1
by brian
clean slate |
1 |
/* Copyright (C) 2000 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 |
/* This file is originally from the mysql distribution. Coded by monty */
|
|
17 |
||
212.5.39
by Monty Taylor
Phew. Moved my_base and my_global. |
18 |
#include "global.h" |
212.5.13
by Monty Taylor
Moved my_sys/my_pthread/my_nosys and mysys_err to mysys. |
19 |
#include <mysys/my_sys.h> |
212.5.45
by Monty Taylor
Removed excess AM_CPPFLAGS from the tree. Now the only thing that should be in the include path should be -I${top_srcdir} and -I${top_builddir}w |
20 |
#include <mystrings/m_string.h> |
1
by brian
clean slate |
21 |
|
398.1.5
by Monty Taylor
Removed C++ includes and std namespace from global.h. |
22 |
#include <algorithm> |
23 |
||
1
by brian
clean slate |
24 |
/*
|
25 |
The following extern declarations are ok as these are interface functions
|
|
26 |
required by the string function
|
|
27 |
*/
|
|
28 |
||
481
by Brian Aker
Remove all of uchar. |
29 |
extern unsigned char* sql_alloc(unsigned size); |
1
by brian
clean slate |
30 |
extern void sql_element_free(void *ptr); |
31 |
||
32 |
#include "sql_string.h" |
|
33 |
||
34 |
/*****************************************************************************
|
|
35 |
** String functions
|
|
36 |
*****************************************************************************/
|
|
37 |
||
205
by Brian Aker
uint32 -> uin32_t |
38 |
bool String::real_alloc(uint32_t arg_length) |
1
by brian
clean slate |
39 |
{
|
40 |
arg_length=ALIGN_SIZE(arg_length+1); |
|
41 |
str_length=0; |
|
42 |
if (Alloced_length < arg_length) |
|
43 |
{
|
|
44 |
free(); |
|
641.3.6
by Monty Taylor
Removed some my_malloc calls. |
45 |
if (!(Ptr=(char*) malloc(arg_length))) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
46 |
return true; |
1
by brian
clean slate |
47 |
Alloced_length=arg_length; |
48 |
alloced=1; |
|
49 |
}
|
|
50 |
Ptr[0]=0; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
51 |
return false; |
1
by brian
clean slate |
52 |
}
|
53 |
||
54 |
||
55 |
/*
|
|
56 |
** Check that string is big enough. Set string[alloc_length] to 0
|
|
57 |
** (for C functions)
|
|
58 |
*/
|
|
59 |
||
205
by Brian Aker
uint32 -> uin32_t |
60 |
bool String::realloc(uint32_t alloc_length) |
1
by brian
clean slate |
61 |
{
|
205
by Brian Aker
uint32 -> uin32_t |
62 |
uint32_t len=ALIGN_SIZE(alloc_length+1); |
1
by brian
clean slate |
63 |
if (Alloced_length < len) |
64 |
{
|
|
65 |
char *new_ptr; |
|
66 |
if (alloced) |
|
67 |
{
|
|
656.1.26
by Monty Taylor
Finally removed all of the my_malloc stuff. |
68 |
if ((new_ptr= (char*) ::realloc(Ptr,len))) |
1
by brian
clean slate |
69 |
{
|
70 |
Ptr=new_ptr; |
|
71 |
Alloced_length=len; |
|
72 |
}
|
|
73 |
else
|
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
74 |
return true; // Signal error |
1
by brian
clean slate |
75 |
}
|
641.3.6
by Monty Taylor
Removed some my_malloc calls. |
76 |
else if ((new_ptr= (char*) malloc(len))) |
1
by brian
clean slate |
77 |
{
|
78 |
if (str_length) // Avoid bugs in memcpy on AIX |
|
79 |
memcpy(new_ptr,Ptr,str_length); |
|
80 |
new_ptr[str_length]=0; |
|
81 |
Ptr=new_ptr; |
|
82 |
Alloced_length=len; |
|
83 |
alloced=1; |
|
84 |
}
|
|
85 |
else
|
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
86 |
return true; // Signal error |
1
by brian
clean slate |
87 |
}
|
88 |
Ptr[alloc_length]=0; // This make other funcs shorter |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
89 |
return false; |
1
by brian
clean slate |
90 |
}
|
91 |
||
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
92 |
bool String::set_int(int64_t num, bool unsigned_flag, const CHARSET_INFO * const cs) |
1
by brian
clean slate |
93 |
{
|
482
by Brian Aker
Remove uint. |
94 |
uint32_t l=20*cs->mbmaxlen+1; |
1
by brian
clean slate |
95 |
int base= unsigned_flag ? 10 : -10; |
96 |
||
97 |
if (alloc(l)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
98 |
return true; |
205
by Brian Aker
uint32 -> uin32_t |
99 |
str_length=(uint32_t) (cs->cset->int64_t10_to_str)(cs,Ptr,l,base,num); |
1
by brian
clean slate |
100 |
str_charset=cs; |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
101 |
return false; |
1
by brian
clean slate |
102 |
}
|
103 |
||
482
by Brian Aker
Remove uint. |
104 |
bool String::set_real(double num,uint32_t decimals, const CHARSET_INFO * const cs) |
1
by brian
clean slate |
105 |
{
|
106 |
char buff[FLOATING_POINT_BUFFER]; |
|
482
by Brian Aker
Remove uint. |
107 |
uint32_t dummy_errors; |
1
by brian
clean slate |
108 |
size_t len; |
109 |
||
110 |
str_charset=cs; |
|
111 |
if (decimals >= NOT_FIXED_DEC) |
|
112 |
{
|
|
113 |
len= my_gcvt(num, MY_GCVT_ARG_DOUBLE, sizeof(buff) - 1, buff, NULL); |
|
383.1.12
by Brian Aker
Much closer toward UTF8 being around all the time... |
114 |
return copy(buff, len, &my_charset_utf8_general_ci, cs, &dummy_errors); |
1
by brian
clean slate |
115 |
}
|
116 |
len= my_fcvt(num, decimals, buff, NULL); |
|
383.1.12
by Brian Aker
Much closer toward UTF8 being around all the time... |
117 |
return copy(buff, (uint32_t) len, &my_charset_utf8_general_ci, cs, |
1
by brian
clean slate |
118 |
&dummy_errors); |
119 |
}
|
|
120 |
||
121 |
||
122 |
bool String::copy() |
|
123 |
{
|
|
124 |
if (!alloced) |
|
125 |
{
|
|
126 |
Alloced_length=0; // Force realloc |
|
127 |
return realloc(str_length); |
|
128 |
}
|
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
129 |
return false; |
1
by brian
clean slate |
130 |
}
|
131 |
||
132 |
bool String::copy(const String &str) |
|
133 |
{
|
|
134 |
if (alloc(str.str_length)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
135 |
return true; |
1
by brian
clean slate |
136 |
str_length=str.str_length; |
212.6.3
by Mats Kindahl
Removing deprecated functions from code and replacing them with C99 equivalents: |
137 |
memmove(Ptr, str.Ptr, str_length); // May be overlapping |
1
by brian
clean slate |
138 |
Ptr[str_length]=0; |
139 |
str_charset=str.str_charset; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
140 |
return false; |
1
by brian
clean slate |
141 |
}
|
142 |
||
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
143 |
bool String::copy(const char *str,uint32_t arg_length, const CHARSET_INFO * const cs) |
1
by brian
clean slate |
144 |
{
|
145 |
if (alloc(arg_length)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
146 |
return true; |
1
by brian
clean slate |
147 |
if ((str_length=arg_length)) |
148 |
memcpy(Ptr,str,arg_length); |
|
149 |
Ptr[arg_length]=0; |
|
150 |
str_charset=cs; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
151 |
return false; |
1
by brian
clean slate |
152 |
}
|
153 |
||
154 |
||
155 |
/*
|
|
156 |
Checks that the source string can be just copied to the destination string
|
|
157 |
without conversion.
|
|
158 |
||
159 |
SYNPOSIS
|
|
160 |
||
161 |
needs_conversion()
|
|
162 |
arg_length Length of string to copy.
|
|
163 |
from_cs Character set to copy from
|
|
164 |
to_cs Character set to copy to
|
|
205
by Brian Aker
uint32 -> uin32_t |
165 |
uint32_t *offset Returns number of unaligned characters.
|
1
by brian
clean slate |
166 |
|
167 |
RETURN
|
|
168 |
0 No conversion needed
|
|
169 |
1 Either character set conversion or adding leading zeros
|
|
170 |
(e.g. for UCS-2) must be done
|
|
171 |
||
172 |
NOTE
|
|
173 |
to_cs may be NULL for "no conversion" if the system variable
|
|
174 |
character_set_results is NULL.
|
|
175 |
*/
|
|
176 |
||
205
by Brian Aker
uint32 -> uin32_t |
177 |
bool String::needs_conversion(uint32_t arg_length, |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
178 |
const CHARSET_INFO * const from_cs, |
179 |
const CHARSET_INFO * const to_cs, |
|
205
by Brian Aker
uint32 -> uin32_t |
180 |
uint32_t *offset) |
1
by brian
clean slate |
181 |
{
|
182 |
*offset= 0; |
|
183 |
if (!to_cs || |
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
184 |
(to_cs == &my_charset_bin) || |
1
by brian
clean slate |
185 |
(to_cs == from_cs) || |
186 |
my_charset_same(from_cs, to_cs) || |
|
187 |
((from_cs == &my_charset_bin) && |
|
188 |
(!(*offset=(arg_length % to_cs->mbminlen))))) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
189 |
return false; |
190 |
return true; |
|
1
by brian
clean slate |
191 |
}
|
192 |
||
193 |
||
194 |
||
195 |
||
205
by Brian Aker
uint32 -> uin32_t |
196 |
bool String::set_or_copy_aligned(const char *str,uint32_t arg_length, |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
197 |
const CHARSET_INFO * const cs) |
1
by brian
clean slate |
198 |
{
|
199 |
/* How many bytes are in incomplete character */
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
200 |
uint32_t offset= (arg_length % cs->mbminlen); |
201 |
||
910.1.5
by Brian Aker
Remove some dead bits of string (and fix the semi_join test). |
202 |
assert(!offset); /* All characters are complete, just copy */ |
203 |
||
204 |
set(str, arg_length, cs); |
|
205 |
return false; |
|
1
by brian
clean slate |
206 |
}
|
207 |
||
208 |
/* Copy with charset conversion */
|
|
209 |
||
205
by Brian Aker
uint32 -> uin32_t |
210 |
bool String::copy(const char *str, uint32_t arg_length, |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
211 |
const CHARSET_INFO * const from_cs, |
482
by Brian Aker
Remove uint. |
212 |
const CHARSET_INFO * const to_cs, uint32_t *errors) |
1
by brian
clean slate |
213 |
{
|
205
by Brian Aker
uint32 -> uin32_t |
214 |
uint32_t offset; |
1
by brian
clean slate |
215 |
if (!needs_conversion(arg_length, from_cs, to_cs, &offset)) |
216 |
{
|
|
217 |
*errors= 0; |
|
218 |
return copy(str, arg_length, to_cs); |
|
219 |
}
|
|
220 |
if ((from_cs == &my_charset_bin) && offset) |
|
221 |
{
|
|
222 |
*errors= 0; |
|
910.1.5
by Brian Aker
Remove some dead bits of string (and fix the semi_join test). |
223 |
assert((from_cs == &my_charset_bin) && offset); |
224 |
return false; //copy_aligned(str, arg_length, offset, to_cs); |
|
1
by brian
clean slate |
225 |
}
|
205
by Brian Aker
uint32 -> uin32_t |
226 |
uint32_t new_length= to_cs->mbmaxlen*arg_length; |
1
by brian
clean slate |
227 |
if (alloc(new_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
228 |
return true; |
1
by brian
clean slate |
229 |
str_length=copy_and_convert((char*) Ptr, new_length, to_cs, |
230 |
str, arg_length, from_cs, errors); |
|
231 |
str_charset=to_cs; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
232 |
return false; |
1
by brian
clean slate |
233 |
}
|
234 |
||
235 |
||
236 |
/*
|
|
237 |
Set a string to the value of a latin1-string, keeping the original charset
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
238 |
|
1
by brian
clean slate |
239 |
SYNOPSIS
|
240 |
copy_or_set()
|
|
241 |
str String of a simple charset (latin1)
|
|
242 |
arg_length Length of string
|
|
243 |
||
244 |
IMPLEMENTATION
|
|
245 |
If string object is of a simple character set, set it to point to the
|
|
246 |
given string.
|
|
247 |
If not, make a copy and convert it to the new character set.
|
|
248 |
||
249 |
RETURN
|
|
250 |
0 ok
|
|
251 |
1 Could not allocate result buffer
|
|
252 |
||
253 |
*/
|
|
254 |
||
205
by Brian Aker
uint32 -> uin32_t |
255 |
bool String::set_ascii(const char *str, uint32_t arg_length) |
1
by brian
clean slate |
256 |
{
|
257 |
if (str_charset->mbminlen == 1) |
|
258 |
{
|
|
259 |
set(str, arg_length, str_charset); |
|
260 |
return 0; |
|
261 |
}
|
|
482
by Brian Aker
Remove uint. |
262 |
uint32_t dummy_errors; |
383.1.12
by Brian Aker
Much closer toward UTF8 being around all the time... |
263 |
return copy(str, arg_length, &my_charset_utf8_general_ci, str_charset, &dummy_errors); |
1
by brian
clean slate |
264 |
}
|
265 |
||
266 |
bool String::append(const String &s) |
|
267 |
{
|
|
268 |
if (s.length()) |
|
269 |
{
|
|
270 |
if (realloc(str_length+s.length())) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
271 |
return true; |
1
by brian
clean slate |
272 |
memcpy(Ptr+str_length,s.ptr(),s.length()); |
273 |
str_length+=s.length(); |
|
274 |
}
|
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
275 |
return false; |
1
by brian
clean slate |
276 |
}
|
277 |
||
278 |
||
279 |
/*
|
|
280 |
Append an ASCII string to the a string of the current character set
|
|
281 |
*/
|
|
282 |
||
205
by Brian Aker
uint32 -> uin32_t |
283 |
bool String::append(const char *s,uint32_t arg_length) |
1
by brian
clean slate |
284 |
{
|
285 |
if (!arg_length) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
286 |
return false; |
1
by brian
clean slate |
287 |
|
288 |
/*
|
|
289 |
For an ASCII incompatible string, e.g. UCS-2, we need to convert
|
|
290 |
*/
|
|
291 |
if (str_charset->mbminlen > 1) |
|
292 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
293 |
uint32_t add_length=arg_length * str_charset->mbmaxlen; |
482
by Brian Aker
Remove uint. |
294 |
uint32_t dummy_errors; |
1
by brian
clean slate |
295 |
if (realloc(str_length+ add_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
296 |
return true; |
1
by brian
clean slate |
297 |
str_length+= copy_and_convert(Ptr+str_length, add_length, str_charset, |
383.1.12
by Brian Aker
Much closer toward UTF8 being around all the time... |
298 |
s, arg_length, &my_charset_utf8_general_ci, |
1
by brian
clean slate |
299 |
&dummy_errors); |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
300 |
return false; |
1
by brian
clean slate |
301 |
}
|
302 |
||
303 |
/*
|
|
304 |
For an ASCII compatinble string we can just append.
|
|
305 |
*/
|
|
306 |
if (realloc(str_length+arg_length)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
307 |
return true; |
1
by brian
clean slate |
308 |
memcpy(Ptr+str_length,s,arg_length); |
309 |
str_length+=arg_length; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
310 |
return false; |
1
by brian
clean slate |
311 |
}
|
312 |
||
313 |
||
314 |
/*
|
|
315 |
Append a 0-terminated ASCII string
|
|
316 |
*/
|
|
317 |
||
318 |
bool String::append(const char *s) |
|
319 |
{
|
|
320 |
return append(s, strlen(s)); |
|
321 |
}
|
|
322 |
||
323 |
||
324 |
/*
|
|
325 |
Append a string in the given charset to the string
|
|
326 |
with character set recoding
|
|
327 |
*/
|
|
328 |
||
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
329 |
bool String::append(const char *s,uint32_t arg_length, const CHARSET_INFO * const cs) |
1
by brian
clean slate |
330 |
{
|
205
by Brian Aker
uint32 -> uin32_t |
331 |
uint32_t dummy_offset; |
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
332 |
|
1
by brian
clean slate |
333 |
if (needs_conversion(arg_length, cs, str_charset, &dummy_offset)) |
334 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
335 |
uint32_t add_length= arg_length / cs->mbminlen * str_charset->mbmaxlen; |
482
by Brian Aker
Remove uint. |
336 |
uint32_t dummy_errors; |
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
337 |
if (realloc(str_length + add_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
338 |
return true; |
1
by brian
clean slate |
339 |
str_length+= copy_and_convert(Ptr+str_length, add_length, str_charset, |
340 |
s, arg_length, cs, &dummy_errors); |
|
341 |
}
|
|
342 |
else
|
|
343 |
{
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
344 |
if (realloc(str_length + arg_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
345 |
return true; |
1
by brian
clean slate |
346 |
memcpy(Ptr + str_length, s, arg_length); |
347 |
str_length+= arg_length; |
|
348 |
}
|
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
349 |
return false; |
1
by brian
clean slate |
350 |
}
|
351 |
||
352 |
||
205
by Brian Aker
uint32 -> uin32_t |
353 |
bool String::append_with_prefill(const char *s,uint32_t arg_length, |
354 |
uint32_t full_length, char fill_char) |
|
1
by brian
clean slate |
355 |
{
|
356 |
int t_length= arg_length > full_length ? arg_length : full_length; |
|
357 |
||
358 |
if (realloc(str_length + t_length)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
359 |
return true; |
1
by brian
clean slate |
360 |
t_length= full_length - arg_length; |
361 |
if (t_length > 0) |
|
362 |
{
|
|
212.6.3
by Mats Kindahl
Removing deprecated functions from code and replacing them with C99 equivalents: |
363 |
memset(Ptr+str_length, fill_char, t_length); |
1
by brian
clean slate |
364 |
str_length=str_length + t_length; |
365 |
}
|
|
366 |
append(s, arg_length); |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
367 |
return false; |
1
by brian
clean slate |
368 |
}
|
369 |
||
205
by Brian Aker
uint32 -> uin32_t |
370 |
uint32_t String::numchars() |
1
by brian
clean slate |
371 |
{
|
372 |
return str_charset->cset->numchars(str_charset, Ptr, Ptr+str_length); |
|
373 |
}
|
|
374 |
||
205
by Brian Aker
uint32 -> uin32_t |
375 |
int String::charpos(int i,uint32_t offset) |
1
by brian
clean slate |
376 |
{
|
377 |
if (i <= 0) |
|
378 |
return i; |
|
379 |
return str_charset->cset->charpos(str_charset,Ptr+offset,Ptr+str_length,i); |
|
380 |
}
|
|
381 |
||
205
by Brian Aker
uint32 -> uin32_t |
382 |
int String::strstr(const String &s,uint32_t offset) |
1
by brian
clean slate |
383 |
{
|
384 |
if (s.length()+offset <= str_length) |
|
385 |
{
|
|
386 |
if (!s.length()) |
|
387 |
return ((int) offset); // Empty string is always found |
|
388 |
||
389 |
register const char *str = Ptr+offset; |
|
390 |
register const char *search=s.ptr(); |
|
391 |
const char *end=Ptr+str_length-s.length()+1; |
|
392 |
const char *search_end=s.ptr()+s.length(); |
|
393 |
skip: |
|
394 |
while (str != end) |
|
395 |
{
|
|
396 |
if (*str++ == *search) |
|
397 |
{
|
|
398 |
register char *i,*j; |
|
399 |
i=(char*) str; j=(char*) search+1; |
|
400 |
while (j != search_end) |
|
401 |
if (*i++ != *j++) goto skip; |
|
402 |
return (int) (str-Ptr) -1; |
|
403 |
}
|
|
404 |
}
|
|
405 |
}
|
|
406 |
return -1; |
|
407 |
}
|
|
408 |
||
409 |
/*
|
|
410 |
** Search string from end. Offset is offset to the end of string
|
|
411 |
*/
|
|
412 |
||
205
by Brian Aker
uint32 -> uin32_t |
413 |
int String::strrstr(const String &s,uint32_t offset) |
1
by brian
clean slate |
414 |
{
|
415 |
if (s.length() <= offset && offset <= str_length) |
|
416 |
{
|
|
417 |
if (!s.length()) |
|
418 |
return offset; // Empty string is always found |
|
419 |
register const char *str = Ptr+offset-1; |
|
420 |
register const char *search=s.ptr()+s.length()-1; |
|
421 |
||
422 |
const char *end=Ptr+s.length()-2; |
|
423 |
const char *search_end=s.ptr()-1; |
|
424 |
skip: |
|
425 |
while (str != end) |
|
426 |
{
|
|
427 |
if (*str-- == *search) |
|
428 |
{
|
|
429 |
register char *i,*j; |
|
430 |
i=(char*) str; j=(char*) search-1; |
|
431 |
while (j != search_end) |
|
432 |
if (*i-- != *j--) goto skip; |
|
433 |
return (int) (i-Ptr) +1; |
|
434 |
}
|
|
435 |
}
|
|
436 |
}
|
|
437 |
return -1; |
|
438 |
}
|
|
439 |
||
440 |
/*
|
|
441 |
Replace substring with string
|
|
442 |
If wrong parameter or not enough memory, do nothing
|
|
443 |
*/
|
|
444 |
||
205
by Brian Aker
uint32 -> uin32_t |
445 |
bool String::replace(uint32_t offset,uint32_t arg_length,const String &to) |
1
by brian
clean slate |
446 |
{
|
447 |
return replace(offset,arg_length,to.ptr(),to.length()); |
|
448 |
}
|
|
449 |
||
205
by Brian Aker
uint32 -> uin32_t |
450 |
bool String::replace(uint32_t offset,uint32_t arg_length, |
451 |
const char *to, uint32_t to_length) |
|
1
by brian
clean slate |
452 |
{
|
453 |
long diff = (long) to_length-(long) arg_length; |
|
454 |
if (offset+arg_length <= str_length) |
|
455 |
{
|
|
456 |
if (diff < 0) |
|
457 |
{
|
|
458 |
if (to_length) |
|
459 |
memcpy(Ptr+offset,to,to_length); |
|
629.3.4
by Kristian Nielsen
Take Mats'es changes from bmove()->memcpy(), and fix all of them to be |
460 |
memmove(Ptr+offset+to_length, Ptr+offset+arg_length, |
461 |
str_length-offset-arg_length); |
|
1
by brian
clean slate |
462 |
}
|
463 |
else
|
|
464 |
{
|
|
465 |
if (diff) |
|
466 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
467 |
if (realloc(str_length+(uint32_t) diff)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
468 |
return true; |
481
by Brian Aker
Remove all of uchar. |
469 |
bmove_upp((unsigned char*) Ptr+str_length+diff, (unsigned char*) Ptr+str_length, |
1
by brian
clean slate |
470 |
str_length-offset-arg_length); |
471 |
}
|
|
472 |
if (to_length) |
|
473 |
memcpy(Ptr+offset,to,to_length); |
|
474 |
}
|
|
205
by Brian Aker
uint32 -> uin32_t |
475 |
str_length+=(uint32_t) diff; |
1
by brian
clean slate |
476 |
}
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
477 |
return false; |
1
by brian
clean slate |
478 |
}
|
479 |
||
480 |
||
481 |
||
482 |
/*
|
|
483 |
Compare strings according to collation, without end space.
|
|
484 |
||
485 |
SYNOPSIS
|
|
486 |
sortcmp()
|
|
487 |
s First string
|
|
488 |
t Second string
|
|
489 |
cs Collation
|
|
490 |
||
491 |
NOTE:
|
|
492 |
Normally this is case sensitive comparison
|
|
493 |
||
494 |
RETURN
|
|
495 |
< 0 s < t
|
|
496 |
0 s == t
|
|
497 |
> 0 s > t
|
|
498 |
*/
|
|
499 |
||
500 |
||
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
501 |
int sortcmp(const String *s,const String *t, const CHARSET_INFO * const cs) |
1
by brian
clean slate |
502 |
{
|
503 |
return cs->coll->strnncollsp(cs, |
|
481
by Brian Aker
Remove all of uchar. |
504 |
(unsigned char *) s->ptr(),s->length(), |
505 |
(unsigned char *) t->ptr(),t->length(), 0); |
|
1
by brian
clean slate |
506 |
}
|
507 |
||
508 |
||
509 |
/*
|
|
510 |
Compare strings byte by byte. End spaces are also compared.
|
|
511 |
||
512 |
SYNOPSIS
|
|
513 |
stringcmp()
|
|
514 |
s First string
|
|
515 |
t Second string
|
|
516 |
||
517 |
NOTE:
|
|
481
by Brian Aker
Remove all of uchar. |
518 |
Strings are compared as a stream of unsigned chars
|
1
by brian
clean slate |
519 |
|
520 |
RETURN
|
|
521 |
< 0 s < t
|
|
522 |
0 s == t
|
|
523 |
> 0 s > t
|
|
524 |
*/
|
|
525 |
||
526 |
||
527 |
int stringcmp(const String *s,const String *t) |
|
528 |
{
|
|
398.1.4
by Monty Taylor
Renamed max/min. |
529 |
uint32_t s_len=s->length(),t_len=t->length(),len=cmin(s_len,t_len); |
1
by brian
clean slate |
530 |
int cmp= memcmp(s->ptr(), t->ptr(), len); |
531 |
return (cmp) ? cmp : (int) (s_len - t_len); |
|
532 |
}
|
|
533 |
||
534 |
||
205
by Brian Aker
uint32 -> uin32_t |
535 |
String *copy_if_not_alloced(String *to,String *from,uint32_t from_length) |
1
by brian
clean slate |
536 |
{
|
537 |
if (from->Alloced_length >= from_length) |
|
538 |
return from; |
|
539 |
if (from->alloced || !to || from == to) |
|
540 |
{
|
|
541 |
(void) from->realloc(from_length); |
|
542 |
return from; |
|
543 |
}
|
|
544 |
if (to->realloc(from_length)) |
|
545 |
return from; // Actually an error |
|
398.1.4
by Monty Taylor
Renamed max/min. |
546 |
if ((to->str_length=cmin(from->str_length,from_length))) |
1
by brian
clean slate |
547 |
memcpy(to->Ptr,from->Ptr,to->str_length); |
548 |
to->str_charset=from->str_charset; |
|
549 |
return to; |
|
550 |
}
|
|
551 |
||
552 |
||
553 |
/****************************************************************************
|
|
554 |
Help functions
|
|
555 |
****************************************************************************/
|
|
556 |
||
557 |
/*
|
|
558 |
copy a string from one character set to another
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
559 |
|
1
by brian
clean slate |
560 |
SYNOPSIS
|
561 |
copy_and_convert()
|
|
562 |
to Store result here
|
|
563 |
to_cs Character set of result string
|
|
564 |
from Copy from here
|
|
565 |
from_length Length of from string
|
|
566 |
from_cs From character set
|
|
567 |
||
568 |
NOTES
|
|
569 |
'to' must be big enough as form_length * to_cs->mbmaxlen
|
|
570 |
||
571 |
RETURN
|
|
572 |
length of bytes copied to 'to'
|
|
573 |
*/
|
|
574 |
||
575 |
||
205
by Brian Aker
uint32 -> uin32_t |
576 |
static uint32_t |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
577 |
copy_and_convert_extended(char *to, uint32_t to_length, |
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
578 |
const CHARSET_INFO * const to_cs, |
205
by Brian Aker
uint32 -> uin32_t |
579 |
const char *from, uint32_t from_length, |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
580 |
const CHARSET_INFO * const from_cs, |
482
by Brian Aker
Remove uint. |
581 |
uint32_t *errors) |
1
by brian
clean slate |
582 |
{
|
583 |
int cnvres; |
|
584 |
my_wc_t wc; |
|
481
by Brian Aker
Remove all of uchar. |
585 |
const unsigned char *from_end= (const unsigned char*) from+from_length; |
1
by brian
clean slate |
586 |
char *to_start= to; |
481
by Brian Aker
Remove all of uchar. |
587 |
unsigned char *to_end= (unsigned char*) to+to_length; |
1
by brian
clean slate |
588 |
my_charset_conv_mb_wc mb_wc= from_cs->cset->mb_wc; |
589 |
my_charset_conv_wc_mb wc_mb= to_cs->cset->wc_mb; |
|
482
by Brian Aker
Remove uint. |
590 |
uint32_t error_count= 0; |
1
by brian
clean slate |
591 |
|
592 |
while (1) |
|
593 |
{
|
|
481
by Brian Aker
Remove all of uchar. |
594 |
if ((cnvres= (*mb_wc)(from_cs, &wc, (unsigned char*) from, |
1
by brian
clean slate |
595 |
from_end)) > 0) |
596 |
from+= cnvres; |
|
597 |
else if (cnvres == MY_CS_ILSEQ) |
|
598 |
{
|
|
599 |
error_count++; |
|
600 |
from++; |
|
601 |
wc= '?'; |
|
602 |
}
|
|
603 |
else if (cnvres > MY_CS_TOOSMALL) |
|
604 |
{
|
|
605 |
/*
|
|
606 |
A correct multibyte sequence detected
|
|
607 |
But it doesn't have Unicode mapping.
|
|
608 |
*/
|
|
609 |
error_count++; |
|
610 |
from+= (-cnvres); |
|
611 |
wc= '?'; |
|
612 |
}
|
|
613 |
else
|
|
614 |
break; // Not enough characters |
|
615 |
||
616 |
outp: |
|
481
by Brian Aker
Remove all of uchar. |
617 |
if ((cnvres= (*wc_mb)(to_cs, wc, (unsigned char*) to, to_end)) > 0) |
1
by brian
clean slate |
618 |
to+= cnvres; |
619 |
else if (cnvres == MY_CS_ILUNI && wc != '?') |
|
620 |
{
|
|
621 |
error_count++; |
|
622 |
wc= '?'; |
|
623 |
goto outp; |
|
624 |
}
|
|
625 |
else
|
|
626 |
break; |
|
627 |
}
|
|
628 |
*errors= error_count; |
|
205
by Brian Aker
uint32 -> uin32_t |
629 |
return (uint32_t) (to - to_start); |
1
by brian
clean slate |
630 |
}
|
631 |
||
632 |
||
633 |
/*
|
|
634 |
Optimized for quick copying of ASCII characters in the range 0x00..0x7F.
|
|
635 |
*/
|
|
205
by Brian Aker
uint32 -> uin32_t |
636 |
uint32_t
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
637 |
copy_and_convert(char *to, uint32_t to_length, const CHARSET_INFO * const to_cs, |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
638 |
const char *from, uint32_t from_length, |
482
by Brian Aker
Remove uint. |
639 |
const CHARSET_INFO * const from_cs, uint32_t *errors) |
1
by brian
clean slate |
640 |
{
|
641 |
/*
|
|
642 |
If any of the character sets is not ASCII compatible,
|
|
643 |
immediately switch to slow mb_wc->wc_mb method.
|
|
644 |
*/
|
|
645 |
if ((to_cs->state | from_cs->state) & MY_CS_NONASCII) |
|
646 |
return copy_and_convert_extended(to, to_length, to_cs, |
|
647 |
from, from_length, from_cs, errors); |
|
648 |
||
398.1.4
by Monty Taylor
Renamed max/min. |
649 |
uint32_t length= cmin(to_length, from_length), length2= length; |
1
by brian
clean slate |
650 |
|
651 |
#if defined(__i386__)
|
|
652 |
/*
|
|
653 |
Special loop for i386, it allows to refer to a
|
|
654 |
non-aligned memory block as UINT32, which makes
|
|
655 |
it possible to copy four bytes at once. This
|
|
656 |
gives about 10% performance improvement comparing
|
|
657 |
to byte-by-byte loop.
|
|
658 |
*/
|
|
659 |
for ( ; length >= 4; length-= 4, from+= 4, to+= 4) |
|
660 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
661 |
if ((*(uint32_t*)from) & 0x80808080) |
1
by brian
clean slate |
662 |
break; |
205
by Brian Aker
uint32 -> uin32_t |
663 |
*((uint32_t*) to)= *((const uint32_t*) from); |
1
by brian
clean slate |
664 |
}
|
665 |
#endif
|
|
666 |
||
667 |
for (; ; *to++= *from++, length--) |
|
668 |
{
|
|
669 |
if (!length) |
|
670 |
{
|
|
671 |
*errors= 0; |
|
672 |
return length2; |
|
673 |
}
|
|
674 |
if (*((unsigned char*) from) > 0x7F) /* A non-ASCII character */ |
|
675 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
676 |
uint32_t copied_length= length2 - length; |
1
by brian
clean slate |
677 |
to_length-= copied_length; |
678 |
from_length-= copied_length; |
|
679 |
return copied_length + copy_and_convert_extended(to, to_length, |
|
680 |
to_cs, |
|
681 |
from, from_length, |
|
682 |
from_cs, |
|
683 |
errors); |
|
684 |
}
|
|
685 |
}
|
|
686 |
||
892.2.4
by Monty Taylor
Fixed more warnings. |
687 |
#ifndef __sun
|
1
by brian
clean slate |
688 |
return 0; // Make compiler happy |
892.2.4
by Monty Taylor
Fixed more warnings. |
689 |
#endif
|
1
by brian
clean slate |
690 |
}
|
691 |
||
692 |
||
693 |
/**
|
|
694 |
Copy string with HEX-encoding of "bad" characters.
|
|
695 |
||
696 |
@details This functions copies the string pointed by "src"
|
|
697 |
to the string pointed by "dst". Not more than "srclen" bytes
|
|
698 |
are read from "src". Any sequences of bytes representing
|
|
699 |
a not-well-formed substring (according to cs) are hex-encoded,
|
|
700 |
and all well-formed substrings (according to cs) are copied as is.
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
701 |
Not more than "dstlen" bytes are written to "dst". The number
|
1
by brian
clean slate |
702 |
of bytes written to "dst" is returned.
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
703 |
|
1
by brian
clean slate |
704 |
@param cs character set pointer of the destination string
|
705 |
@param[out] dst destination string
|
|
706 |
@param dstlen size of dst
|
|
707 |
@param src source string
|
|
708 |
@param srclen length of src
|
|
709 |
||
710 |
@retval result length
|
|
711 |
*/
|
|
712 |
||
713 |
size_t
|
|
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
714 |
my_copy_with_hex_escaping(const CHARSET_INFO * const cs, |
1
by brian
clean slate |
715 |
char *dst, size_t dstlen, |
716 |
const char *src, size_t srclen) |
|
717 |
{
|
|
718 |
const char *srcend= src + srclen; |
|
719 |
char *dst0= dst; |
|
720 |
||
721 |
for ( ; src < srcend ; ) |
|
722 |
{
|
|
723 |
size_t chlen; |
|
724 |
if ((chlen= my_ismbchar(cs, src, srcend))) |
|
725 |
{
|
|
726 |
if (dstlen < chlen) |
|
727 |
break; /* purecov: inspected */ |
|
728 |
memcpy(dst, src, chlen); |
|
729 |
src+= chlen; |
|
730 |
dst+= chlen; |
|
731 |
dstlen-= chlen; |
|
732 |
}
|
|
733 |
else if (*src & 0x80) |
|
734 |
{
|
|
735 |
if (dstlen < 4) |
|
736 |
break; /* purecov: inspected */ |
|
737 |
*dst++= '\\'; |
|
738 |
*dst++= 'x'; |
|
739 |
*dst++= _dig_vec_upper[((unsigned char) *src) >> 4]; |
|
740 |
*dst++= _dig_vec_upper[((unsigned char) *src) & 15]; |
|
741 |
src++; |
|
742 |
dstlen-= 4; |
|
743 |
}
|
|
744 |
else
|
|
745 |
{
|
|
746 |
if (dstlen < 1) |
|
747 |
break; /* purecov: inspected */ |
|
748 |
*dst++= *src++; |
|
749 |
dstlen--; |
|
750 |
}
|
|
751 |
}
|
|
752 |
return dst - dst0; |
|
753 |
}
|
|
754 |
||
755 |
/*
|
|
756 |
copy a string,
|
|
757 |
with optional character set conversion,
|
|
758 |
with optional left padding (for binary -> UCS2 conversion)
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
759 |
|
1
by brian
clean slate |
760 |
SYNOPSIS
|
761 |
well_formed_copy_nchars()
|
|
762 |
to Store result here
|
|
763 |
to_length Maxinum length of "to" string
|
|
764 |
to_cs Character set of "to" string
|
|
765 |
from Copy from here
|
|
766 |
from_length Length of from string
|
|
767 |
from_cs From character set
|
|
768 |
nchars Copy not more that nchars characters
|
|
769 |
well_formed_error_pos Return position when "from" is not well formed
|
|
770 |
or NULL otherwise.
|
|
771 |
cannot_convert_error_pos Return position where a not convertable
|
|
772 |
character met, or NULL otherwise.
|
|
773 |
from_end_pos Return position where scanning of "from"
|
|
774 |
string stopped.
|
|
775 |
NOTES
|
|
776 |
||
777 |
RETURN
|
|
778 |
length of bytes copied to 'to'
|
|
779 |
*/
|
|
780 |
||
781 |
||
205
by Brian Aker
uint32 -> uin32_t |
782 |
uint32_t
|
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
783 |
well_formed_copy_nchars(const CHARSET_INFO * const to_cs, |
482
by Brian Aker
Remove uint. |
784 |
char *to, uint32_t to_length, |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
785 |
const CHARSET_INFO * const from_cs, |
482
by Brian Aker
Remove uint. |
786 |
const char *from, uint32_t from_length, |
787 |
uint32_t nchars, |
|
1
by brian
clean slate |
788 |
const char **well_formed_error_pos, |
789 |
const char **cannot_convert_error_pos, |
|
790 |
const char **from_end_pos) |
|
791 |
{
|
|
482
by Brian Aker
Remove uint. |
792 |
uint32_t res; |
1
by brian
clean slate |
793 |
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
794 |
if ((to_cs == &my_charset_bin) || |
1
by brian
clean slate |
795 |
(from_cs == &my_charset_bin) || |
796 |
(to_cs == from_cs) || |
|
797 |
my_charset_same(from_cs, to_cs)) |
|
798 |
{
|
|
799 |
if (to_length < to_cs->mbminlen || !nchars) |
|
800 |
{
|
|
801 |
*from_end_pos= from; |
|
802 |
*cannot_convert_error_pos= NULL; |
|
803 |
*well_formed_error_pos= NULL; |
|
804 |
return 0; |
|
805 |
}
|
|
806 |
||
807 |
if (to_cs == &my_charset_bin) |
|
808 |
{
|
|
398.1.4
by Monty Taylor
Renamed max/min. |
809 |
res= cmin(cmin(nchars, to_length), from_length); |
1
by brian
clean slate |
810 |
memmove(to, from, res); |
811 |
*from_end_pos= from + res; |
|
812 |
*well_formed_error_pos= NULL; |
|
813 |
*cannot_convert_error_pos= NULL; |
|
814 |
}
|
|
815 |
else
|
|
816 |
{
|
|
817 |
int well_formed_error; |
|
482
by Brian Aker
Remove uint. |
818 |
uint32_t from_offset; |
1
by brian
clean slate |
819 |
|
820 |
if ((from_offset= (from_length % to_cs->mbminlen)) && |
|
821 |
(from_cs == &my_charset_bin)) |
|
822 |
{
|
|
823 |
/*
|
|
824 |
Copying from BINARY to UCS2 needs to prepend zeros sometimes:
|
|
825 |
INSERT INTO t1 (ucs2_column) VALUES (0x01);
|
|
826 |
0x01 -> 0x0001
|
|
827 |
*/
|
|
482
by Brian Aker
Remove uint. |
828 |
uint32_t pad_length= to_cs->mbminlen - from_offset; |
212.6.1
by Mats Kindahl
Replacing all bzero() calls with memset() calls and removing the bzero.c file. |
829 |
memset(to, 0, pad_length); |
1
by brian
clean slate |
830 |
memmove(to + pad_length, from, from_offset); |
831 |
nchars--; |
|
832 |
from+= from_offset; |
|
833 |
from_length-= from_offset; |
|
834 |
to+= to_cs->mbminlen; |
|
835 |
to_length-= to_cs->mbminlen; |
|
836 |
}
|
|
837 |
||
838 |
set_if_smaller(from_length, to_length); |
|
839 |
res= to_cs->cset->well_formed_len(to_cs, from, from + from_length, |
|
840 |
nchars, &well_formed_error); |
|
841 |
memmove(to, from, res); |
|
842 |
*from_end_pos= from + res; |
|
843 |
*well_formed_error_pos= well_formed_error ? from + res : NULL; |
|
844 |
*cannot_convert_error_pos= NULL; |
|
845 |
if (from_offset) |
|
846 |
res+= to_cs->mbminlen; |
|
847 |
}
|
|
848 |
}
|
|
849 |
else
|
|
850 |
{
|
|
851 |
int cnvres; |
|
852 |
my_wc_t wc; |
|
853 |
my_charset_conv_mb_wc mb_wc= from_cs->cset->mb_wc; |
|
854 |
my_charset_conv_wc_mb wc_mb= to_cs->cset->wc_mb; |
|
481
by Brian Aker
Remove all of uchar. |
855 |
const unsigned char *from_end= (const unsigned char*) from + from_length; |
856 |
unsigned char *to_end= (unsigned char*) to + to_length; |
|
1
by brian
clean slate |
857 |
char *to_start= to; |
858 |
*well_formed_error_pos= NULL; |
|
859 |
*cannot_convert_error_pos= NULL; |
|
860 |
||
861 |
for ( ; nchars; nchars--) |
|
862 |
{
|
|
863 |
const char *from_prev= from; |
|
481
by Brian Aker
Remove all of uchar. |
864 |
if ((cnvres= (*mb_wc)(from_cs, &wc, (unsigned char*) from, from_end)) > 0) |
1
by brian
clean slate |
865 |
from+= cnvres; |
866 |
else if (cnvres == MY_CS_ILSEQ) |
|
867 |
{
|
|
868 |
if (!*well_formed_error_pos) |
|
869 |
*well_formed_error_pos= from; |
|
870 |
from++; |
|
871 |
wc= '?'; |
|
872 |
}
|
|
873 |
else if (cnvres > MY_CS_TOOSMALL) |
|
874 |
{
|
|
875 |
/*
|
|
876 |
A correct multibyte sequence detected
|
|
877 |
But it doesn't have Unicode mapping.
|
|
878 |
*/
|
|
879 |
if (!*cannot_convert_error_pos) |
|
880 |
*cannot_convert_error_pos= from; |
|
881 |
from+= (-cnvres); |
|
882 |
wc= '?'; |
|
883 |
}
|
|
884 |
else
|
|
885 |
break; // Not enough characters |
|
886 |
||
887 |
outp: |
|
481
by Brian Aker
Remove all of uchar. |
888 |
if ((cnvres= (*wc_mb)(to_cs, wc, (unsigned char*) to, to_end)) > 0) |
1
by brian
clean slate |
889 |
to+= cnvres; |
890 |
else if (cnvres == MY_CS_ILUNI && wc != '?') |
|
891 |
{
|
|
892 |
if (!*cannot_convert_error_pos) |
|
893 |
*cannot_convert_error_pos= from_prev; |
|
894 |
wc= '?'; |
|
895 |
goto outp; |
|
896 |
}
|
|
897 |
else
|
|
898 |
{
|
|
899 |
from= from_prev; |
|
900 |
break; |
|
901 |
}
|
|
902 |
}
|
|
903 |
*from_end_pos= from; |
|
904 |
res= to - to_start; |
|
905 |
}
|
|
205
by Brian Aker
uint32 -> uin32_t |
906 |
return (uint32_t) res; |
1
by brian
clean slate |
907 |
}
|
908 |
||
909 |
||
910 |
||
911 |
||
912 |
void String::print(String *str) |
|
913 |
{
|
|
914 |
char *st= (char*)Ptr, *end= st+str_length; |
|
915 |
for (; st < end; st++) |
|
916 |
{
|
|
481
by Brian Aker
Remove all of uchar. |
917 |
unsigned char c= *st; |
1
by brian
clean slate |
918 |
switch (c) |
919 |
{
|
|
920 |
case '\\': |
|
520.4.32
by Monty Taylor
Fixed oops. |
921 |
str->append("\\\\", sizeof("\\\\")-1); |
1
by brian
clean slate |
922 |
break; |
923 |
case '\0': |
|
520.4.32
by Monty Taylor
Fixed oops. |
924 |
str->append("\\0", sizeof("\\0")-1); |
1
by brian
clean slate |
925 |
break; |
926 |
case '\'': |
|
520.4.32
by Monty Taylor
Fixed oops. |
927 |
str->append("\\'", sizeof("\\'")-1); |
1
by brian
clean slate |
928 |
break; |
929 |
case '\n': |
|
520.4.32
by Monty Taylor
Fixed oops. |
930 |
str->append("\\n", sizeof("\\n")-1); |
1
by brian
clean slate |
931 |
break; |
932 |
case '\r': |
|
520.4.32
by Monty Taylor
Fixed oops. |
933 |
str->append("\\r", sizeof("\\r")-1); |
1
by brian
clean slate |
934 |
break; |
935 |
case '\032': // Ctrl-Z |
|
520.4.32
by Monty Taylor
Fixed oops. |
936 |
str->append("\\Z", sizeof("\\Z")-1); |
1
by brian
clean slate |
937 |
break; |
938 |
default: |
|
939 |
str->append(c); |
|
940 |
}
|
|
941 |
}
|
|
942 |
}
|
|
943 |
||
794
by Brian Aker
Refactor append_identifier and remove dead OPTION_QUOTE_SHOW_CREATE option |
944 |
/*
|
945 |
Quote the given identifier.
|
|
946 |
If the given identifier is empty, it will be quoted.
|
|
947 |
||
948 |
SYNOPSIS
|
|
949 |
append_identifier()
|
|
950 |
name the identifier to be appended
|
|
951 |
name_length length of the appending identifier
|
|
952 |
*/
|
|
953 |
||
954 |
/* Factor the extern out */
|
|
955 |
extern const CHARSET_INFO *system_charset_info, *files_charset_info; |
|
956 |
||
779.3.10
by Monty Taylor
Turned on -Wshadow. |
957 |
void String::append_identifier(const char *name, uint32_t in_length) |
794
by Brian Aker
Refactor append_identifier and remove dead OPTION_QUOTE_SHOW_CREATE option |
958 |
{
|
959 |
const char *name_end; |
|
960 |
char quote_char; |
|
961 |
int q= '`'; |
|
962 |
||
963 |
/*
|
|
964 |
The identifier must be quoted as it includes a quote character or
|
|
965 |
it's a keyword
|
|
966 |
*/
|
|
967 |
||
779.3.10
by Monty Taylor
Turned on -Wshadow. |
968 |
reserve(in_length*2 + 2); |
794
by Brian Aker
Refactor append_identifier and remove dead OPTION_QUOTE_SHOW_CREATE option |
969 |
quote_char= (char) q; |
970 |
append("e_char, 1, system_charset_info); |
|
971 |
||
779.3.10
by Monty Taylor
Turned on -Wshadow. |
972 |
for (name_end= name+in_length ; name < name_end ; name+= in_length) |
794
by Brian Aker
Refactor append_identifier and remove dead OPTION_QUOTE_SHOW_CREATE option |
973 |
{
|
974 |
unsigned char chr= (unsigned char) *name; |
|
779.3.10
by Monty Taylor
Turned on -Wshadow. |
975 |
in_length= my_mbcharlen(system_charset_info, chr); |
794
by Brian Aker
Refactor append_identifier and remove dead OPTION_QUOTE_SHOW_CREATE option |
976 |
/*
|
977 |
my_mbcharlen can return 0 on a wrong multibyte
|
|
978 |
sequence. It is possible when upgrading from 4.0,
|
|
979 |
and identifier contains some accented characters.
|
|
980 |
The manual says it does not work. So we'll just
|
|
981 |
change length to 1 not to hang in the endless loop.
|
|
982 |
*/
|
|
779.3.10
by Monty Taylor
Turned on -Wshadow. |
983 |
if (!in_length) |
984 |
in_length= 1; |
|
985 |
if (in_length == 1 && chr == (unsigned char) quote_char) |
|
794
by Brian Aker
Refactor append_identifier and remove dead OPTION_QUOTE_SHOW_CREATE option |
986 |
append("e_char, 1, system_charset_info); |
779.3.10
by Monty Taylor
Turned on -Wshadow. |
987 |
append(name, in_length, system_charset_info); |
794
by Brian Aker
Refactor append_identifier and remove dead OPTION_QUOTE_SHOW_CREATE option |
988 |
}
|
989 |
append("e_char, 1, system_charset_info); |
|
990 |
}
|
|
991 |
||
1
by brian
clean slate |
992 |
|
993 |
/*
|
|
994 |
Exchange state of this object and argument.
|
|
995 |
||
996 |
SYNOPSIS
|
|
997 |
String::swap()
|
|
998 |
||
999 |
RETURN
|
|
1000 |
Target string will contain state of this object and vice versa.
|
|
1001 |
*/
|
|
1002 |
||
1003 |
void String::swap(String &s) |
|
1004 |
{
|
|
322.2.2
by Mats Kindahl
Hiding THD::proc_info field and providing a setter and getter. |
1005 |
std::swap(Ptr, s.Ptr); |
1006 |
std::swap(str_length, s.str_length); |
|
1007 |
std::swap(Alloced_length, s.Alloced_length); |
|
1008 |
std::swap(alloced, s.alloced); |
|
1009 |
std::swap(str_charset, s.str_charset); |
|
1
by brian
clean slate |
1010 |
}
|
598.1.1
by Super-User
Fixed solaris build crap. |
1011 |
|
1012 |
||
1013 |
bool operator==(const String &s1, const String &s2) |
|
1014 |
{
|
|
1015 |
return stringcmp(&s1,&s2) == 0; |
|
1016 |
}
|
|
1017 |
||
1018 |
bool operator!=(const String &s1, const String &s2) |
|
1019 |
{
|
|
1020 |
return !(s1 == s2); |
|
1021 |
}
|
|
1022 |