 189 | U8("\xE0\xA0\x80", "0x800"); | |
| 190 | U8("\xE0\xBF\xBF", "0xFFF"); | |
| 191 | insist(!validutf8("\xE0\xC0\xBF")); | |
| 192 | ||
| 193 | insist(!validutf8("\xE1\x80\x7F")); | |
| 194 | U8("\xE1\x80\x80", "0x1000"); | |
| 195 | U8("\xEC\xBF\xBF", "0xCFFF"); | |
| 196 | insist(!validutf8("\xEC\xC0\xBF")); | |
| 197 | ||
| 198 | U8("\xED\x80\x80", "0xD000"); | |
| 199 | U8("\xED\x9F\xBF", "0xD7FF"); | |
| 200 | insist(!validutf8("\xED\xA0\xBF")); | |
| 201 | ||
| 202 | insist(!validutf8("\xEE\x7f\x80")); | |
| 203 | U8("\xEE\x80\x80", "0xE000"); | |
| 204 | U8("\xEF\xBF\xBF", "0xFFFF"); | |
| 205 | insist(!validutf8("\xEF\xC0\xBF")); | |
| 206 | ||
| 207 | /* UTF8-4 = %xF0 %x90-BF 2( UTF8-tail ) / %xF1-F3 3( UTF8-tail ) / | |
| 208 | * %xF4 %x80-8F 2( UTF8-tail ) | |
| 209 | */ | |
| 210 | insist(!validutf8("\xF0\x8F\x80\x80")); | |
| 211 | U8("\xF0\x90\x80\x80", "0x10000"); | |
| 212 | U8("\xF0\xBF\xBF\xBF", "0x3FFFF"); | |
| 213 | insist(!validutf8("\xF0\xC0\x80\x80")); | |
| 214 | ||
| 215 | insist(!validutf8("\xF1\x80\x80\x7F")); | |
| 216 | U8("\xF1\x80\x80\x80", "0x40000"); | |
| 217 | U8("\xF3\xBF\xBF\xBF", "0xFFFFF"); | |
| 218 | insist(!validutf8("\xF3\xC0\x80\x80")); | |
| 219 | ||
| 220 | insist(!validutf8("\xF4\x80\x80\x7F")); | |
| 221 | U8("\xF4\x80\x80\x80", "0x100000"); | |
| 222 | U8("\xF4\x8F\xBF\xBF", "0x10FFFF"); | |
| 223 | insist(!validutf8("\xF4\x90\x80\x80")); | |
| c35e83d9 | 224 | insist(!validutf8("\xF4\x80\xFF\x80")); |
| 460b9539 | 225 | |
| 226 | /* miscellaneous non-UTF-8 rubbish */ | |
| 227 | insist(!validutf8("\x80")); | |
| 228 | insist(!validutf8("\xBF")); | |
| 229 | insist(!validutf8("\xC0")); | |
| 230 | insist(!validutf8("\xC0\x7F")); | |
| 231 | insist(!validutf8("\xC0\xC0")); | |
| 232 | insist(!validutf8("\xE0")); | |
| 233 | insist(!validutf8("\xE0\x7F")); | |
| 234 | insist(!validutf8("\xE0\xC0")); | |
| 235 | insist(!validutf8("\xE0\x80")); | |
| 236 | insist(!validutf8("\xE0\x80\x7f")); | |
| 237 | insist(!validutf8("\xE0\x80\xC0")); | |
| 238 | insist(!validutf8("\xF0")); | |
| 239 | insist(!validutf8("\xF0\x7F")); | |
| 240 | insist(!validutf8("\xF0\xC0")); | |
| 241 | insist(!validutf8("\xF0\x80")); | |
| 242 | insist(!validutf8("\xF0\x80\x7f")); | |
| 243 | insist(!validutf8("\xF0\x80\xC0")); | |
| 244 | insist(!validutf8("\xF0\x80\x80\x7f")); | |
| 245 | insist(!validutf8("\xF0\x80\x80\xC0")); | |
| 246 | insist(!validutf8("\xF5\x80\x80\x80")); | |
| 247 | insist(!validutf8("\xF8")); | |
| 248 | } | |
| 249 | ||
| 250 | static void test_mime(void) { | |
| 251 | char *t, *n, *v; | |
| 252 | ||
| 033fd4e3 RK |
253 | fprintf(stderr, "test_mime\n"); |
| 254 | ||
| 460b9539 | 255 | t = n = v = 0; |
| 256 | insist(!mime_content_type("text/plain", &t, &n, &v)); | |
| 257 | insist(!strcmp(t, "text/plain")); | |
| 258 | insist(n == 0); | |
| 259 | insist(v == 0); | |
| 260 | ||
| 261 | t = n = v = 0; | |
| 262 | insist(!mime_content_type("TEXT ((nested) comment) /plain", &t, &n, &v)); | |
| 263 | insist(!strcmp(t, "text/plain")); | |
| 264 | insist(n == 0); | |
| 265 | insist(v == 0); | |
| 266 | ||
| 267 | t = n = v = 0; | |
| 268 | insist(!mime_content_type(" text/plain ; Charset=utf-8", &t, &n, &v)); | |
| 269 | insist(!strcmp(t, "text/plain")); | |
| 270 | insist(!strcmp(n, "charset")); | |
| 271 | insist(!strcmp(v, "utf-8")); | |
| 272 | ||
| 273 | t = n = v = 0; | |
| 274 | insist(!mime_content_type("text/plain;charset = ISO-8859-1 ", &t, &n, &v)); | |
| 275 | insist(!strcmp(t, "text/plain")); | |
| 276 | insist(!strcmp(n, "charset")); | |
| 277 | insist(!strcmp(v, "ISO-8859-1")); | |
| 278 | ||
| 279 | /* XXX mime_parse */ | |
| 280 | /* XXX mime_multipart */ | |
| 281 | /* XXX mime_rfc2388_content_disposition */ | |
| 282 | ||
| 283 | check_string(mime_qp(""), ""); | |
| 284 | check_string(mime_qp("foobar"), "foobar"); | |
| 285 | check_string(mime_qp("foo=20bar"), "foo bar"); | |
| 286 | check_string(mime_qp("x \r\ny"), "x\r\ny"); | |
| 287 | check_string(mime_qp("x=\r\ny"), "xy"); | |
| 288 | check_string(mime_qp("x= \r\ny"), "xy"); | |
| 289 | check_string(mime_qp("x =\r\ny"), "x y"); | |
| 290 | check_string(mime_qp("x = \r\ny"), "x y"); | |
| 291 | ||
| 292 | /* from RFC2045 */ | |
| 293 | check_string(mime_qp("Now's the time =\r\n" | |
| 294 | "for all folk to come=\r\n" | |
| 295 | " to the aid of their country."), | |
| 296 | "Now's the time for all folk to come to the aid of their country."); | |
| 297 | ||
| 298 | check_string(mime_base64(""), ""); | |
| 299 | check_string(mime_base64("BBBB"), "\x04\x10\x41"); | |
| 300 | check_string(mime_base64("////"), "\xFF\xFF\xFF"); | |
| 301 | check_string(mime_base64("//BB"), "\xFF\xF0\x41"); | |
| 302 | check_string(mime_base64("BBBB//BB////"), | |
| 303 | "\x04\x10\x41" "\xFF\xF0\x41" "\xFF\xFF\xFF"); | |
| 304 | check_string(mime_base64("B B B B / / B B / / / /"), | |
| 305 | "\x04\x10\x41" "\xFF\xF0\x41" "\xFF\xFF\xFF"); | |
| 306 | check_string(mime_base64("B\r\nBBB.// B-B//~//"), | |
| 307 | "\x04\x10\x41" "\xFF\xF0\x41" "\xFF\xFF\xFF"); | |
| 308 | check_string(mime_base64("BBBB="), | |
| 309 | "\x04\x10\x41"); | |
| 310 | check_string(mime_base64("BBBBx="), /* not actually valid base64 */ | |
| 311 | "\x04\x10\x41"); | |
| 312 | check_string(mime_base64("BBBB BB=="), | |
| 313 | "\x04\x10\x41" "\x04"); | |
| 314 | check_string(mime_base64("BBBB BBB="), | |
| 315 | "\x04\x10\x41" "\x04\x10"); | |
| 316 | } | |
| 317 | ||
| 318 | static void test_hex(void) { | |
| 319 | unsigned n; | |
| 320 | static const unsigned char h[] = { 0x00, 0xFF, 0x80, 0x7F }; | |
| 321 | uint8_t *u; | |
| 322 | size_t ul; | |
| 323 | ||
| 033fd4e3 RK |
324 | fprintf(stderr, "test_hex\n"); |
| 325 | ||
| 460b9539 | 326 | for(n = 0; n <= UCHAR_MAX; ++n) { |
| 327 | if(!isxdigit(n)) | |
| 328 | insist(unhexdigitq(n) == -1); | |
| 329 | } | |
| 330 | insist(unhexdigitq('0') == 0); | |
| 331 | insist(unhexdigitq('1') == 1); | |
| 332 | insist(unhexdigitq('2') == 2); | |
| 333 | insist(unhexdigitq('3') == 3); | |
| 334 | insist(unhexdigitq('4') == 4); | |
| 335 | insist(unhexdigitq('5') == 5); | |
| 336 | insist(unhexdigitq('6') == 6); | |
| 337 | insist(unhexdigitq('7') == 7); | |
| 338 | insist(unhexdigitq('8') == 8); | |
| 339 | insist(unhexdigitq('9') == 9); | |
| 340 | insist(unhexdigitq('a') == 10); | |
| 341 | insist(unhexdigitq('b') == 11); | |
| 342 | insist(unhexdigitq('c') == 12); | |
| 343 | insist(unhexdigitq('d') == 13); | |
| 344 | insist(unhexdigitq('e') == 14); | |
| 345 | insist(unhexdigitq('f') == 15); | |
| 346 | insist(unhexdigitq('A') == 10); | |
| 347 | insist(unhexdigitq('B') == 11); | |
| 348 | insist(unhexdigitq('C') == 12); | |
| 349 | insist(unhexdigitq('D') == 13); | |
| 350 | insist(unhexdigitq('E') == 14); | |
| 351 | insist(unhexdigitq('F') == 15); | |
| 352 | check_string(hex(h, sizeof h), "00ff807f"); | |
| 353 | check_string(hex(0, 0), ""); | |
| 354 | u = unhex("00ff807f", &ul); | |
| 355 | insist(ul == 4); | |
| 356 | insist(memcmp(u, h, 4) == 0); | |
| 357 | u = unhex("00FF807F", &ul); | |
| 358 | insist(ul == 4); | |
| 359 | insist(memcmp(u, h, 4) == 0); | |
| 360 | u = unhex("", &ul); | |
| 361 | insist(ul == 0); | |
| 033fd4e3 | 362 | fprintf(stderr, "2 ERROR reports expected {\n"); |
| 460b9539 | 363 | insist(unhex("F", 0) == 0); |
| 364 | insist(unhex("az", 0) == 0); | |
| 033fd4e3 | 365 | fprintf(stderr, "}\n"); |
| 460b9539 | 366 | } |
| 367 | ||
| 368 | static void test_casefold(void) { | |
| e5a5a138 | 369 | uint32_t c, l; |
| 56fd389c | 370 | const char *input, *canon_folded, *compat_folded, *canon_expected, *compat_expected; |
| 460b9539 | 371 | |
| 033fd4e3 | 372 | fprintf(stderr, "test_casefold\n"); |
| 56fd389c RK |
373 | |
| 374 | /* This isn't a very exhaustive test. Unlike for normalization, there don't | |
| 375 | * seem to be any public test vectors for these algorithms. */ | |
| e5a5a138 | 376 | |
| 460b9539 | 377 | for(c = 1; c < 256; ++c) { |
| e5a5a138 | 378 | input = utf32_to_utf8(&c, 1, 0); |
| 56fd389c RK |
379 | canon_folded = utf8_casefold_canon(input, strlen(input), 0); |
| 380 | compat_folded = utf8_casefold_compat(input, strlen(input), 0); | |
| 460b9539 | 381 | switch(c) { |
| 382 | default: | |
| 383 | if((c >= 'A' && c <= 'Z') | |
| 384 | || (c >= 0xC0 && c <= 0xDE && c != 0xD7)) | |
| 385 | l = c ^ 0x20; | |
| 386 | else | |
| 387 | l = c; | |
| 388 | break; | |
| 389 | case 0xB5: /* MICRO SIGN */ | |
| e5a5a138 | 390 | l = 0x3BC; /* GREEK SMALL LETTER MU */ |
| 460b9539 | 391 | break; |
| 392 | case 0xDF: /* LATIN SMALL LETTER SHARP S */ | |
| 56fd389c RK |
393 | insist(!strcmp(canon_folded, "ss")); |
| 394 | insist(!strcmp(compat_folded, "ss")); | |
| 460b9539 | 395 | l = 0; |
| 396 | break; | |
| 397 | } | |
| 398 | if(l) { | |
| caecd4f4 | 399 | uint32_t *d; |
| e5a5a138 | 400 | /* Case-folded data is now normalized */ |
| caecd4f4 RK |
401 | d = utf32_decompose_canon(&l, 1, 0); |
| 402 | canon_expected = utf32_to_utf8(d, utf32_len(d), 0); | |
| 56fd389c RK |
403 | if(strcmp(canon_folded, canon_expected)) { |
| 404 | fprintf(stderr, "%s:%d: canon-casefolding %#lx got '%s', expected '%s'\n", | |
| 405 | __FILE__, __LINE__, (unsigned long)c, | |
| 406 | format(canon_folded), format(canon_expected)); | |
| bb48024f | 407 | count_error(); |
| 56fd389c RK |
408 | } |
| 409 | ++tests; | |
| caecd4f4 RK |
410 | d = utf32_decompose_compat(&l, 1, 0); |
| 411 | compat_expected = utf32_to_utf8(d, utf32_len(d), 0); | |
| 56fd389c RK |
412 | if(strcmp(compat_folded, compat_expected)) { |
| 413 | fprintf(stderr, "%s:%d: compat-casefolding %#lx got '%s', expected '%s'\n", | |
| 460b9539 | 414 | __FILE__, __LINE__, (unsigned long)c, |
| 56fd389c | 415 | format(compat_folded), format(compat_expected)); |
| bb48024f | 416 | count_error(); |
| 460b9539 | 417 | } |
| 418 | ++tests; | |
| 419 | } | |
| 420 | } | |
| c85b7022 | 421 | check_string(utf8_casefold_canon("", 0, 0), ""); |
| 460b9539 | 422 | } |
| 423 | ||
| 8818b7fc RK |
424 | struct { |
| 425 | const char *in; | |
| 426 | const char *expect[10]; | |
| 427 | } wtest[] = { | |
| 428 | /* Empty string */ | |
| 429 | { "", { 0 } }, | |
| 430 | /* Only whitespace and punctuation */ | |
| 431 | { " ", { 0 } }, | |
| 432 | { " ' ", { 0 } }, | |
| 433 | { " ! ", { 0 } }, | |
| 434 | { " \"\" ", { 0 } }, | |
| 435 | { " @ ", { 0 } }, | |
| 436 | /* Basics */ | |
| 437 | { "wibble", { "wibble", 0 } }, | |
| 438 | { " wibble", { "wibble", 0 } }, | |
| 439 | { " wibble ", { "wibble", 0 } }, | |
| 440 | { "wibble ", { "wibble", 0 } }, | |
| 441 | { "wibble spong", { "wibble", "spong", 0 } }, | |
| 442 | { " wibble spong", { "wibble", "spong", 0 } }, | |
| 443 | { " wibble spong ", { "wibble", "spong", 0 } }, | |
| 444 | { "wibble spong ", { "wibble", "spong", 0 } }, | |
| 445 | { "wibble spong splat foo zot ", { "wibble", "spong", "splat", "foo", "zot", 0 } }, | |
| 446 | /* Apostrophes */ | |
| 447 | { "wibble 'spong", { "wibble", "spong", 0 } }, | |
| 448 | { " wibble's", { "wibble's", 0 } }, | |
| 449 | { " wibblespong' ", { "wibblespong", 0 } }, | |
| 450 | { "wibble sp''ong ", { "wibble", "sp", "ong", 0 } }, | |
| 451 | }; | |
| 452 | #define NWTEST (sizeof wtest / sizeof *wtest) | |
| 453 | ||
| 454 | static void test_words(void) { | |
| 455 | size_t t, nexpect, ngot, i; | |
| 456 | int right; | |
| 457 | ||
| 458 | fprintf(stderr, "test_words\n"); | |
| 459 | for(t = 0; t < NWTEST; ++t) { | |
| c85b7022 | 460 | char **got = utf8_word_split(wtest[t].in, strlen(wtest[t].in), &ngot, 0); |
| 8818b7fc RK |
461 | |
| 462 | for(nexpect = 0; wtest[t].expect[nexpect]; ++nexpect) | |
| 463 | ; | |
| 464 | if(nexpect == ngot) { | |
| 465 | for(i = 0; i < ngot; ++i) | |
| 466 | if(strcmp(wtest[t].expect[i], got[i])) | |
| 467 | break; | |
| 468 | right = i == ngot; | |
| 469 | } else | |
| 470 | right = 0; | |
| 471 | if(!right) { | |
| 472 | fprintf(stderr, "word split %zu failed\n", t); | |
| 473 | fprintf(stderr, "input: %s\n", wtest[t].in); | |
| 474 | fprintf(stderr, " | %-30s | %-30s\n", | |
| 475 | "expected", "got"); | |
| 476 | for(i = 0; i < nexpect || i < ngot; ++i) { | |
| 477 | const char *e = i < nexpect ? wtest[t].expect[i] : "<none>"; | |
| 478 | const char *g = i < ngot ? got[i] : "<none>"; | |
| 479 | fprintf(stderr, " %2zu | %-30s | %-30s\n", i, e, g); | |
| 480 | } | |
| 481 | count_error(); | |
| 482 | } | |
| 483 | ++tests; | |
| 484 | } | |
| 485 | } | |
| 486 | ||
| 033fd4e3 RK |
487 | /** @brief Less-than comparison function for integer heap */ |
| 488 | static inline int int_lt(int a, int b) { return a < b; } | |
| 489 | ||
| dab22732 RK |
490 | /** @struct iheap |
| 491 | * @brief A heap with @c int elements */ | |
| 033fd4e3 | 492 | HEAP_TYPE(iheap, int, int_lt); |
| 8e3fe3d8 | 493 | HEAP_DEFINE(iheap, int, int_lt); |
| 033fd4e3 RK |
494 | |
| 495 | /** @brief Tests for @ref heap.h */ | |
| 496 | static void test_heap(void) { | |
| 497 | struct iheap h[1]; | |
| 498 | int n; | |
| 499 | int last = -1; | |
| 500 | ||
| 501 | fprintf(stderr, "test_heap\n"); | |
| 502 | ||
| 503 | iheap_init(h); | |
| 504 | for(n = 0; n < 1000; ++n) | |
| 505 | iheap_insert(h, random() % 100); | |
| 506 | for(n = 0; n < 1000; ++n) { | |
| 507 | const int latest = iheap_remove(h); | |
| 508 | if(last > latest) | |
| 509 | fprintf(stderr, "should have %d <= %d\n", last, latest); | |
| 510 | insist(last <= latest); | |
| 511 | last = latest; | |
| 512 | } | |
| 513 | putchar('\n'); | |
| 514 | } | |
| 515 | ||
| e2452add RK |
516 | /** @brief Open a Unicode test file */ |
| 517 | static FILE *open_unicode_test(const char *path) { | |
| 518 | const char *base; | |
| 519 | FILE *fp; | |
| 520 | char buffer[1024]; | |
| 521 | int w; | |
| 522 | ||
| 523 | if((base = strrchr(path, '/'))) | |
| 524 | ++base; | |
| 525 | else | |
| 526 | base = path; | |
| 527 | if(!(fp = fopen(base, "r"))) { | |
| 528 | snprintf(buffer, sizeof buffer, | |
| 529 | "wget http://www.unicode.org/Public/5.0.0/ucd/%s", path); | |
| 530 | if((w = system(buffer))) | |
| 531 | fatal(0, "%s: %s", buffer, wstat(w)); | |
| 532 | if(chmod(base, 0444) < 0) | |
| 533 | fatal(errno, "chmod %s", base); | |
| 534 | if(!(fp = fopen(base, "r"))) | |
| 535 | fatal(errno, "%s", base); | |
| 536 | } | |
| 537 | return fp; | |
| 538 | } | |
| 539 | ||
| 1625e11a | 540 | /** @brief Run breaking tests for utf32_grapheme_boundary() etc */ |
| bb48024f RK |
541 | static void breaktest(const char *path, |
| 542 | int (*breakfn)(const uint32_t *, size_t, size_t)) { | |
| 543 | FILE *fp = open_unicode_test(path); | |
| 544 | int lineno = 0; | |
| 545 | char *l, *lp; | |
| 546 | size_t bn, n; | |
| 547 | char break_allowed[1024]; | |
| 548 | uint32_t buffer[1024]; | |
| 549 | ||
| 550 | while(!inputline(path, fp, &l, '\n')) { | |
| 551 | ++lineno; | |
| 552 | if(l[0] == '#') continue; | |
| 553 | bn = 0; | |
| 554 | lp = l; | |
| 555 | while(*lp) { | |
| 556 | if(*lp == ' ' || *lp == '\t') { | |
| 557 | ++lp; | |
| 558 | continue; | |
| 559 | } | |
| 560 | if(*lp == '#') | |
| 561 | break; | |
| 562 | if((unsigned char)*lp == 0xC3 && (unsigned char)lp[1] == 0xB7) { | |
| 563 | /* 00F7 DIVISION SIGN */ | |
| 564 | break_allowed[bn] = 1; | |
| 565 | lp += 2; | |
| 566 | continue; | |
| 567 | } | |
| 568 | if((unsigned char)*lp == 0xC3 && (unsigned char)lp[1] == 0x97) { | |
| 569 | /* 00D7 MULTIPLICATION SIGN */ | |
| 570 | break_allowed[bn] = 0; | |
| 571 | lp += 2; | |
| 572 | continue; | |
| 573 | } | |
| 574 | if(isxdigit((unsigned char)*lp)) { | |
| 575 | buffer[bn++] = strtoul(lp, &lp, 16); | |
| 576 | continue; | |
| 577 | } | |
| 578 | fatal(0, "%s:%d: evil line: %s", path, lineno, l); | |
| 579 | } | |
| 580 | for(n = 0; n <= bn; ++n) { | |
| 581 | if(breakfn(buffer, bn, n) != break_allowed[n]) { | |
| 582 | fprintf(stderr, | |
| b21a155c RK |
583 | "%s:%d: offset %zu: mismatch\n" |
| 584 | "%s\n" | |
| 585 | "\n", | |
| 586 | path, lineno, n, l); | |
| bb48024f RK |
587 | count_error(); |
| 588 | } | |
| 589 | ++tests; | |
| 590 | } | |
| 591 | xfree(l); | |
| 592 | } | |
| 593 | fclose(fp); | |
| 594 | } | |
| 595 | ||
| e5a5a138 RK |
596 | /** @brief Tests for @ref lib/unicode.h */ |
| 597 | static void test_unicode(void) { | |
| 598 | FILE *fp; | |
| 599 | int lineno = 0; | |
| 600 | char *l, *lp; | |
| 601 | uint32_t buffer[1024]; | |
| 16506c9d | 602 | uint32_t *c[6], *NFD_c[6], *NFKD_c[6], *NFC_c[6], *NFKC_c[6]; /* 1-indexed */ |
| e5a5a138 RK |
603 | int cn, bn; |
| 604 | ||
| 605 | fprintf(stderr, "test_unicode\n"); | |
| e2452add | 606 | fp = open_unicode_test("NormalizationTest.txt"); |
| e5a5a138 RK |
607 | while(!inputline("NormalizationTest.txt", fp, &l, '\n')) { |
| 608 | ++lineno; | |
| 609 | if(*l == '#' || *l == '@') | |
| 610 | continue; | |
| 611 | bn = 0; | |
| 612 | cn = 1; | |
| 613 | lp = l; | |
| 614 | c[cn++] = &buffer[bn]; | |
| 615 | while(*lp && *lp != '#') { | |
| 616 | if(*lp == ' ') { | |
| 617 | ++lp; | |
| 618 | continue; | |
| 619 | } | |
| 620 | if(*lp == ';') { | |
| 621 | buffer[bn++] = 0; | |
| 622 | if(cn == 6) | |
| 623 | break; | |
| 624 | c[cn++] = &buffer[bn]; | |
| 625 | ++lp; | |
| 626 | continue; | |
| 627 | } | |
| 628 | buffer[bn++] = strtoul(lp, &lp, 16); | |
| 629 | } | |
| 630 | buffer[bn] = 0; | |
| 631 | assert(cn == 6); | |
| 632 | for(cn = 1; cn <= 5; ++cn) { | |
| 633 | NFD_c[cn] = utf32_decompose_canon(c[cn], utf32_len(c[cn]), 0); | |
| 634 | NFKD_c[cn] = utf32_decompose_compat(c[cn], utf32_len(c[cn]), 0); | |
| 16506c9d RK |
635 | NFC_c[cn] = utf32_compose_canon(c[cn], utf32_len(c[cn]), 0); |
| 636 | NFKC_c[cn] = utf32_compose_compat(c[cn], utf32_len(c[cn]), 0); | |
| e5a5a138 RK |
637 | } |
| 638 | #define unt_check(T, A, B) do { \ | |
| 639 | ++tests; \ | |
| 640 | if(utf32_cmp(c[A], T##_c[B])) { \ | |
| e2452add RK |
641 | fprintf(stderr, \ |
| 642 | "NormalizationTest.txt:%d: c%d != "#T"(c%d)\n", \ | |
| 643 | lineno, A, B); \ | |
| 16506c9d | 644 | fprintf(stderr, " c%d:%s\n", \ |
| e5a5a138 | 645 | A, format_utf32(c[A])); \ |
| 16506c9d RK |
646 | fprintf(stderr, " c%d:%s\n", \ |
| 647 | B, format_utf32(c[B])); \ | |
| 648 | fprintf(stderr, "%4s(c%d):%s\n", \ | |
| e5a5a138 | 649 | #T, B, format_utf32(T##_c[B])); \ |
| bcf9ed7f | 650 | count_error(); \ |
| e5a5a138 RK |
651 | } \ |
| 652 | } while(0) | |
| 653 | unt_check(NFD, 3, 1); | |
| 654 | unt_check(NFD, 3, 2); | |
| 655 | unt_check(NFD, 3, 3); | |
| 656 | unt_check(NFD, 5, 4); | |
| 657 | unt_check(NFD, 5, 5); | |
| 658 | unt_check(NFKD, 5, 1); | |
| 659 | unt_check(NFKD, 5, 2); | |
| 660 | unt_check(NFKD, 5, 3); | |
| 661 | unt_check(NFKD, 5, 4); | |
| 662 | unt_check(NFKD, 5, 5); | |
| 16506c9d RK |
663 | unt_check(NFC, 2, 1); |
| 664 | unt_check(NFC, 2, 2); | |
| 665 | unt_check(NFC, 2, 3); | |
| 666 | unt_check(NFC, 4, 4); | |
| 667 | unt_check(NFC, 4, 5); | |
| 668 | unt_check(NFKC, 4, 1); | |
| 669 | unt_check(NFKC, 4, 2); | |
| 670 | unt_check(NFKC, 4, 3); | |
| 671 | unt_check(NFKC, 4, 4); | |
| 672 | unt_check(NFKC, 4, 5); | |
| e5a5a138 RK |
673 | for(cn = 1; cn <= 5; ++cn) { |
| 674 | xfree(NFD_c[cn]); | |
| 675 | xfree(NFKD_c[cn]); | |
| 676 | } | |
| 677 | xfree(l); | |
| 678 | } | |
| e2452add | 679 | fclose(fp); |
| 1625e11a | 680 | breaktest("auxiliary/GraphemeBreakTest.txt", utf32_is_grapheme_boundary); |
| bb48024f | 681 | breaktest("auxiliary/WordBreakTest.txt", utf32_is_word_boundary); |
| c35e83d9 RK |
682 | insist(utf32_combining_class(0x40000) == 0); |
| 683 | insist(utf32_combining_class(0xE0000) == 0); | |
| e5a5a138 RK |
684 | } |
| 685 | ||
| ea387d53 RK |
686 | static void test_signame(void) { |
| 687 | fprintf(stderr, "test_signame\n"); | |
| 688 | insist(find_signal("SIGTERM") == SIGTERM); | |
| 689 | insist(find_signal("SIGHUP") == SIGHUP); | |
| 690 | insist(find_signal("SIGINT") == SIGINT); | |
| 691 | insist(find_signal("SIGQUIT") == SIGQUIT); | |
| 692 | insist(find_signal("SIGKILL") == SIGKILL); | |
| 693 | insist(find_signal("SIGYOURMUM") == -1); | |
| 694 | } | |
| 695 | ||
| 9f28e855 RK |
696 | static void test_cache(void) { |
| 697 | const struct cache_type t1 = { 1 }, t2 = { 10 }; | |
| 698 | const char v11[] = "spong", v12[] = "wibble", v2[] = "blat"; | |
| 699 | fprintf(stderr, "test_cache\n"); | |
| 700 | cache_put(&t1, "1_1", v11); | |
| 701 | cache_put(&t1, "1_2", v12); | |
| 702 | cache_put(&t2, "2", v2); | |
| 703 | insist(cache_count() == 3); | |
| 704 | insist(cache_get(&t2, "2") == v2); | |
| 705 | insist(cache_get(&t1, "1_1") == v11); | |
| 706 | insist(cache_get(&t1, "1_2") == v12); | |
| 707 | insist(cache_get(&t1, "2") == 0); | |
| 708 | insist(cache_get(&t2, "1_1") == 0); | |
| 709 | insist(cache_get(&t2, "1_2") == 0); | |
| 710 | insist(cache_get(&t1, "2") == 0); | |
| 711 | insist(cache_get(&t2, "1_1") == 0); | |
| 712 | insist(cache_get(&t2, "1_2") == 0); | |
| 713 | sleep(2); | |
| 714 | cache_expire(); | |
| 715 | insist(cache_count() == 1); | |
| 716 | insist(cache_get(&t1, "1_1") == 0); | |
| 717 | insist(cache_get(&t1, "1_2") == 0); | |
| 718 | insist(cache_get(&t2, "2") == v2); | |
| 719 | cache_clean(0); | |
| 720 | insist(cache_count() == 0); | |
| 721 | insist(cache_get(&t2, "2") == 0); | |
| 722 | } | |
| 723 | ||
| 00e36cd0 RK |
724 | static void test_filepart(void) { |
| 725 | fprintf(stderr, "test_filepart\n"); | |
| 726 | check_string(d_dirname("/"), "/"); | |
| 727 | check_string(d_dirname("/spong"), "/"); | |
| 728 | check_string(d_dirname("/foo/bar"), "/foo"); | |
| 729 | check_string(d_dirname("./bar"), "."); | |
| 730 | check_string(d_dirname("."), "."); | |
| 731 | check_string(d_dirname(".."), "."); | |
| 732 | check_string(d_dirname("../blat"), ".."); | |
| 733 | check_string(d_dirname("wibble"), "."); | |
| 734 | check_string(extension("foo.c"), ".c"); | |
| 735 | check_string(extension(".c"), ".c"); | |
| 736 | check_string(extension("."), "."); | |
| 737 | check_string(extension("foo"), ""); | |
| 738 | check_string(extension("./foo"), ""); | |
| 739 | check_string(extension("./foo.c"), ".c"); | |
| 740 | } | |
| 741 | ||
| 65bb0fff RK |
742 | static void test_selection(void) { |
| 743 | hash *h; | |
| 744 | fprintf(stderr, "test_selection\n"); | |
| 745 | insist((h = selection_new()) != 0); | |
| 746 | selection_set(h, "one", 1); | |
| 747 | selection_set(h, "two", 1); | |
| 748 | selection_set(h, "three", 0); | |
| 749 | selection_set(h, "four", 1); | |
| 750 | insist(selection_selected(h, "one") == 1); | |
| 751 | insist(selection_selected(h, "two") == 1); | |
| 752 | insist(selection_selected(h, "three") == 0); | |
| 753 | insist(selection_selected(h, "four") == 1); | |
| 754 | insist(selection_selected(h, "five") == 0); | |
| 755 | insist(hash_count(h) == 3); | |
| 756 | selection_flip(h, "one"); | |
| 757 | selection_flip(h, "three"); | |
| 758 | insist(selection_selected(h, "one") == 0); | |
| 759 | insist(selection_selected(h, "three") == 1); | |
| 760 | insist(hash_count(h) == 3); | |
| 761 | selection_live(h, "one"); | |
| 762 | selection_live(h, "two"); | |
| 763 | selection_live(h, "three"); | |
| 764 | selection_cleanup(h); | |
| 765 | insist(selection_selected(h, "one") == 0); | |
| 766 | insist(selection_selected(h, "two") == 1); | |
| 767 | insist(selection_selected(h, "three") == 1); | |
| 768 | insist(selection_selected(h, "four") == 0); | |
| 769 | insist(selection_selected(h, "five") == 0); | |
| 770 | insist(hash_count(h) == 2); | |
| 771 | selection_empty(h); | |
| 772 | insist(selection_selected(h, "one") == 0); | |
| 773 | insist(selection_selected(h, "two") == 0); | |
| 774 | insist(selection_selected(h, "three") == 0); | |
| 775 | insist(selection_selected(h, "four") == 0); | |
| 776 | insist(selection_selected(h, "five") == 0); | |
| 777 | insist(hash_count(h) == 0); | |
| 778 | } | |
| 779 | ||
| 460b9539 | 780 | int main(void) { |
| bb48024f | 781 | fail_first = !!getenv("FAIL_FIRST"); |
| 460b9539 | 782 | insist('\n' == 0x0A); |
| 783 | insist('\r' == 0x0D); | |
| 784 | insist(' ' == 0x20); | |
| 785 | insist('0' == 0x30); | |
| 786 | insist('9' == 0x39); | |
| 787 | insist('A' == 0x41); | |
| 788 | insist('Z' == 0x5A); | |
| 789 | insist('a' == 0x61); | |
| 790 | insist('z' == 0x7A); | |
| 791 | /* addr.c */ | |
| 792 | /* asprintf.c */ | |
| 793 | /* authhash.c */ | |
| 794 | /* basen.c */ | |
| 795 | /* charset.c */ | |
| 796 | /* client.c */ | |
| 797 | /* configuration.c */ | |
| 798 | /* event.c */ | |
| 00e36cd0 RK |
799 | /* filepart.c */ |
| 800 | test_filepart(); | |
| 460b9539 | 801 | /* fprintf.c */ |
| 033fd4e3 RK |
802 | /* heap.c */ |
| 803 | test_heap(); | |
| 460b9539 | 804 | /* hex.c */ |
| 805 | test_hex(); | |
| 806 | /* inputline.c */ | |
| 807 | /* kvp.c */ | |
| 808 | /* log.c */ | |
| 809 | /* mem.c */ | |
| 810 | /* mime.c */ | |
| 811 | test_mime(); | |
| 812 | /* mixer.c */ | |
| 813 | /* plugin.c */ | |
| 814 | /* printf.c */ | |
| 815 | /* queue.c */ | |
| 816 | /* sink.c */ | |
| 817 | /* snprintf.c */ | |
| 818 | /* split.c */ | |
| 819 | /* syscalls.c */ | |
| 820 | /* table.c */ | |
| e5a5a138 RK |
821 | /* unicode.c */ |
| 822 | test_unicode(); | |
| 460b9539 | 823 | /* utf8.c */ |
| 824 | test_utf8(); | |
| 825 | /* vector.c */ | |
| 826 | /* words.c */ | |
| 827 | test_casefold(); | |
| 8818b7fc | 828 | test_words(); |
| 460b9539 | 829 | /* wstat.c */ |
| ea387d53 RK |
830 | /* signame.c */ |
| 831 | test_signame(); | |
| 9f28e855 RK |
832 | /* cache.c */ |
| 833 | test_cache(); | |
| 65bb0fff RK |
834 | /* selection.c */ |
| 835 | test_selection(); | |
| 460b9539 | 836 | fprintf(stderr, "%d errors out of %d tests\n", errors, tests); |
| 837 | return !!errors; | |
| 838 | } | |
| 839 | ||
| 840 | /* | |
| 841 | Local Variables: | |
| 842 | c-basic-offset:2 | |
| 843 | comment-column:40 | |
| 56fd389c RK |
844 | fill-column:79 |
| 845 | indent-tabs-mode:nil | |
| 460b9539 | 846 | End: |
| 847 | */ |