1
by brian
clean slate |
1 |
/* Copyright (C) 2002 MySQL AB
|
2 |
||
3 |
This program is free software; you can redistribute it and/or modify
|
|
4 |
it under the terms of the GNU General Public License as published by
|
|
5 |
the Free Software Foundation; version 2 of the License.
|
|
6 |
||
7 |
This program is distributed in the hope that it will be useful,
|
|
8 |
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
9 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
10 |
GNU General Public License for more details.
|
|
11 |
||
12 |
You should have received a copy of the GNU General Public License
|
|
13 |
along with this program; if not, write to the Free Software
|
|
14 |
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
|
|
15 |
||
16 |
#include <my_global.h> |
|
17 |
#include "m_string.h" |
|
18 |
#include "m_ctype.h" |
|
19 |
#include "my_sys.h" /* Needed for MY_ERRNO_ERANGE */ |
|
20 |
#include <errno.h> |
|
21 |
||
22 |
#include "stdarg.h" |
|
23 |
||
24 |
/*
|
|
25 |
Returns the number of bytes required for strnxfrm().
|
|
26 |
*/
|
|
27 |
||
28 |
size_t my_strnxfrmlen_simple(CHARSET_INFO *cs, size_t len) |
|
29 |
{
|
|
30 |
return len * (cs->strxfrm_multiply ? cs->strxfrm_multiply : 1); |
|
31 |
}
|
|
32 |
||
33 |
||
34 |
/*
|
|
35 |
Converts a string into its sort key.
|
|
36 |
|
|
37 |
SYNOPSIS
|
|
38 |
my_strnxfrm_xxx()
|
|
39 |
|
|
40 |
IMPLEMENTATION
|
|
41 |
|
|
42 |
The my_strxfrm_xxx() function transforms a string pointed to by
|
|
43 |
'src' with length 'srclen' according to the charset+collation
|
|
44 |
pair 'cs' and copies the result key into 'dest'.
|
|
45 |
|
|
46 |
Comparing two strings using memcmp() after my_strnxfrm_xxx()
|
|
47 |
is equal to comparing two original strings with my_strnncollsp_xxx().
|
|
48 |
|
|
49 |
Not more than 'dstlen' bytes are written into 'dst'.
|
|
50 |
To garantee that the whole string is transformed, 'dstlen' must be
|
|
51 |
at least srclen*cs->strnxfrm_multiply bytes long. Otherwise,
|
|
52 |
consequent memcmp() may return a non-accurate result.
|
|
53 |
|
|
54 |
If the source string is too short to fill whole 'dstlen' bytes,
|
|
55 |
then the 'dest' string is padded up to 'dstlen', ensuring that:
|
|
56 |
|
|
57 |
"a" == "a "
|
|
58 |
"a\0" < "a"
|
|
59 |
"a\0" < "a "
|
|
60 |
|
|
61 |
my_strnxfrm_simple() is implemented for 8bit charsets and
|
|
62 |
simple collations with one-to-one string->key transformation.
|
|
63 |
|
|
64 |
See also implementations for various charsets/collations in
|
|
65 |
other ctype-xxx.c files.
|
|
66 |
|
|
67 |
RETURN
|
|
68 |
|
|
69 |
Target len 'dstlen'.
|
|
70 |
|
|
71 |
*/
|
|
72 |
||
73 |
||
74 |
size_t my_strnxfrm_simple(CHARSET_INFO * cs, |
|
75 |
uchar *dst, size_t dstlen, uint nweights, |
|
76 |
const uchar *src, size_t srclen, uint flags) |
|
77 |
{
|
|
78 |
uchar *map= cs->sort_order; |
|
79 |
uchar *d0= dst; |
|
80 |
uint frmlen; |
|
81 |
if ((frmlen= min(dstlen, nweights)) > srclen) |
|
82 |
frmlen= srclen; |
|
83 |
if (dst != src) |
|
84 |
{
|
|
85 |
const uchar *end; |
|
86 |
for (end= src + frmlen; src < end;) |
|
87 |
*dst++= map[*src++]; |
|
88 |
}
|
|
89 |
else
|
|
90 |
{
|
|
91 |
const uchar *end; |
|
92 |
for (end= dst + frmlen; dst < end; dst++) |
|
93 |
*dst= map[(uchar) *dst]; |
|
94 |
}
|
|
95 |
return my_strxfrm_pad_desc_and_reverse(cs, d0, dst, d0 + dstlen, |
|
96 |
nweights - frmlen, flags, 0); |
|
97 |
}
|
|
98 |
||
99 |
||
100 |
int my_strnncoll_simple(CHARSET_INFO * cs, const uchar *s, size_t slen, |
|
101 |
const uchar *t, size_t tlen, |
|
102 |
my_bool t_is_prefix) |
|
103 |
{
|
|
104 |
size_t len = ( slen > tlen ) ? tlen : slen; |
|
105 |
uchar *map= cs->sort_order; |
|
106 |
if (t_is_prefix && slen > tlen) |
|
107 |
slen=tlen; |
|
108 |
while (len--) |
|
109 |
{
|
|
110 |
if (map[*s++] != map[*t++]) |
|
111 |
return ((int) map[s[-1]] - (int) map[t[-1]]); |
|
112 |
}
|
|
113 |
/*
|
|
114 |
We can't use (slen - tlen) here as the result may be outside of the
|
|
115 |
precision of a signed int
|
|
116 |
*/
|
|
117 |
return slen > tlen ? 1 : slen < tlen ? -1 : 0 ; |
|
118 |
}
|
|
119 |
||
120 |
||
121 |
/*
|
|
122 |
Compare strings, discarding end space
|
|
123 |
||
124 |
SYNOPSIS
|
|
125 |
my_strnncollsp_simple()
|
|
126 |
cs character set handler
|
|
127 |
a First string to compare
|
|
128 |
a_length Length of 'a'
|
|
129 |
b Second string to compare
|
|
130 |
b_length Length of 'b'
|
|
131 |
diff_if_only_endspace_difference
|
|
132 |
Set to 1 if the strings should be regarded as different
|
|
133 |
if they only difference in end space
|
|
134 |
||
135 |
IMPLEMENTATION
|
|
136 |
If one string is shorter as the other, then we space extend the other
|
|
137 |
so that the strings have equal length.
|
|
138 |
||
139 |
This will ensure that the following things hold:
|
|
140 |
||
141 |
"a" == "a "
|
|
142 |
"a\0" < "a"
|
|
143 |
"a\0" < "a "
|
|
144 |
||
145 |
RETURN
|
|
146 |
< 0 a < b
|
|
147 |
= 0 a == b
|
|
148 |
> 0 a > b
|
|
149 |
*/
|
|
150 |
||
151 |
int my_strnncollsp_simple(CHARSET_INFO * cs, const uchar *a, size_t a_length, |
|
152 |
const uchar *b, size_t b_length, |
|
153 |
my_bool diff_if_only_endspace_difference) |
|
154 |
{
|
|
155 |
const uchar *map= cs->sort_order, *end; |
|
156 |
size_t length; |
|
157 |
int res; |
|
158 |
||
159 |
#ifndef VARCHAR_WITH_DIFF_ENDSPACE_ARE_DIFFERENT_FOR_UNIQUE
|
|
160 |
diff_if_only_endspace_difference= 0; |
|
161 |
#endif
|
|
162 |
||
163 |
end= a + (length= min(a_length, b_length)); |
|
164 |
while (a < end) |
|
165 |
{
|
|
166 |
if (map[*a++] != map[*b++]) |
|
167 |
return ((int) map[a[-1]] - (int) map[b[-1]]); |
|
168 |
}
|
|
169 |
res= 0; |
|
170 |
if (a_length != b_length) |
|
171 |
{
|
|
172 |
int swap= 1; |
|
173 |
if (diff_if_only_endspace_difference) |
|
174 |
res= 1; /* Assume 'a' is bigger */ |
|
175 |
/*
|
|
176 |
Check the next not space character of the longer key. If it's < ' ',
|
|
177 |
then it's smaller than the other key.
|
|
178 |
*/
|
|
179 |
if (a_length < b_length) |
|
180 |
{
|
|
181 |
/* put shorter key in s */
|
|
182 |
a_length= b_length; |
|
183 |
a= b; |
|
184 |
swap= -1; /* swap sign of result */ |
|
185 |
res= -res; |
|
186 |
}
|
|
187 |
for (end= a + a_length-length; a < end ; a++) |
|
188 |
{
|
|
189 |
if (map[*a] != ' ') |
|
190 |
return (map[*a] < ' ') ? -swap : swap; |
|
191 |
}
|
|
192 |
}
|
|
193 |
return res; |
|
194 |
}
|
|
195 |
||
196 |
||
197 |
size_t my_caseup_str_8bit(CHARSET_INFO * cs,char *str) |
|
198 |
{
|
|
199 |
register uchar *map= cs->to_upper; |
|
200 |
char *str_orig= str; |
|
201 |
while ((*str= (char) map[(uchar) *str]) != 0) |
|
202 |
str++; |
|
203 |
return (size_t) (str - str_orig); |
|
204 |
}
|
|
205 |
||
206 |
||
207 |
size_t my_casedn_str_8bit(CHARSET_INFO * cs,char *str) |
|
208 |
{
|
|
209 |
register uchar *map= cs->to_lower; |
|
210 |
char *str_orig= str; |
|
211 |
while ((*str= (char) map[(uchar) *str]) != 0) |
|
212 |
str++; |
|
213 |
return (size_t) (str - str_orig); |
|
214 |
}
|
|
215 |
||
216 |
||
217 |
size_t my_caseup_8bit(CHARSET_INFO * cs, char *src, size_t srclen, |
|
218 |
char *dst __attribute__((unused)), |
|
219 |
size_t dstlen __attribute__((unused))) |
|
220 |
{
|
|
221 |
char *end= src + srclen; |
|
222 |
register uchar *map= cs->to_upper; |
|
223 |
DBUG_ASSERT(src == dst && srclen == dstlen); |
|
224 |
for ( ; src != end ; src++) |
|
225 |
*src= (char) map[(uchar) *src]; |
|
226 |
return srclen; |
|
227 |
}
|
|
228 |
||
229 |
||
230 |
size_t my_casedn_8bit(CHARSET_INFO * cs, char *src, size_t srclen, |
|
231 |
char *dst __attribute__((unused)), |
|
232 |
size_t dstlen __attribute__((unused))) |
|
233 |
{
|
|
234 |
char *end= src + srclen; |
|
235 |
register uchar *map=cs->to_lower; |
|
236 |
DBUG_ASSERT(src == dst && srclen == dstlen); |
|
237 |
for ( ; src != end ; src++) |
|
238 |
*src= (char) map[(uchar) *src]; |
|
239 |
return srclen; |
|
240 |
}
|
|
241 |
||
242 |
int my_strcasecmp_8bit(CHARSET_INFO * cs,const char *s, const char *t) |
|
243 |
{
|
|
244 |
register uchar *map=cs->to_upper; |
|
245 |
while (map[(uchar) *s] == map[(uchar) *t++]) |
|
246 |
if (!*s++) return 0; |
|
247 |
return ((int) map[(uchar) s[0]] - (int) map[(uchar) t[-1]]); |
|
248 |
}
|
|
249 |
||
250 |
||
251 |
int my_mb_wc_8bit(CHARSET_INFO *cs,my_wc_t *wc, |
|
252 |
const uchar *str, |
|
253 |
const uchar *end __attribute__((unused))) |
|
254 |
{
|
|
255 |
if (str >= end) |
|
256 |
return MY_CS_TOOSMALL; |
|
257 |
||
258 |
*wc=cs->tab_to_uni[*str]; |
|
259 |
return (!wc[0] && str[0]) ? -1 : 1; |
|
260 |
}
|
|
261 |
||
262 |
int my_wc_mb_8bit(CHARSET_INFO *cs,my_wc_t wc, |
|
263 |
uchar *str, |
|
264 |
uchar *end) |
|
265 |
{
|
|
266 |
MY_UNI_IDX *idx; |
|
267 |
||
268 |
if (str >= end) |
|
269 |
return MY_CS_TOOSMALL; |
|
270 |
||
271 |
for (idx=cs->tab_from_uni; idx->tab ; idx++) |
|
272 |
{
|
|
273 |
if (idx->from <= wc && idx->to >= wc) |
|
274 |
{
|
|
275 |
str[0]= idx->tab[wc - idx->from]; |
|
276 |
return (!str[0] && wc) ? MY_CS_ILUNI : 1; |
|
277 |
}
|
|
278 |
}
|
|
279 |
return MY_CS_ILUNI; |
|
280 |
}
|
|
281 |
||
282 |
||
283 |
/*
|
|
284 |
We can't use vsprintf here as it's not guaranteed to return
|
|
285 |
the length on all operating systems.
|
|
286 |
This function is also not called in a safe environment, so the
|
|
287 |
end buffer must be checked.
|
|
288 |
*/
|
|
289 |
||
290 |
size_t my_snprintf_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
291 |
char* to, size_t n __attribute__((unused)), |
|
292 |
const char* fmt, ...) |
|
293 |
{
|
|
294 |
va_list args; |
|
295 |
int result; |
|
296 |
va_start(args,fmt); |
|
77.1.18
by Monty Taylor
Removed my_vsnprintf and my_snprintf. |
297 |
result= vsnprintf(to, n, fmt, args); |
1
by brian
clean slate |
298 |
va_end(args); |
299 |
return result; |
|
300 |
}
|
|
301 |
||
302 |
||
303 |
void my_hash_sort_simple(CHARSET_INFO *cs, |
|
304 |
const uchar *key, size_t len, |
|
305 |
ulong *nr1, ulong *nr2) |
|
306 |
{
|
|
307 |
register uchar *sort_order=cs->sort_order; |
|
308 |
const uchar *end; |
|
309 |
||
310 |
/*
|
|
311 |
Remove end space. We have to do this to be able to compare
|
|
312 |
'A ' and 'A' as identical
|
|
313 |
*/
|
|
314 |
end= skip_trailing_space(key, len); |
|
315 |
||
316 |
for (; key < (uchar*) end ; key++) |
|
317 |
{
|
|
318 |
nr1[0]^=(ulong) ((((uint) nr1[0] & 63)+nr2[0]) * |
|
319 |
((uint) sort_order[(uint) *key])) + (nr1[0] << 8); |
|
320 |
nr2[0]+=3; |
|
321 |
}
|
|
322 |
}
|
|
323 |
||
324 |
||
325 |
long my_strntol_8bit(CHARSET_INFO *cs, |
|
326 |
const char *nptr, size_t l, int base, |
|
327 |
char **endptr, int *err) |
|
328 |
{
|
|
329 |
int negative; |
|
330 |
register uint32 cutoff; |
|
331 |
register uint cutlim; |
|
332 |
register uint32 i; |
|
333 |
register const char *s; |
|
334 |
register uchar c; |
|
335 |
const char *save, *e; |
|
336 |
int overflow; |
|
337 |
||
338 |
*err= 0; /* Initialize error indicator */ |
|
339 |
#ifdef NOT_USED
|
|
340 |
if (base < 0 || base == 1 || base > 36) |
|
341 |
base = 10; |
|
342 |
#endif
|
|
343 |
||
344 |
s = nptr; |
|
345 |
e = nptr+l; |
|
346 |
||
347 |
for ( ; s<e && my_isspace(cs, *s) ; s++); |
|
348 |
||
349 |
if (s == e) |
|
350 |
{
|
|
351 |
goto noconv; |
|
352 |
}
|
|
353 |
||
354 |
/* Check for a sign. */
|
|
355 |
if (*s == '-') |
|
356 |
{
|
|
357 |
negative = 1; |
|
358 |
++s; |
|
359 |
}
|
|
360 |
else if (*s == '+') |
|
361 |
{
|
|
362 |
negative = 0; |
|
363 |
++s; |
|
364 |
}
|
|
365 |
else
|
|
366 |
negative = 0; |
|
367 |
||
368 |
#ifdef NOT_USED
|
|
369 |
if (base == 16 && s[0] == '0' && (s[1]=='X' || s[1]=='x')) |
|
370 |
s += 2; |
|
371 |
#endif
|
|
372 |
||
373 |
#ifdef NOT_USED
|
|
374 |
if (base == 0) |
|
375 |
{
|
|
376 |
if (*s == '0') |
|
377 |
{
|
|
378 |
if (s[1]=='X' || s[1]=='x') |
|
379 |
{
|
|
380 |
s += 2; |
|
381 |
base = 16; |
|
382 |
}
|
|
383 |
else
|
|
384 |
base = 8; |
|
385 |
}
|
|
386 |
else
|
|
387 |
base = 10; |
|
388 |
}
|
|
389 |
#endif
|
|
390 |
||
391 |
save = s; |
|
392 |
cutoff = ((uint32)~0L) / (uint32) base; |
|
393 |
cutlim = (uint) (((uint32)~0L) % (uint32) base); |
|
394 |
||
395 |
overflow = 0; |
|
396 |
i = 0; |
|
397 |
for (c = *s; s != e; c = *++s) |
|
398 |
{
|
|
399 |
if (c>='0' && c<='9') |
|
400 |
c -= '0'; |
|
401 |
else if (c>='A' && c<='Z') |
|
402 |
c = c - 'A' + 10; |
|
403 |
else if (c>='a' && c<='z') |
|
404 |
c = c - 'a' + 10; |
|
405 |
else
|
|
406 |
break; |
|
407 |
if (c >= base) |
|
408 |
break; |
|
409 |
if (i > cutoff || (i == cutoff && c > cutlim)) |
|
410 |
overflow = 1; |
|
411 |
else
|
|
412 |
{
|
|
413 |
i *= (uint32) base; |
|
414 |
i += c; |
|
415 |
}
|
|
416 |
}
|
|
417 |
||
418 |
if (s == save) |
|
419 |
goto noconv; |
|
420 |
||
421 |
if (endptr != NULL) |
|
422 |
*endptr = (char *) s; |
|
423 |
||
424 |
if (negative) |
|
425 |
{
|
|
426 |
if (i > (uint32) INT_MIN32) |
|
427 |
overflow = 1; |
|
428 |
}
|
|
429 |
else if (i > INT_MAX32) |
|
430 |
overflow = 1; |
|
431 |
||
432 |
if (overflow) |
|
433 |
{
|
|
434 |
err[0]= ERANGE; |
|
435 |
return negative ? INT_MIN32 : INT_MAX32; |
|
436 |
}
|
|
437 |
||
438 |
return (negative ? -((long) i) : (long) i); |
|
439 |
||
440 |
noconv: |
|
441 |
err[0]= EDOM; |
|
442 |
if (endptr != NULL) |
|
443 |
*endptr = (char *) nptr; |
|
444 |
return 0L; |
|
445 |
}
|
|
446 |
||
447 |
||
448 |
ulong my_strntoul_8bit(CHARSET_INFO *cs, |
|
449 |
const char *nptr, size_t l, int base, |
|
450 |
char **endptr, int *err) |
|
451 |
{
|
|
452 |
int negative; |
|
453 |
register uint32 cutoff; |
|
454 |
register uint cutlim; |
|
455 |
register uint32 i; |
|
456 |
register const char *s; |
|
457 |
register uchar c; |
|
458 |
const char *save, *e; |
|
459 |
int overflow; |
|
460 |
||
461 |
*err= 0; /* Initialize error indicator */ |
|
462 |
#ifdef NOT_USED
|
|
463 |
if (base < 0 || base == 1 || base > 36) |
|
464 |
base = 10; |
|
465 |
#endif
|
|
466 |
||
467 |
s = nptr; |
|
468 |
e = nptr+l; |
|
469 |
||
470 |
for( ; s<e && my_isspace(cs, *s); s++); |
|
471 |
||
472 |
if (s==e) |
|
473 |
{
|
|
474 |
goto noconv; |
|
475 |
}
|
|
476 |
||
477 |
if (*s == '-') |
|
478 |
{
|
|
479 |
negative = 1; |
|
480 |
++s; |
|
481 |
}
|
|
482 |
else if (*s == '+') |
|
483 |
{
|
|
484 |
negative = 0; |
|
485 |
++s; |
|
486 |
}
|
|
487 |
else
|
|
488 |
negative = 0; |
|
489 |
||
490 |
#ifdef NOT_USED
|
|
491 |
if (base == 16 && s[0] == '0' && (s[1]=='X' || s[1]=='x')) |
|
492 |
s += 2; |
|
493 |
#endif
|
|
494 |
||
495 |
#ifdef NOT_USED
|
|
496 |
if (base == 0) |
|
497 |
{
|
|
498 |
if (*s == '0') |
|
499 |
{
|
|
500 |
if (s[1]=='X' || s[1]=='x') |
|
501 |
{
|
|
502 |
s += 2; |
|
503 |
base = 16; |
|
504 |
}
|
|
505 |
else
|
|
506 |
base = 8; |
|
507 |
}
|
|
508 |
else
|
|
509 |
base = 10; |
|
510 |
}
|
|
511 |
#endif
|
|
512 |
||
513 |
save = s; |
|
514 |
cutoff = ((uint32)~0L) / (uint32) base; |
|
515 |
cutlim = (uint) (((uint32)~0L) % (uint32) base); |
|
516 |
overflow = 0; |
|
517 |
i = 0; |
|
518 |
||
519 |
for (c = *s; s != e; c = *++s) |
|
520 |
{
|
|
521 |
if (c>='0' && c<='9') |
|
522 |
c -= '0'; |
|
523 |
else if (c>='A' && c<='Z') |
|
524 |
c = c - 'A' + 10; |
|
525 |
else if (c>='a' && c<='z') |
|
526 |
c = c - 'a' + 10; |
|
527 |
else
|
|
528 |
break; |
|
529 |
if (c >= base) |
|
530 |
break; |
|
531 |
if (i > cutoff || (i == cutoff && c > cutlim)) |
|
532 |
overflow = 1; |
|
533 |
else
|
|
534 |
{
|
|
535 |
i *= (uint32) base; |
|
536 |
i += c; |
|
537 |
}
|
|
538 |
}
|
|
539 |
||
540 |
if (s == save) |
|
541 |
goto noconv; |
|
542 |
||
543 |
if (endptr != NULL) |
|
544 |
*endptr = (char *) s; |
|
545 |
||
546 |
if (overflow) |
|
547 |
{
|
|
548 |
err[0]= ERANGE; |
|
549 |
return (~(uint32) 0); |
|
550 |
}
|
|
551 |
||
552 |
return (negative ? -((long) i) : (long) i); |
|
553 |
||
554 |
noconv: |
|
555 |
err[0]= EDOM; |
|
556 |
if (endptr != NULL) |
|
557 |
*endptr = (char *) nptr; |
|
558 |
return 0L; |
|
559 |
}
|
|
560 |
||
561 |
||
562 |
longlong my_strntoll_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
563 |
const char *nptr, size_t l, int base, |
|
564 |
char **endptr,int *err) |
|
565 |
{
|
|
566 |
int negative; |
|
567 |
register ulonglong cutoff; |
|
568 |
register uint cutlim; |
|
569 |
register ulonglong i; |
|
570 |
register const char *s, *e; |
|
571 |
const char *save; |
|
572 |
int overflow; |
|
573 |
||
574 |
*err= 0; /* Initialize error indicator */ |
|
575 |
#ifdef NOT_USED
|
|
576 |
if (base < 0 || base == 1 || base > 36) |
|
577 |
base = 10; |
|
578 |
#endif
|
|
579 |
||
580 |
s = nptr; |
|
581 |
e = nptr+l; |
|
582 |
||
583 |
for(; s<e && my_isspace(cs,*s); s++); |
|
584 |
||
585 |
if (s == e) |
|
586 |
{
|
|
587 |
goto noconv; |
|
588 |
}
|
|
589 |
||
590 |
if (*s == '-') |
|
591 |
{
|
|
592 |
negative = 1; |
|
593 |
++s; |
|
594 |
}
|
|
595 |
else if (*s == '+') |
|
596 |
{
|
|
597 |
negative = 0; |
|
598 |
++s; |
|
599 |
}
|
|
600 |
else
|
|
601 |
negative = 0; |
|
602 |
||
603 |
#ifdef NOT_USED
|
|
604 |
if (base == 16 && s[0] == '0' && (s[1]=='X'|| s[1]=='x')) |
|
605 |
s += 2; |
|
606 |
#endif
|
|
607 |
||
608 |
#ifdef NOT_USED
|
|
609 |
if (base == 0) |
|
610 |
{
|
|
611 |
if (*s == '0') |
|
612 |
{
|
|
613 |
if (s[1]=='X' || s[1]=='x') |
|
614 |
{
|
|
615 |
s += 2; |
|
616 |
base = 16; |
|
617 |
}
|
|
618 |
else
|
|
619 |
base = 8; |
|
620 |
}
|
|
621 |
else
|
|
622 |
base = 10; |
|
623 |
}
|
|
624 |
#endif
|
|
625 |
||
626 |
save = s; |
|
627 |
||
628 |
cutoff = (~(ulonglong) 0) / (unsigned long int) base; |
|
629 |
cutlim = (uint) ((~(ulonglong) 0) % (unsigned long int) base); |
|
630 |
||
631 |
overflow = 0; |
|
632 |
i = 0; |
|
633 |
for ( ; s != e; s++) |
|
634 |
{
|
|
635 |
register uchar c= *s; |
|
636 |
if (c>='0' && c<='9') |
|
637 |
c -= '0'; |
|
638 |
else if (c>='A' && c<='Z') |
|
639 |
c = c - 'A' + 10; |
|
640 |
else if (c>='a' && c<='z') |
|
641 |
c = c - 'a' + 10; |
|
642 |
else
|
|
643 |
break; |
|
644 |
if (c >= base) |
|
645 |
break; |
|
646 |
if (i > cutoff || (i == cutoff && c > cutlim)) |
|
647 |
overflow = 1; |
|
648 |
else
|
|
649 |
{
|
|
650 |
i *= (ulonglong) base; |
|
651 |
i += c; |
|
652 |
}
|
|
653 |
}
|
|
654 |
||
655 |
if (s == save) |
|
656 |
goto noconv; |
|
657 |
||
658 |
if (endptr != NULL) |
|
659 |
*endptr = (char *) s; |
|
660 |
||
661 |
if (negative) |
|
662 |
{
|
|
663 |
if (i > (ulonglong) LONGLONG_MIN) |
|
664 |
overflow = 1; |
|
665 |
}
|
|
666 |
else if (i > (ulonglong) LONGLONG_MAX) |
|
667 |
overflow = 1; |
|
668 |
||
669 |
if (overflow) |
|
670 |
{
|
|
671 |
err[0]= ERANGE; |
|
672 |
return negative ? LONGLONG_MIN : LONGLONG_MAX; |
|
673 |
}
|
|
674 |
||
675 |
return (negative ? -((longlong) i) : (longlong) i); |
|
676 |
||
677 |
noconv: |
|
678 |
err[0]= EDOM; |
|
679 |
if (endptr != NULL) |
|
680 |
*endptr = (char *) nptr; |
|
681 |
return 0L; |
|
682 |
}
|
|
683 |
||
684 |
||
685 |
ulonglong my_strntoull_8bit(CHARSET_INFO *cs, |
|
686 |
const char *nptr, size_t l, int base, |
|
687 |
char **endptr, int *err) |
|
688 |
{
|
|
689 |
int negative; |
|
690 |
register ulonglong cutoff; |
|
691 |
register uint cutlim; |
|
692 |
register ulonglong i; |
|
693 |
register const char *s, *e; |
|
694 |
const char *save; |
|
695 |
int overflow; |
|
696 |
||
697 |
*err= 0; /* Initialize error indicator */ |
|
698 |
#ifdef NOT_USED
|
|
699 |
if (base < 0 || base == 1 || base > 36) |
|
700 |
base = 10; |
|
701 |
#endif
|
|
702 |
||
703 |
s = nptr; |
|
704 |
e = nptr+l; |
|
705 |
||
706 |
for(; s<e && my_isspace(cs,*s); s++); |
|
707 |
||
708 |
if (s == e) |
|
709 |
{
|
|
710 |
goto noconv; |
|
711 |
}
|
|
712 |
||
713 |
if (*s == '-') |
|
714 |
{
|
|
715 |
negative = 1; |
|
716 |
++s; |
|
717 |
}
|
|
718 |
else if (*s == '+') |
|
719 |
{
|
|
720 |
negative = 0; |
|
721 |
++s; |
|
722 |
}
|
|
723 |
else
|
|
724 |
negative = 0; |
|
725 |
||
726 |
#ifdef NOT_USED
|
|
727 |
if (base == 16 && s[0] == '0' && (s[1]=='X' || s[1]=='x')) |
|
728 |
s += 2; |
|
729 |
#endif
|
|
730 |
||
731 |
#ifdef NOT_USED
|
|
732 |
if (base == 0) |
|
733 |
{
|
|
734 |
if (*s == '0') |
|
735 |
{
|
|
736 |
if (s[1]=='X' || s[1]=='x') |
|
737 |
{
|
|
738 |
s += 2; |
|
739 |
base = 16; |
|
740 |
}
|
|
741 |
else
|
|
742 |
base = 8; |
|
743 |
}
|
|
744 |
else
|
|
745 |
base = 10; |
|
746 |
}
|
|
747 |
#endif
|
|
748 |
||
749 |
save = s; |
|
750 |
||
751 |
cutoff = (~(ulonglong) 0) / (unsigned long int) base; |
|
752 |
cutlim = (uint) ((~(ulonglong) 0) % (unsigned long int) base); |
|
753 |
||
754 |
overflow = 0; |
|
755 |
i = 0; |
|
756 |
for ( ; s != e; s++) |
|
757 |
{
|
|
758 |
register uchar c= *s; |
|
759 |
||
760 |
if (c>='0' && c<='9') |
|
761 |
c -= '0'; |
|
762 |
else if (c>='A' && c<='Z') |
|
763 |
c = c - 'A' + 10; |
|
764 |
else if (c>='a' && c<='z') |
|
765 |
c = c - 'a' + 10; |
|
766 |
else
|
|
767 |
break; |
|
768 |
if (c >= base) |
|
769 |
break; |
|
770 |
if (i > cutoff || (i == cutoff && c > cutlim)) |
|
771 |
overflow = 1; |
|
772 |
else
|
|
773 |
{
|
|
774 |
i *= (ulonglong) base; |
|
775 |
i += c; |
|
776 |
}
|
|
777 |
}
|
|
778 |
||
779 |
if (s == save) |
|
780 |
goto noconv; |
|
781 |
||
782 |
if (endptr != NULL) |
|
783 |
*endptr = (char *) s; |
|
784 |
||
785 |
if (overflow) |
|
786 |
{
|
|
787 |
err[0]= ERANGE; |
|
788 |
return (~(ulonglong) 0); |
|
789 |
}
|
|
790 |
||
791 |
return (negative ? -((longlong) i) : (longlong) i); |
|
792 |
||
793 |
noconv: |
|
794 |
err[0]= EDOM; |
|
795 |
if (endptr != NULL) |
|
796 |
*endptr = (char *) nptr; |
|
797 |
return 0L; |
|
798 |
}
|
|
799 |
||
800 |
||
801 |
/*
|
|
802 |
Read double from string
|
|
803 |
||
804 |
SYNOPSIS:
|
|
805 |
my_strntod_8bit()
|
|
806 |
cs Character set information
|
|
807 |
str String to convert to double
|
|
808 |
length Optional length for string.
|
|
809 |
end result pointer to end of converted string
|
|
810 |
err Error number if failed conversion
|
|
811 |
|
|
812 |
NOTES:
|
|
813 |
If length is not INT_MAX32 or str[length] != 0 then the given str must
|
|
814 |
be writeable
|
|
815 |
If length == INT_MAX32 the str must be \0 terminated.
|
|
816 |
||
817 |
It's implemented this way to save a buffer allocation and a memory copy.
|
|
818 |
||
819 |
RETURN
|
|
820 |
Value of number in string
|
|
821 |
*/
|
|
822 |
||
823 |
||
824 |
double my_strntod_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
825 |
char *str, size_t length, |
|
826 |
char **end, int *err) |
|
827 |
{
|
|
828 |
if (length == INT_MAX32) |
|
829 |
length= 65535; /* Should be big enough */ |
|
830 |
*end= str + length; |
|
831 |
return my_strtod(str, end, err); |
|
832 |
}
|
|
833 |
||
834 |
||
835 |
/*
|
|
836 |
This is a fast version optimized for the case of radix 10 / -10
|
|
837 |
||
838 |
Assume len >= 1
|
|
839 |
*/
|
|
840 |
||
841 |
size_t my_long10_to_str_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
842 |
char *dst, size_t len, int radix, long int val) |
|
843 |
{
|
|
844 |
char buffer[66]; |
|
845 |
register char *p, *e; |
|
846 |
long int new_val; |
|
847 |
uint sign=0; |
|
848 |
unsigned long int uval = (unsigned long int) val; |
|
849 |
||
850 |
e = p = &buffer[sizeof(buffer)-1]; |
|
851 |
*p= 0; |
|
852 |
||
853 |
if (radix < 0) |
|
854 |
{
|
|
855 |
if (val < 0) |
|
856 |
{
|
|
857 |
/* Avoid integer overflow in (-val) for LONGLONG_MIN (BUG#31799). */
|
|
858 |
uval= (unsigned long int)0 - uval; |
|
859 |
*dst++= '-'; |
|
860 |
len--; |
|
861 |
sign= 1; |
|
862 |
}
|
|
863 |
}
|
|
864 |
||
865 |
new_val = (long) (uval / 10); |
|
866 |
*--p = '0'+ (char) (uval - (unsigned long) new_val * 10); |
|
867 |
val = new_val; |
|
868 |
||
869 |
while (val != 0) |
|
870 |
{
|
|
871 |
new_val=val/10; |
|
872 |
*--p = '0' + (char) (val-new_val*10); |
|
873 |
val= new_val; |
|
874 |
}
|
|
875 |
||
876 |
len= min(len, (size_t) (e-p)); |
|
877 |
memcpy(dst, p, len); |
|
878 |
return len+sign; |
|
879 |
}
|
|
880 |
||
881 |
||
882 |
size_t my_longlong10_to_str_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
883 |
char *dst, size_t len, int radix, |
|
884 |
longlong val) |
|
885 |
{
|
|
886 |
char buffer[65]; |
|
887 |
register char *p, *e; |
|
888 |
long long_val; |
|
889 |
uint sign= 0; |
|
890 |
ulonglong uval = (ulonglong)val; |
|
891 |
||
892 |
if (radix < 0) |
|
893 |
{
|
|
894 |
if (val < 0) |
|
895 |
{
|
|
896 |
/* Avoid integer overflow in (-val) for LONGLONG_MIN (BUG#31799). */
|
|
897 |
uval = (ulonglong)0 - uval; |
|
898 |
*dst++= '-'; |
|
899 |
len--; |
|
900 |
sign= 1; |
|
901 |
}
|
|
902 |
}
|
|
903 |
||
904 |
e = p = &buffer[sizeof(buffer)-1]; |
|
905 |
*p= 0; |
|
906 |
||
907 |
if (uval == 0) |
|
908 |
{
|
|
909 |
*--p= '0'; |
|
910 |
len= 1; |
|
911 |
goto cnv; |
|
912 |
}
|
|
913 |
||
914 |
while (uval > (ulonglong) LONG_MAX) |
|
915 |
{
|
|
916 |
ulonglong quo= uval/(uint) 10; |
|
917 |
uint rem= (uint) (uval- quo* (uint) 10); |
|
918 |
*--p = '0' + rem; |
|
919 |
uval= quo; |
|
920 |
}
|
|
921 |
||
922 |
long_val= (long) uval; |
|
923 |
while (long_val != 0) |
|
924 |
{
|
|
925 |
long quo= long_val/10; |
|
926 |
*--p = (char) ('0' + (long_val - quo*10)); |
|
927 |
long_val= quo; |
|
928 |
}
|
|
929 |
||
930 |
len= min(len, (size_t) (e-p)); |
|
931 |
cnv: |
|
932 |
memcpy(dst, p, len); |
|
933 |
return len+sign; |
|
934 |
}
|
|
935 |
||
936 |
||
937 |
/*
|
|
938 |
** Compare string against string with wildcard
|
|
939 |
** 0 if matched
|
|
940 |
** -1 if not matched with wildcard
|
|
941 |
** 1 if matched with wildcard
|
|
942 |
*/
|
|
943 |
||
944 |
#ifdef LIKE_CMP_TOUPPER
|
|
945 |
#define likeconv(s,A) (uchar) my_toupper(s,A)
|
|
946 |
#else
|
|
947 |
#define likeconv(s,A) (uchar) (s)->sort_order[(uchar) (A)]
|
|
948 |
#endif
|
|
949 |
||
950 |
#define INC_PTR(cs,A,B) (A)++
|
|
951 |
||
952 |
||
953 |
int my_wildcmp_8bit(CHARSET_INFO *cs, |
|
954 |
const char *str,const char *str_end, |
|
955 |
const char *wildstr,const char *wildend, |
|
956 |
int escape, int w_one, int w_many) |
|
957 |
{
|
|
958 |
int result= -1; /* Not found, using wildcards */ |
|
959 |
||
960 |
while (wildstr != wildend) |
|
961 |
{
|
|
962 |
while (*wildstr != w_many && *wildstr != w_one) |
|
963 |
{
|
|
964 |
if (*wildstr == escape && wildstr+1 != wildend) |
|
965 |
wildstr++; |
|
966 |
||
967 |
if (str == str_end || likeconv(cs,*wildstr++) != likeconv(cs,*str++)) |
|
968 |
return(1); /* No match */ |
|
969 |
if (wildstr == wildend) |
|
970 |
return(str != str_end); /* Match if both are at end */ |
|
971 |
result=1; /* Found an anchor char */ |
|
972 |
}
|
|
973 |
if (*wildstr == w_one) |
|
974 |
{
|
|
975 |
do
|
|
976 |
{
|
|
977 |
if (str == str_end) /* Skip one char if possible */ |
|
978 |
return(result); |
|
979 |
INC_PTR(cs,str,str_end); |
|
980 |
} while (++wildstr < wildend && *wildstr == w_one); |
|
981 |
if (wildstr == wildend) |
|
982 |
break; |
|
983 |
}
|
|
984 |
if (*wildstr == w_many) |
|
985 |
{ /* Found w_many */ |
|
986 |
uchar cmp; |
|
987 |
||
988 |
wildstr++; |
|
989 |
/* Remove any '%' and '_' from the wild search string */
|
|
990 |
for (; wildstr != wildend ; wildstr++) |
|
991 |
{
|
|
992 |
if (*wildstr == w_many) |
|
993 |
continue; |
|
994 |
if (*wildstr == w_one) |
|
995 |
{
|
|
996 |
if (str == str_end) |
|
997 |
return(-1); |
|
998 |
INC_PTR(cs,str,str_end); |
|
999 |
continue; |
|
1000 |
}
|
|
1001 |
break; /* Not a wild character */ |
|
1002 |
}
|
|
1003 |
if (wildstr == wildend) |
|
1004 |
return(0); /* Ok if w_many is last */ |
|
1005 |
if (str == str_end) |
|
1006 |
return(-1); |
|
1007 |
||
1008 |
if ((cmp= *wildstr) == escape && wildstr+1 != wildend) |
|
1009 |
cmp= *++wildstr; |
|
1010 |
||
1011 |
INC_PTR(cs,wildstr,wildend); /* This is compared trough cmp */ |
|
1012 |
cmp=likeconv(cs,cmp); |
|
1013 |
do
|
|
1014 |
{
|
|
1015 |
while (str != str_end && (uchar) likeconv(cs,*str) != cmp) |
|
1016 |
str++; |
|
1017 |
if (str++ == str_end) return(-1); |
|
1018 |
{
|
|
1019 |
int tmp=my_wildcmp_8bit(cs,str,str_end,wildstr,wildend,escape,w_one, |
|
1020 |
w_many); |
|
1021 |
if (tmp <= 0) |
|
1022 |
return(tmp); |
|
1023 |
}
|
|
1024 |
} while (str != str_end && wildstr[0] != w_many); |
|
1025 |
return(-1); |
|
1026 |
}
|
|
1027 |
}
|
|
1028 |
return(str != str_end ? 1 : 0); |
|
1029 |
}
|
|
1030 |
||
1031 |
||
1032 |
/*
|
|
1033 |
** Calculate min_str and max_str that ranges a LIKE string.
|
|
1034 |
** Arguments:
|
|
1035 |
** ptr Pointer to LIKE string.
|
|
1036 |
** ptr_length Length of LIKE string.
|
|
1037 |
** escape Escape character in LIKE. (Normally '\').
|
|
1038 |
** All escape characters should be removed from min_str and max_str
|
|
1039 |
** res_length Length of min_str and max_str.
|
|
1040 |
** min_str Smallest case sensitive string that ranges LIKE.
|
|
1041 |
** Should be space padded to res_length.
|
|
1042 |
** max_str Largest case sensitive string that ranges LIKE.
|
|
1043 |
** Normally padded with the biggest character sort value.
|
|
1044 |
**
|
|
1045 |
** The function should return 0 if ok and 1 if the LIKE string can't be
|
|
1046 |
** optimized !
|
|
1047 |
*/
|
|
1048 |
||
1049 |
my_bool my_like_range_simple(CHARSET_INFO *cs, |
|
1050 |
const char *ptr, size_t ptr_length, |
|
1051 |
pbool escape, pbool w_one, pbool w_many, |
|
1052 |
size_t res_length, |
|
1053 |
char *min_str,char *max_str, |
|
1054 |
size_t *min_length, size_t *max_length) |
|
1055 |
{
|
|
1056 |
const char *end= ptr + ptr_length; |
|
1057 |
char *min_org=min_str; |
|
1058 |
char *min_end=min_str+res_length; |
|
1059 |
size_t charlen= res_length / cs->mbmaxlen; |
|
1060 |
||
1061 |
for (; ptr != end && min_str != min_end && charlen > 0 ; ptr++, charlen--) |
|
1062 |
{
|
|
1063 |
if (*ptr == escape && ptr+1 != end) |
|
1064 |
{
|
|
1065 |
ptr++; /* Skip escape */ |
|
1066 |
*min_str++= *max_str++ = *ptr; |
|
1067 |
continue; |
|
1068 |
}
|
|
1069 |
if (*ptr == w_one) /* '_' in SQL */ |
|
1070 |
{
|
|
1071 |
*min_str++='\0'; /* This should be min char */ |
|
1072 |
*max_str++= (char) cs->max_sort_char; |
|
1073 |
continue; |
|
1074 |
}
|
|
1075 |
if (*ptr == w_many) /* '%' in SQL */ |
|
1076 |
{
|
|
1077 |
/* Calculate length of keys */
|
|
1078 |
*min_length= ((cs->state & MY_CS_BINSORT) ? |
|
1079 |
(size_t) (min_str - min_org) : |
|
1080 |
res_length); |
|
1081 |
*max_length= res_length; |
|
1082 |
do
|
|
1083 |
{
|
|
1084 |
*min_str++= 0; |
|
1085 |
*max_str++= (char) cs->max_sort_char; |
|
1086 |
} while (min_str != min_end); |
|
1087 |
return 0; |
|
1088 |
}
|
|
1089 |
*min_str++= *max_str++ = *ptr; |
|
1090 |
}
|
|
1091 |
||
1092 |
*min_length= *max_length = (size_t) (min_str - min_org); |
|
1093 |
while (min_str != min_end) |
|
1094 |
*min_str++= *max_str++ = ' '; /* Because if key compression */ |
|
1095 |
return 0; |
|
1096 |
}
|
|
1097 |
||
1098 |
||
1099 |
size_t my_scan_8bit(CHARSET_INFO *cs, const char *str, const char *end, int sq) |
|
1100 |
{
|
|
1101 |
const char *str0= str; |
|
1102 |
switch (sq) |
|
1103 |
{
|
|
1104 |
case MY_SEQ_INTTAIL: |
|
1105 |
if (*str == '.') |
|
1106 |
{
|
|
1107 |
for(str++ ; str != end && *str == '0' ; str++); |
|
1108 |
return (size_t) (str - str0); |
|
1109 |
}
|
|
1110 |
return 0; |
|
1111 |
||
1112 |
case MY_SEQ_SPACES: |
|
1113 |
for ( ; str < end ; str++) |
|
1114 |
{
|
|
1115 |
if (!my_isspace(cs,*str)) |
|
1116 |
break; |
|
1117 |
}
|
|
1118 |
return (size_t) (str - str0); |
|
1119 |
default: |
|
1120 |
return 0; |
|
1121 |
}
|
|
1122 |
}
|
|
1123 |
||
1124 |
||
1125 |
void my_fill_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
1126 |
char *s, size_t l, int fill) |
|
1127 |
{
|
|
1128 |
bfill((uchar*) s,l,fill); |
|
1129 |
}
|
|
1130 |
||
1131 |
||
1132 |
size_t my_numchars_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
1133 |
const char *b, const char *e) |
|
1134 |
{
|
|
1135 |
return (size_t) (e - b); |
|
1136 |
}
|
|
1137 |
||
1138 |
||
1139 |
size_t my_numcells_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
1140 |
const char *b, const char *e) |
|
1141 |
{
|
|
1142 |
return (size_t) (e - b); |
|
1143 |
}
|
|
1144 |
||
1145 |
||
1146 |
size_t my_charpos_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
1147 |
const char *b __attribute__((unused)), |
|
1148 |
const char *e __attribute__((unused)), |
|
1149 |
size_t pos) |
|
1150 |
{
|
|
1151 |
return pos; |
|
1152 |
}
|
|
1153 |
||
1154 |
||
1155 |
size_t my_well_formed_len_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
1156 |
const char *start, const char *end, |
|
1157 |
size_t nchars, int *error) |
|
1158 |
{
|
|
1159 |
size_t nbytes= (size_t) (end-start); |
|
1160 |
*error= 0; |
|
1161 |
return min(nbytes, nchars); |
|
1162 |
}
|
|
1163 |
||
1164 |
||
1165 |
size_t my_lengthsp_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
1166 |
const char *ptr, size_t length) |
|
1167 |
{
|
|
1168 |
const char *end; |
|
1169 |
end= (const char *) skip_trailing_space((const uchar *)ptr, length); |
|
1170 |
return (size_t) (end-ptr); |
|
1171 |
}
|
|
1172 |
||
1173 |
||
1174 |
uint my_instr_simple(CHARSET_INFO *cs, |
|
1175 |
const char *b, size_t b_length, |
|
1176 |
const char *s, size_t s_length, |
|
1177 |
my_match_t *match, uint nmatch) |
|
1178 |
{
|
|
1179 |
register const uchar *str, *search, *end, *search_end; |
|
1180 |
||
1181 |
if (s_length <= b_length) |
|
1182 |
{
|
|
1183 |
if (!s_length) |
|
1184 |
{
|
|
1185 |
if (nmatch) |
|
1186 |
{
|
|
1187 |
match->beg= 0; |
|
1188 |
match->end= 0; |
|
1189 |
match->mb_len= 0; |
|
1190 |
}
|
|
1191 |
return 1; /* Empty string is always found */ |
|
1192 |
}
|
|
1193 |
||
1194 |
str= (const uchar*) b; |
|
1195 |
search= (const uchar*) s; |
|
1196 |
end= (const uchar*) b+b_length-s_length+1; |
|
1197 |
search_end= (const uchar*) s + s_length; |
|
1198 |
||
1199 |
skip: |
|
1200 |
while (str != end) |
|
1201 |
{
|
|
1202 |
if (cs->sort_order[*str++] == cs->sort_order[*search]) |
|
1203 |
{
|
|
1204 |
register const uchar *i,*j; |
|
1205 |
||
1206 |
i= str; |
|
1207 |
j= search+1; |
|
1208 |
||
1209 |
while (j != search_end) |
|
1210 |
if (cs->sort_order[*i++] != cs->sort_order[*j++]) |
|
1211 |
goto skip; |
|
1212 |
||
1213 |
if (nmatch > 0) |
|
1214 |
{
|
|
1215 |
match[0].beg= 0; |
|
1216 |
match[0].end= (size_t) (str- (const uchar*)b-1); |
|
1217 |
match[0].mb_len= match[0].end; |
|
1218 |
||
1219 |
if (nmatch > 1) |
|
1220 |
{
|
|
1221 |
match[1].beg= match[0].end; |
|
1222 |
match[1].end= match[0].end+s_length; |
|
1223 |
match[1].mb_len= match[1].end-match[1].beg; |
|
1224 |
}
|
|
1225 |
}
|
|
1226 |
return 2; |
|
1227 |
}
|
|
1228 |
}
|
|
1229 |
}
|
|
1230 |
return 0; |
|
1231 |
}
|
|
1232 |
||
1233 |
||
1234 |
typedef struct |
|
1235 |
{
|
|
1236 |
int nchars; |
|
1237 |
MY_UNI_IDX uidx; |
|
1238 |
} uni_idx; |
|
1239 |
||
1240 |
#define PLANE_SIZE 0x100
|
|
1241 |
#define PLANE_NUM 0x100
|
|
1242 |
#define PLANE_NUMBER(x) (((x)>>8) % PLANE_NUM)
|
|
1243 |
||
1244 |
static int pcmp(const void * f, const void * s) |
|
1245 |
{
|
|
1246 |
const uni_idx *F= (const uni_idx*) f; |
|
1247 |
const uni_idx *S= (const uni_idx*) s; |
|
1248 |
int res; |
|
1249 |
||
1250 |
if (!(res=((S->nchars)-(F->nchars)))) |
|
1251 |
res=((F->uidx.from)-(S->uidx.to)); |
|
1252 |
return res; |
|
1253 |
}
|
|
1254 |
||
1255 |
static my_bool create_fromuni(CHARSET_INFO *cs, void *(*alloc)(size_t)) |
|
1256 |
{
|
|
1257 |
uni_idx idx[PLANE_NUM]; |
|
1258 |
int i,n; |
|
1259 |
||
1260 |
/*
|
|
1261 |
Check that Unicode map is loaded.
|
|
1262 |
It can be not loaded when the collation is
|
|
1263 |
listed in Index.xml but not specified
|
|
1264 |
in the character set specific XML file.
|
|
1265 |
*/
|
|
1266 |
if (!cs->tab_to_uni) |
|
1267 |
return TRUE; |
|
1268 |
||
1269 |
/* Clear plane statistics */
|
|
1270 |
bzero(idx,sizeof(idx)); |
|
1271 |
||
1272 |
/* Count number of characters in each plane */
|
|
1273 |
for (i=0; i< 0x100; i++) |
|
1274 |
{
|
|
1275 |
uint16 wc=cs->tab_to_uni[i]; |
|
1276 |
int pl= PLANE_NUMBER(wc); |
|
1277 |
||
1278 |
if (wc || !i) |
|
1279 |
{
|
|
1280 |
if (!idx[pl].nchars) |
|
1281 |
{
|
|
1282 |
idx[pl].uidx.from=wc; |
|
1283 |
idx[pl].uidx.to=wc; |
|
1284 |
}else |
|
1285 |
{
|
|
1286 |
idx[pl].uidx.from=wc<idx[pl].uidx.from?wc:idx[pl].uidx.from; |
|
1287 |
idx[pl].uidx.to=wc>idx[pl].uidx.to?wc:idx[pl].uidx.to; |
|
1288 |
}
|
|
1289 |
idx[pl].nchars++; |
|
1290 |
}
|
|
1291 |
}
|
|
1292 |
||
1293 |
/* Sort planes in descending order */
|
|
1294 |
qsort(&idx,PLANE_NUM,sizeof(uni_idx),&pcmp); |
|
1295 |
||
1296 |
for (i=0; i < PLANE_NUM; i++) |
|
1297 |
{
|
|
1298 |
int ch,numchars; |
|
1299 |
||
1300 |
/* Skip empty plane */
|
|
1301 |
if (!idx[i].nchars) |
|
1302 |
break; |
|
1303 |
||
1304 |
numchars=idx[i].uidx.to-idx[i].uidx.from+1; |
|
1305 |
if (!(idx[i].uidx.tab=(uchar*) alloc(numchars * sizeof(*idx[i].uidx.tab)))) |
|
1306 |
return TRUE; |
|
1307 |
||
1308 |
bzero(idx[i].uidx.tab,numchars*sizeof(*idx[i].uidx.tab)); |
|
1309 |
||
1310 |
for (ch=1; ch < PLANE_SIZE; ch++) |
|
1311 |
{
|
|
1312 |
uint16 wc=cs->tab_to_uni[ch]; |
|
1313 |
if (wc >= idx[i].uidx.from && wc <= idx[i].uidx.to && wc) |
|
1314 |
{
|
|
1315 |
int ofs= wc - idx[i].uidx.from; |
|
1316 |
idx[i].uidx.tab[ofs]= ch; |
|
1317 |
}
|
|
1318 |
}
|
|
1319 |
}
|
|
1320 |
||
1321 |
/* Allocate and fill reverse table for each plane */
|
|
1322 |
n=i; |
|
1323 |
if (!(cs->tab_from_uni= (MY_UNI_IDX*) alloc(sizeof(MY_UNI_IDX)*(n+1)))) |
|
1324 |
return TRUE; |
|
1325 |
||
1326 |
for (i=0; i< n; i++) |
|
1327 |
cs->tab_from_uni[i]= idx[i].uidx; |
|
1328 |
||
1329 |
/* Set end-of-list marker */
|
|
1330 |
bzero(&cs->tab_from_uni[i],sizeof(MY_UNI_IDX)); |
|
1331 |
return FALSE; |
|
1332 |
}
|
|
1333 |
||
1334 |
static my_bool my_cset_init_8bit(CHARSET_INFO *cs, void *(*alloc)(size_t)) |
|
1335 |
{
|
|
1336 |
cs->caseup_multiply= 1; |
|
1337 |
cs->casedn_multiply= 1; |
|
1338 |
cs->pad_char= ' '; |
|
1339 |
return create_fromuni(cs, alloc); |
|
1340 |
}
|
|
1341 |
||
1342 |
static void set_max_sort_char(CHARSET_INFO *cs) |
|
1343 |
{
|
|
1344 |
uchar max_char; |
|
1345 |
uint i; |
|
1346 |
||
1347 |
if (!cs->sort_order) |
|
1348 |
return; |
|
1349 |
||
1350 |
max_char=cs->sort_order[(uchar) cs->max_sort_char]; |
|
1351 |
for (i= 0; i < 256; i++) |
|
1352 |
{
|
|
1353 |
if ((uchar) cs->sort_order[i] > max_char) |
|
1354 |
{
|
|
1355 |
max_char=(uchar) cs->sort_order[i]; |
|
1356 |
cs->max_sort_char= i; |
|
1357 |
}
|
|
1358 |
}
|
|
1359 |
}
|
|
1360 |
||
1361 |
static my_bool my_coll_init_simple(CHARSET_INFO *cs, |
|
1362 |
void *(*alloc)(size_t) __attribute__((unused))) |
|
1363 |
{
|
|
1364 |
set_max_sort_char(cs); |
|
1365 |
return FALSE; |
|
1366 |
}
|
|
1367 |
||
1368 |
||
1369 |
longlong my_strtoll10_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
1370 |
const char *nptr, char **endptr, int *error) |
|
1371 |
{
|
|
1372 |
return my_strtoll10(nptr, endptr, error); |
|
1373 |
}
|
|
1374 |
||
1375 |
||
1376 |
int my_mb_ctype_8bit(CHARSET_INFO *cs, int *ctype, |
|
1377 |
const uchar *s, const uchar *e) |
|
1378 |
{
|
|
1379 |
if (s >= e) |
|
1380 |
{
|
|
1381 |
*ctype= 0; |
|
1382 |
return MY_CS_TOOSMALL; |
|
1383 |
}
|
|
1384 |
*ctype= cs->ctype[*s + 1]; |
|
1385 |
return 1; |
|
1386 |
}
|
|
1387 |
||
1388 |
||
1389 |
#undef ULONGLONG_MAX
|
|
1390 |
/*
|
|
1391 |
Needed under MetroWerks Compiler, since MetroWerks compiler does not
|
|
1392 |
properly handle a constant expression containing a mod operator
|
|
1393 |
*/
|
|
1394 |
#if defined(__NETWARE__) && defined(__MWERKS__)
|
|
1395 |
static ulonglong ulonglong_max= ~(ulonglong) 0; |
|
1396 |
#define ULONGLONG_MAX ulonglong_max
|
|
1397 |
#else
|
|
1398 |
#define ULONGLONG_MAX (~(ulonglong) 0)
|
|
1399 |
#endif /* __NETWARE__ && __MWERKS__ */ |
|
1400 |
||
1401 |
||
1402 |
#define CUTOFF (ULONGLONG_MAX / 10)
|
|
1403 |
#define CUTLIM (ULONGLONG_MAX % 10)
|
|
1404 |
#define DIGITS_IN_ULONGLONG 20
|
|
1405 |
||
1406 |
static ulonglong d10[DIGITS_IN_ULONGLONG]= |
|
1407 |
{
|
|
1408 |
1, |
|
1409 |
10, |
|
1410 |
100, |
|
1411 |
1000, |
|
1412 |
10000, |
|
1413 |
100000, |
|
1414 |
1000000, |
|
1415 |
10000000, |
|
1416 |
100000000, |
|
1417 |
1000000000, |
|
1418 |
10000000000ULL, |
|
1419 |
100000000000ULL, |
|
1420 |
1000000000000ULL, |
|
1421 |
10000000000000ULL, |
|
1422 |
100000000000000ULL, |
|
1423 |
1000000000000000ULL, |
|
1424 |
10000000000000000ULL, |
|
1425 |
100000000000000000ULL, |
|
1426 |
1000000000000000000ULL, |
|
1427 |
10000000000000000000ULL
|
|
1428 |
};
|
|
1429 |
||
1430 |
||
1431 |
/*
|
|
1432 |
||
1433 |
Convert a string to unsigned long long integer value
|
|
1434 |
with rounding.
|
|
1435 |
|
|
1436 |
SYNOPSYS
|
|
1437 |
my_strntoull10_8bit()
|
|
1438 |
cs in pointer to character set
|
|
1439 |
str in pointer to the string to be converted
|
|
1440 |
length in string length
|
|
1441 |
unsigned_flag in whether the number is unsigned
|
|
1442 |
endptr out pointer to the stop character
|
|
1443 |
error out returned error code
|
|
1444 |
||
1445 |
DESCRIPTION
|
|
1446 |
This function takes the decimal representation of integer number
|
|
1447 |
from string str and converts it to an signed or unsigned
|
|
1448 |
long long integer value.
|
|
1449 |
Space characters and tab are ignored.
|
|
1450 |
A sign character might precede the digit characters.
|
|
1451 |
The number may have any number of pre-zero digits.
|
|
1452 |
The number may have decimal point and exponent.
|
|
1453 |
Rounding is always done in "away from zero" style:
|
|
1454 |
0.5 -> 1
|
|
1455 |
-0.5 -> -1
|
|
1456 |
||
1457 |
The function stops reading the string str after "length" bytes
|
|
1458 |
or at the first character that is not a part of correct number syntax:
|
|
1459 |
||
1460 |
<signed numeric literal> ::=
|
|
1461 |
[ <sign> ] <exact numeric literal> [ E [ <sign> ] <unsigned integer> ]
|
|
1462 |
||
1463 |
<exact numeric literal> ::=
|
|
1464 |
<unsigned integer> [ <period> [ <unsigned integer> ] ]
|
|
1465 |
| <period> <unsigned integer>
|
|
1466 |
<unsigned integer> ::= <digit>...
|
|
1467 |
|
|
1468 |
RETURN VALUES
|
|
1469 |
Value of string as a signed/unsigned longlong integer
|
|
1470 |
||
1471 |
endptr cannot be NULL. The function will store the end pointer
|
|
1472 |
to the stop character here.
|
|
1473 |
||
1474 |
The error parameter contains information how things went:
|
|
1475 |
0 ok
|
|
1476 |
ERANGE If the the value of the converted number is out of range
|
|
1477 |
In this case the return value is:
|
|
1478 |
- ULONGLONG_MAX if unsigned_flag and the number was too big
|
|
1479 |
- 0 if unsigned_flag and the number was negative
|
|
1480 |
- LONGLONG_MAX if no unsigned_flag and the number is too big
|
|
1481 |
- LONGLONG_MIN if no unsigned_flag and the number it too big negative
|
|
1482 |
|
|
1483 |
EDOM If the string didn't contain any digits.
|
|
1484 |
In this case the return value is 0.
|
|
1485 |
*/
|
|
1486 |
||
1487 |
ulonglong
|
|
1488 |
my_strntoull10rnd_8bit(CHARSET_INFO *cs __attribute__((unused)), |
|
1489 |
const char *str, size_t length, int unsigned_flag, |
|
1490 |
char **endptr, int *error) |
|
1491 |
{
|
|
1492 |
const char *dot, *end9, *beg, *end= str + length; |
|
1493 |
ulonglong ull; |
|
1494 |
ulong ul; |
|
1495 |
uchar ch; |
|
1496 |
int shift= 0, digits= 0, negative, addon; |
|
1497 |
||
1498 |
/* Skip leading spaces and tabs */
|
|
1499 |
for ( ; str < end && (*str == ' ' || *str == '\t') ; str++); |
|
1500 |
||
1501 |
if (str >= end) |
|
1502 |
goto ret_edom; |
|
1503 |
||
1504 |
if ((negative= (*str == '-')) || *str=='+') /* optional sign */ |
|
1505 |
{
|
|
1506 |
if (++str == end) |
|
1507 |
goto ret_edom; |
|
1508 |
}
|
|
1509 |
||
1510 |
beg= str; |
|
1511 |
end9= (str + 9) > end ? end : (str + 9); |
|
1512 |
/* Accumulate small number into ulong, for performance purposes */
|
|
1513 |
for (ul= 0 ; str < end9 && (ch= (uchar) (*str - '0')) < 10; str++) |
|
1514 |
{
|
|
1515 |
ul= ul * 10 + ch; |
|
1516 |
}
|
|
1517 |
||
1518 |
if (str >= end) /* Small number without dots and expanents */ |
|
1519 |
{
|
|
1520 |
*endptr= (char*) str; |
|
1521 |
if (negative) |
|
1522 |
{
|
|
1523 |
if (unsigned_flag) |
|
1524 |
{
|
|
1525 |
*error= ul ? MY_ERRNO_ERANGE : 0; |
|
1526 |
return 0; |
|
1527 |
}
|
|
1528 |
else
|
|
1529 |
{
|
|
1530 |
*error= 0; |
|
1531 |
return (ulonglong) (longlong) -(long) ul; |
|
1532 |
}
|
|
1533 |
}
|
|
1534 |
else
|
|
1535 |
{
|
|
1536 |
*error=0; |
|
1537 |
return (ulonglong) ul; |
|
1538 |
}
|
|
1539 |
}
|
|
1540 |
||
1541 |
digits= str - beg; |
|
1542 |
||
1543 |
/* Continue to accumulate into ulonglong */
|
|
1544 |
for (dot= NULL, ull= ul; str < end; str++) |
|
1545 |
{
|
|
1546 |
if ((ch= (uchar) (*str - '0')) < 10) |
|
1547 |
{
|
|
1548 |
if (ull < CUTOFF || (ull == CUTOFF && ch <= CUTLIM)) |
|
1549 |
{
|
|
1550 |
ull= ull * 10 + ch; |
|
1551 |
digits++; |
|
1552 |
continue; |
|
1553 |
}
|
|
1554 |
/*
|
|
1555 |
Adding the next digit would overflow.
|
|
1556 |
Remember the next digit in "addon", for rounding.
|
|
1557 |
Scan all digits with an optional single dot.
|
|
1558 |
*/
|
|
1559 |
if (ull == CUTOFF) |
|
1560 |
{
|
|
1561 |
ull= ULONGLONG_MAX; |
|
1562 |
addon= 1; |
|
1563 |
str++; |
|
1564 |
}
|
|
1565 |
else
|
|
1566 |
addon= (*str >= '5'); |
|
1567 |
if (!dot) |
|
1568 |
{
|
|
1569 |
for ( ; str < end && (ch= (uchar) (*str - '0')) < 10; shift++, str++); |
|
1570 |
if (str < end && *str == '.') |
|
1571 |
{
|
|
1572 |
str++; |
|
1573 |
for ( ; str < end && (ch= (uchar) (*str - '0')) < 10; str++); |
|
1574 |
}
|
|
1575 |
}
|
|
1576 |
else
|
|
1577 |
{
|
|
1578 |
shift= dot - str; |
|
1579 |
for ( ; str < end && (ch= (uchar) (*str - '0')) < 10; str++); |
|
1580 |
}
|
|
1581 |
goto exp; |
|
1582 |
}
|
|
1583 |
||
1584 |
if (*str == '.') |
|
1585 |
{
|
|
1586 |
if (dot) |
|
1587 |
{
|
|
1588 |
/* The second dot character */
|
|
1589 |
addon= 0; |
|
1590 |
goto exp; |
|
1591 |
}
|
|
1592 |
else
|
|
1593 |
{
|
|
1594 |
dot= str + 1; |
|
1595 |
}
|
|
1596 |
continue; |
|
1597 |
}
|
|
1598 |
||
1599 |
/* Unknown character, exit the loop */
|
|
1600 |
break; |
|
1601 |
}
|
|
1602 |
shift= dot ? dot - str : 0; /* Right shift */ |
|
1603 |
addon= 0; |
|
1604 |
||
1605 |
exp: /* [ E [ <sign> ] <unsigned integer> ] */ |
|
1606 |
||
1607 |
if (!digits) |
|
1608 |
{
|
|
1609 |
str= beg; |
|
1610 |
goto ret_edom; |
|
1611 |
}
|
|
1612 |
||
1613 |
if (str < end && (*str == 'e' || *str == 'E')) |
|
1614 |
{
|
|
1615 |
str++; |
|
1616 |
if (str < end) |
|
1617 |
{
|
|
1618 |
int negative_exp, exponent; |
|
1619 |
if ((negative_exp= (*str == '-')) || *str=='+') |
|
1620 |
{
|
|
1621 |
if (++str == end) |
|
1622 |
goto ret_sign; |
|
1623 |
}
|
|
1624 |
for (exponent= 0 ; |
|
1625 |
str < end && (ch= (uchar) (*str - '0')) < 10; |
|
1626 |
str++) |
|
1627 |
{
|
|
1628 |
exponent= exponent * 10 + ch; |
|
1629 |
}
|
|
1630 |
shift+= negative_exp ? -exponent : exponent; |
|
1631 |
}
|
|
1632 |
}
|
|
1633 |
||
1634 |
if (shift == 0) /* No shift, check addon digit */ |
|
1635 |
{
|
|
1636 |
if (addon) |
|
1637 |
{
|
|
1638 |
if (ull == ULONGLONG_MAX) |
|
1639 |
goto ret_too_big; |
|
1640 |
ull++; |
|
1641 |
}
|
|
1642 |
goto ret_sign; |
|
1643 |
}
|
|
1644 |
||
1645 |
if (shift < 0) /* Right shift */ |
|
1646 |
{
|
|
1647 |
ulonglong d, r; |
|
1648 |
||
1649 |
if (-shift >= DIGITS_IN_ULONGLONG) |
|
1650 |
goto ret_zero; /* Exponent is a big negative number, return 0 */ |
|
1651 |
||
1652 |
d= d10[-shift]; |
|
1653 |
r= (ull % d) * 2; |
|
1654 |
ull /= d; |
|
1655 |
if (r >= d) |
|
1656 |
ull++; |
|
1657 |
goto ret_sign; |
|
1658 |
}
|
|
1659 |
||
1660 |
if (shift > DIGITS_IN_ULONGLONG) /* Huge left shift */ |
|
1661 |
{
|
|
1662 |
if (!ull) |
|
1663 |
goto ret_sign; |
|
1664 |
goto ret_too_big; |
|
1665 |
}
|
|
1666 |
||
1667 |
for ( ; shift > 0; shift--, ull*= 10) /* Left shift */ |
|
1668 |
{
|
|
1669 |
if (ull > CUTOFF) |
|
1670 |
goto ret_too_big; /* Overflow, number too big */ |
|
1671 |
}
|
|
1672 |
||
1673 |
ret_sign: |
|
1674 |
*endptr= (char*) str; |
|
1675 |
||
1676 |
if (!unsigned_flag) |
|
1677 |
{
|
|
1678 |
if (negative) |
|
1679 |
{
|
|
1680 |
if (ull > (ulonglong) LONGLONG_MIN) |
|
1681 |
{
|
|
1682 |
*error= MY_ERRNO_ERANGE; |
|
1683 |
return (ulonglong) LONGLONG_MIN; |
|
1684 |
}
|
|
1685 |
*error= 0; |
|
1686 |
return (ulonglong) -(longlong) ull; |
|
1687 |
}
|
|
1688 |
else
|
|
1689 |
{
|
|
1690 |
if (ull > (ulonglong) LONGLONG_MAX) |
|
1691 |
{
|
|
1692 |
*error= MY_ERRNO_ERANGE; |
|
1693 |
return (ulonglong) LONGLONG_MAX; |
|
1694 |
}
|
|
1695 |
*error= 0; |
|
1696 |
return ull; |
|
1697 |
}
|
|
1698 |
}
|
|
1699 |
||
1700 |
/* Unsigned number */
|
|
1701 |
if (negative && ull) |
|
1702 |
{
|
|
1703 |
*error= MY_ERRNO_ERANGE; |
|
1704 |
return 0; |
|
1705 |
}
|
|
1706 |
*error= 0; |
|
1707 |
return ull; |
|
1708 |
||
1709 |
ret_zero: |
|
1710 |
*endptr= (char*) str; |
|
1711 |
*error= 0; |
|
1712 |
return 0; |
|
1713 |
||
1714 |
ret_edom: |
|
1715 |
*endptr= (char*) str; |
|
1716 |
*error= MY_ERRNO_EDOM; |
|
1717 |
return 0; |
|
1718 |
||
1719 |
ret_too_big: |
|
1720 |
*endptr= (char*) str; |
|
1721 |
*error= MY_ERRNO_ERANGE; |
|
1722 |
return unsigned_flag ? |
|
1723 |
ULONGLONG_MAX : |
|
1724 |
negative ? (ulonglong) LONGLONG_MIN : (ulonglong) LONGLONG_MAX; |
|
1725 |
}
|
|
1726 |
||
1727 |
||
1728 |
/*
|
|
1729 |
Check if a constant can be propagated
|
|
1730 |
||
1731 |
SYNOPSIS:
|
|
1732 |
my_propagate_simple()
|
|
1733 |
cs Character set information
|
|
1734 |
str String to convert to double
|
|
1735 |
length Optional length for string.
|
|
1736 |
|
|
1737 |
NOTES:
|
|
1738 |
Takes the string in the given charset and check
|
|
1739 |
if it can be safely propagated in the optimizer.
|
|
1740 |
|
|
1741 |
create table t1 (
|
|
1742 |
s char(5) character set latin1 collate latin1_german2_ci);
|
|
1743 |
insert into t1 values (0xf6); -- o-umlaut
|
|
1744 |
select * from t1 where length(s)=1 and s='oe';
|
|
1745 |
||
1746 |
The above query should return one row.
|
|
1747 |
We cannot convert this query into:
|
|
1748 |
select * from t1 where length('oe')=1 and s='oe';
|
|
1749 |
|
|
1750 |
Currently we don't check the constant itself,
|
|
1751 |
and decide not to propagate a constant
|
|
1752 |
just if the collation itself allows tricky things
|
|
1753 |
like expansions and contractions. In the future
|
|
1754 |
we can write a more sophisticated functions to
|
|
1755 |
check the constants. For example, 'oa' can always
|
|
1756 |
be safety propagated in German2 because unlike
|
|
1757 |
'oe' it does not have any special meaning.
|
|
1758 |
||
1759 |
RETURN
|
|
1760 |
1 if constant can be safely propagated
|
|
1761 |
0 if it is not safe to propagate the constant
|
|
1762 |
*/
|
|
1763 |
||
1764 |
||
1765 |
||
1766 |
my_bool my_propagate_simple(CHARSET_INFO *cs __attribute__((unused)), |
|
1767 |
const uchar *str __attribute__((unused)), |
|
1768 |
size_t length __attribute__((unused))) |
|
1769 |
{
|
|
1770 |
return 1; |
|
1771 |
}
|
|
1772 |
||
1773 |
||
1774 |
my_bool my_propagate_complex(CHARSET_INFO *cs __attribute__((unused)), |
|
1775 |
const uchar *str __attribute__((unused)), |
|
1776 |
size_t length __attribute__((unused))) |
|
1777 |
{
|
|
1778 |
return 0; |
|
1779 |
}
|
|
1780 |
||
1781 |
||
1782 |
||
1783 |
/*
|
|
1784 |
Normalize strxfrm flags
|
|
1785 |
||
1786 |
SYNOPSIS:
|
|
1787 |
my_strxfrm_flag_normalize()
|
|
1788 |
flags - non-normalized flags
|
|
1789 |
nlevels - number of levels
|
|
1790 |
|
|
1791 |
NOTES:
|
|
1792 |
If levels are omitted, then 1-maximum is assumed.
|
|
1793 |
If any level number is greater than the maximum,
|
|
1794 |
it is treated as the maximum.
|
|
1795 |
||
1796 |
RETURN
|
|
1797 |
normalized flags
|
|
1798 |
*/
|
|
1799 |
||
1800 |
uint my_strxfrm_flag_normalize(uint flags, uint maximum) |
|
1801 |
{
|
|
1802 |
DBUG_ASSERT(maximum >= 1 && maximum <= MY_STRXFRM_NLEVELS); |
|
1803 |
||
1804 |
/* If levels are omitted, then 1-maximum is assumed*/
|
|
1805 |
if (!(flags & MY_STRXFRM_LEVEL_ALL)) |
|
1806 |
{
|
|
1807 |
static uint def_level_flags[]= {0, 0x01, 0x03, 0x07, 0x0F, 0x1F, 0x3F }; |
|
1808 |
uint flag_pad= flags & MY_STRXFRM_PAD_WITH_SPACE; |
|
1809 |
flags= def_level_flags[maximum] | flag_pad; |
|
1810 |
}
|
|
1811 |
else
|
|
1812 |
{
|
|
1813 |
uint i; |
|
1814 |
uint flag_lev= flags & MY_STRXFRM_LEVEL_ALL; |
|
1815 |
uint flag_dsc= (flags >> MY_STRXFRM_DESC_SHIFT) & MY_STRXFRM_LEVEL_ALL; |
|
1816 |
uint flag_rev= (flags >> MY_STRXFRM_REVERSE_SHIFT) & MY_STRXFRM_LEVEL_ALL; |
|
1817 |
uint flag_pad= flags & MY_STRXFRM_PAD_WITH_SPACE; |
|
1818 |
||
1819 |
/*
|
|
1820 |
If any level number is greater than the maximum,
|
|
1821 |
it is treated as the maximum.
|
|
1822 |
*/
|
|
1823 |
for (maximum--, flags= 0, i= 0; i < MY_STRXFRM_NLEVELS; i++) |
|
1824 |
{
|
|
1825 |
uint src_bit= 1 << i; |
|
1826 |
uint dst_bit= 1 << min(i, maximum); |
|
1827 |
if (flag_lev & src_bit) |
|
1828 |
{
|
|
1829 |
flags|= dst_bit; |
|
1830 |
flags|= (flag_dsc & dst_bit) << MY_STRXFRM_DESC_SHIFT; |
|
1831 |
flags|= (flag_rev & dst_bit) << MY_STRXFRM_REVERSE_SHIFT; |
|
1832 |
}
|
|
1833 |
else
|
|
1834 |
{
|
|
1835 |
/* Check that there are no DESC or REVERSE flag for skipped level */
|
|
1836 |
DBUG_ASSERT(!(flag_dsc & src_bit) && !(flag_rev & src_bit)); |
|
1837 |
}
|
|
1838 |
}
|
|
1839 |
flags|= flag_pad; |
|
1840 |
}
|
|
1841 |
||
1842 |
return flags; |
|
1843 |
}
|
|
1844 |
||
1845 |
/*
|
|
1846 |
Apply DESC and REVERSE collation rules.
|
|
1847 |
||
1848 |
SYNOPSIS:
|
|
1849 |
my_strxfrm_desc_and_reverse()
|
|
1850 |
str - pointer to string
|
|
1851 |
strend - end of string
|
|
1852 |
flags - flags
|
|
1853 |
level - which level, starting from 0.
|
|
1854 |
|
|
1855 |
NOTES:
|
|
1856 |
Apply DESC or REVERSE or both flags.
|
|
1857 |
|
|
1858 |
If DESC flag is given, then the weights
|
|
1859 |
come out NOTed or negated for that level.
|
|
1860 |
|
|
1861 |
If REVERSE flags is given, then the weights come out in
|
|
1862 |
reverse order for that level, that is, starting with
|
|
1863 |
the last character and ending with the first character.
|
|
1864 |
|
|
1865 |
If nether DESC nor REVERSE flags are give,
|
|
1866 |
the string is not changed.
|
|
1867 |
|
|
1868 |
*/
|
|
1869 |
void my_strxfrm_desc_and_reverse(uchar *str, uchar *strend, |
|
1870 |
uint flags, uint level) |
|
1871 |
{
|
|
1872 |
if (flags & (MY_STRXFRM_DESC_LEVEL1 << level)) |
|
1873 |
{
|
|
1874 |
if (flags & (MY_STRXFRM_REVERSE_LEVEL1 << level)) |
|
1875 |
{
|
|
1876 |
for (strend--; str <= strend;) |
|
1877 |
{
|
|
1878 |
uchar tmp= *str; |
|
1879 |
*str++= ~*strend; |
|
1880 |
*strend--= ~tmp; |
|
1881 |
}
|
|
1882 |
}
|
|
1883 |
else
|
|
1884 |
{
|
|
1885 |
for (; str < strend; str++) |
|
1886 |
*str= ~*str; |
|
1887 |
}
|
|
1888 |
}
|
|
1889 |
else if (flags & (MY_STRXFRM_REVERSE_LEVEL1 << level)) |
|
1890 |
{
|
|
1891 |
for (strend--; str < strend;) |
|
1892 |
{
|
|
1893 |
uchar tmp= *str; |
|
1894 |
*str++= *strend; |
|
1895 |
*strend--= tmp; |
|
1896 |
}
|
|
1897 |
}
|
|
1898 |
}
|
|
1899 |
||
1900 |
||
1901 |
size_t
|
|
1902 |
my_strxfrm_pad_desc_and_reverse(CHARSET_INFO *cs, |
|
1903 |
uchar *str, uchar *frmend, uchar *strend, |
|
1904 |
uint nweights, uint flags, uint level) |
|
1905 |
{
|
|
1906 |
if (nweights && frmend < strend && (flags & MY_STRXFRM_PAD_WITH_SPACE)) |
|
1907 |
{
|
|
1908 |
uint fill_length= min((uint) (strend - frmend), nweights * cs->mbminlen); |
|
1909 |
cs->cset->fill(cs, (char*) frmend, fill_length, cs->pad_char); |
|
1910 |
frmend+= fill_length; |
|
1911 |
}
|
|
1912 |
my_strxfrm_desc_and_reverse(str, frmend, flags, level); |
|
1913 |
return frmend - str; |
|
1914 |
}
|
|
1915 |
||
1916 |
||
1917 |
MY_CHARSET_HANDLER my_charset_8bit_handler= |
|
1918 |
{
|
|
1919 |
my_cset_init_8bit, |
|
1920 |
NULL, /* ismbchar */ |
|
1921 |
my_mbcharlen_8bit, /* mbcharlen */ |
|
1922 |
my_numchars_8bit, |
|
1923 |
my_charpos_8bit, |
|
1924 |
my_well_formed_len_8bit, |
|
1925 |
my_lengthsp_8bit, |
|
1926 |
my_numcells_8bit, |
|
1927 |
my_mb_wc_8bit, |
|
1928 |
my_wc_mb_8bit, |
|
1929 |
my_mb_ctype_8bit, |
|
1930 |
my_caseup_str_8bit, |
|
1931 |
my_casedn_str_8bit, |
|
1932 |
my_caseup_8bit, |
|
1933 |
my_casedn_8bit, |
|
1934 |
my_snprintf_8bit, |
|
1935 |
my_long10_to_str_8bit, |
|
1936 |
my_longlong10_to_str_8bit, |
|
1937 |
my_fill_8bit, |
|
1938 |
my_strntol_8bit, |
|
1939 |
my_strntoul_8bit, |
|
1940 |
my_strntoll_8bit, |
|
1941 |
my_strntoull_8bit, |
|
1942 |
my_strntod_8bit, |
|
1943 |
my_strtoll10_8bit, |
|
1944 |
my_strntoull10rnd_8bit, |
|
1945 |
my_scan_8bit
|
|
1946 |
};
|
|
1947 |
||
1948 |
MY_COLLATION_HANDLER my_collation_8bit_simple_ci_handler = |
|
1949 |
{
|
|
1950 |
my_coll_init_simple, /* init */ |
|
1951 |
my_strnncoll_simple, |
|
1952 |
my_strnncollsp_simple, |
|
1953 |
my_strnxfrm_simple, |
|
1954 |
my_strnxfrmlen_simple, |
|
1955 |
my_like_range_simple, |
|
1956 |
my_wildcmp_8bit, |
|
1957 |
my_strcasecmp_8bit, |
|
1958 |
my_instr_simple, |
|
1959 |
my_hash_sort_simple, |
|
1960 |
my_propagate_simple
|
|
1961 |
};
|