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 |
Copy a multi-byte character sets with adding leading zeros.
|
|
196 |
||
197 |
SYNOPSIS
|
|
198 |
||
199 |
copy_aligned()
|
|
200 |
str String to copy
|
|
201 |
arg_length Length of string. This should NOT be dividable with
|
|
202 |
cs->mbminlen.
|
|
203 |
offset arg_length % cs->mb_minlength
|
|
204 |
cs Character set for 'str'
|
|
205 |
||
206 |
NOTES
|
|
207 |
For real multi-byte, ascii incompatible charactser sets,
|
|
208 |
like UCS-2, add leading zeros if we have an incomplete character.
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
209 |
Thus,
|
210 |
SELECT _ucs2 0xAA
|
|
1
by brian
clean slate |
211 |
will automatically be converted into
|
212 |
SELECT _ucs2 0x00AA
|
|
213 |
||
214 |
RETURN
|
|
215 |
0 ok
|
|
216 |
1 error
|
|
217 |
*/
|
|
218 |
||
205
by Brian Aker
uint32 -> uin32_t |
219 |
bool String::copy_aligned(const char *str,uint32_t arg_length, uint32_t offset, |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
220 |
const CHARSET_INFO * const cs) |
1
by brian
clean slate |
221 |
{
|
222 |
/* How many bytes are in incomplete character */
|
|
223 |
offset= cs->mbmaxlen - offset; /* How many zeros we should prepend */ |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
224 |
assert(offset && offset != cs->mbmaxlen); |
1
by brian
clean slate |
225 |
|
205
by Brian Aker
uint32 -> uin32_t |
226 |
uint32_t aligned_length= arg_length + offset; |
1
by brian
clean slate |
227 |
if (alloc(aligned_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
228 |
return true; |
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
229 |
|
1
by brian
clean slate |
230 |
/*
|
231 |
Note, this is only safe for big-endian UCS-2.
|
|
232 |
If we add little-endian UCS-2 sometimes, this code
|
|
233 |
will be more complicated. But it's OK for now.
|
|
234 |
*/
|
|
212.6.6
by Mats Kindahl
Removing redundant use of casts in drizzled/ for memcmp(), memcpy(), memset(), and memmove(). |
235 |
memset(Ptr, 0, offset); |
1
by brian
clean slate |
236 |
memcpy(Ptr + offset, str, arg_length); |
237 |
Ptr[aligned_length]=0; |
|
238 |
/* str_length is always >= 0 as arg_length is != 0 */
|
|
239 |
str_length= aligned_length; |
|
240 |
str_charset= cs; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
241 |
return false; |
1
by brian
clean slate |
242 |
}
|
243 |
||
244 |
||
205
by Brian Aker
uint32 -> uin32_t |
245 |
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. |
246 |
const CHARSET_INFO * const cs) |
1
by brian
clean slate |
247 |
{
|
248 |
/* How many bytes are in incomplete character */
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
249 |
uint32_t offset= (arg_length % cs->mbminlen); |
250 |
||
1
by brian
clean slate |
251 |
if (!offset) /* All characters are complete, just copy */ |
252 |
{
|
|
253 |
set(str, arg_length, cs); |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
254 |
return false; |
1
by brian
clean slate |
255 |
}
|
256 |
return copy_aligned(str, arg_length, offset, cs); |
|
257 |
}
|
|
258 |
||
259 |
/* Copy with charset conversion */
|
|
260 |
||
205
by Brian Aker
uint32 -> uin32_t |
261 |
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. |
262 |
const CHARSET_INFO * const from_cs, |
482
by Brian Aker
Remove uint. |
263 |
const CHARSET_INFO * const to_cs, uint32_t *errors) |
1
by brian
clean slate |
264 |
{
|
205
by Brian Aker
uint32 -> uin32_t |
265 |
uint32_t offset; |
1
by brian
clean slate |
266 |
if (!needs_conversion(arg_length, from_cs, to_cs, &offset)) |
267 |
{
|
|
268 |
*errors= 0; |
|
269 |
return copy(str, arg_length, to_cs); |
|
270 |
}
|
|
271 |
if ((from_cs == &my_charset_bin) && offset) |
|
272 |
{
|
|
273 |
*errors= 0; |
|
274 |
return copy_aligned(str, arg_length, offset, to_cs); |
|
275 |
}
|
|
205
by Brian Aker
uint32 -> uin32_t |
276 |
uint32_t new_length= to_cs->mbmaxlen*arg_length; |
1
by brian
clean slate |
277 |
if (alloc(new_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
278 |
return true; |
1
by brian
clean slate |
279 |
str_length=copy_and_convert((char*) Ptr, new_length, to_cs, |
280 |
str, arg_length, from_cs, errors); |
|
281 |
str_charset=to_cs; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
282 |
return false; |
1
by brian
clean slate |
283 |
}
|
284 |
||
285 |
||
286 |
/*
|
|
287 |
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: |
288 |
|
1
by brian
clean slate |
289 |
SYNOPSIS
|
290 |
copy_or_set()
|
|
291 |
str String of a simple charset (latin1)
|
|
292 |
arg_length Length of string
|
|
293 |
||
294 |
IMPLEMENTATION
|
|
295 |
If string object is of a simple character set, set it to point to the
|
|
296 |
given string.
|
|
297 |
If not, make a copy and convert it to the new character set.
|
|
298 |
||
299 |
RETURN
|
|
300 |
0 ok
|
|
301 |
1 Could not allocate result buffer
|
|
302 |
||
303 |
*/
|
|
304 |
||
205
by Brian Aker
uint32 -> uin32_t |
305 |
bool String::set_ascii(const char *str, uint32_t arg_length) |
1
by brian
clean slate |
306 |
{
|
307 |
if (str_charset->mbminlen == 1) |
|
308 |
{
|
|
309 |
set(str, arg_length, str_charset); |
|
310 |
return 0; |
|
311 |
}
|
|
482
by Brian Aker
Remove uint. |
312 |
uint32_t dummy_errors; |
383.1.12
by Brian Aker
Much closer toward UTF8 being around all the time... |
313 |
return copy(str, arg_length, &my_charset_utf8_general_ci, str_charset, &dummy_errors); |
1
by brian
clean slate |
314 |
}
|
315 |
||
316 |
||
317 |
/* This is used by mysql.cc */
|
|
318 |
||
205
by Brian Aker
uint32 -> uin32_t |
319 |
bool String::fill(uint32_t max_length,char fill_char) |
1
by brian
clean slate |
320 |
{
|
321 |
if (str_length > max_length) |
|
322 |
Ptr[str_length=max_length]=0; |
|
323 |
else
|
|
324 |
{
|
|
325 |
if (realloc(max_length)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
326 |
return true; |
212.6.3
by Mats Kindahl
Removing deprecated functions from code and replacing them with C99 equivalents: |
327 |
memset(Ptr+str_length, fill_char, max_length-str_length); |
1
by brian
clean slate |
328 |
str_length=max_length; |
329 |
}
|
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
330 |
return false; |
1
by brian
clean slate |
331 |
}
|
332 |
||
333 |
bool String::append(const String &s) |
|
334 |
{
|
|
335 |
if (s.length()) |
|
336 |
{
|
|
337 |
if (realloc(str_length+s.length())) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
338 |
return true; |
1
by brian
clean slate |
339 |
memcpy(Ptr+str_length,s.ptr(),s.length()); |
340 |
str_length+=s.length(); |
|
341 |
}
|
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
342 |
return false; |
1
by brian
clean slate |
343 |
}
|
344 |
||
345 |
||
346 |
/*
|
|
347 |
Append an ASCII string to the a string of the current character set
|
|
348 |
*/
|
|
349 |
||
205
by Brian Aker
uint32 -> uin32_t |
350 |
bool String::append(const char *s,uint32_t arg_length) |
1
by brian
clean slate |
351 |
{
|
352 |
if (!arg_length) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
353 |
return false; |
1
by brian
clean slate |
354 |
|
355 |
/*
|
|
356 |
For an ASCII incompatible string, e.g. UCS-2, we need to convert
|
|
357 |
*/
|
|
358 |
if (str_charset->mbminlen > 1) |
|
359 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
360 |
uint32_t add_length=arg_length * str_charset->mbmaxlen; |
482
by Brian Aker
Remove uint. |
361 |
uint32_t dummy_errors; |
1
by brian
clean slate |
362 |
if (realloc(str_length+ add_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
363 |
return true; |
1
by brian
clean slate |
364 |
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... |
365 |
s, arg_length, &my_charset_utf8_general_ci, |
1
by brian
clean slate |
366 |
&dummy_errors); |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
367 |
return false; |
1
by brian
clean slate |
368 |
}
|
369 |
||
370 |
/*
|
|
371 |
For an ASCII compatinble string we can just append.
|
|
372 |
*/
|
|
373 |
if (realloc(str_length+arg_length)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
374 |
return true; |
1
by brian
clean slate |
375 |
memcpy(Ptr+str_length,s,arg_length); |
376 |
str_length+=arg_length; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
377 |
return false; |
1
by brian
clean slate |
378 |
}
|
379 |
||
380 |
||
381 |
/*
|
|
382 |
Append a 0-terminated ASCII string
|
|
383 |
*/
|
|
384 |
||
385 |
bool String::append(const char *s) |
|
386 |
{
|
|
387 |
return append(s, strlen(s)); |
|
388 |
}
|
|
389 |
||
390 |
||
391 |
/*
|
|
392 |
Append a string in the given charset to the string
|
|
393 |
with character set recoding
|
|
394 |
*/
|
|
395 |
||
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
396 |
bool String::append(const char *s,uint32_t arg_length, const CHARSET_INFO * const cs) |
1
by brian
clean slate |
397 |
{
|
205
by Brian Aker
uint32 -> uin32_t |
398 |
uint32_t dummy_offset; |
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
399 |
|
1
by brian
clean slate |
400 |
if (needs_conversion(arg_length, cs, str_charset, &dummy_offset)) |
401 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
402 |
uint32_t add_length= arg_length / cs->mbminlen * str_charset->mbmaxlen; |
482
by Brian Aker
Remove uint. |
403 |
uint32_t dummy_errors; |
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
404 |
if (realloc(str_length + add_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
405 |
return true; |
1
by brian
clean slate |
406 |
str_length+= copy_and_convert(Ptr+str_length, add_length, str_charset, |
407 |
s, arg_length, cs, &dummy_errors); |
|
408 |
}
|
|
409 |
else
|
|
410 |
{
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
411 |
if (realloc(str_length + arg_length)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
412 |
return true; |
1
by brian
clean slate |
413 |
memcpy(Ptr + str_length, s, arg_length); |
414 |
str_length+= arg_length; |
|
415 |
}
|
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
416 |
return false; |
1
by brian
clean slate |
417 |
}
|
418 |
||
419 |
||
205
by Brian Aker
uint32 -> uin32_t |
420 |
bool String::append(IO_CACHE* file, uint32_t arg_length) |
1
by brian
clean slate |
421 |
{
|
422 |
if (realloc(str_length+arg_length)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
423 |
return true; |
481
by Brian Aker
Remove all of uchar. |
424 |
if (my_b_read(file, (unsigned char*) Ptr + str_length, arg_length)) |
1
by brian
clean slate |
425 |
{
|
426 |
shrink(str_length); |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
427 |
return true; |
1
by brian
clean slate |
428 |
}
|
429 |
str_length+=arg_length; |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
430 |
return false; |
1
by brian
clean slate |
431 |
}
|
432 |
||
205
by Brian Aker
uint32 -> uin32_t |
433 |
bool String::append_with_prefill(const char *s,uint32_t arg_length, |
434 |
uint32_t full_length, char fill_char) |
|
1
by brian
clean slate |
435 |
{
|
436 |
int t_length= arg_length > full_length ? arg_length : full_length; |
|
437 |
||
438 |
if (realloc(str_length + t_length)) |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
439 |
return true; |
1
by brian
clean slate |
440 |
t_length= full_length - arg_length; |
441 |
if (t_length > 0) |
|
442 |
{
|
|
212.6.3
by Mats Kindahl
Removing deprecated functions from code and replacing them with C99 equivalents: |
443 |
memset(Ptr+str_length, fill_char, t_length); |
1
by brian
clean slate |
444 |
str_length=str_length + t_length; |
445 |
}
|
|
446 |
append(s, arg_length); |
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
447 |
return false; |
1
by brian
clean slate |
448 |
}
|
449 |
||
205
by Brian Aker
uint32 -> uin32_t |
450 |
uint32_t String::numchars() |
1
by brian
clean slate |
451 |
{
|
452 |
return str_charset->cset->numchars(str_charset, Ptr, Ptr+str_length); |
|
453 |
}
|
|
454 |
||
205
by Brian Aker
uint32 -> uin32_t |
455 |
int String::charpos(int i,uint32_t offset) |
1
by brian
clean slate |
456 |
{
|
457 |
if (i <= 0) |
|
458 |
return i; |
|
459 |
return str_charset->cset->charpos(str_charset,Ptr+offset,Ptr+str_length,i); |
|
460 |
}
|
|
461 |
||
205
by Brian Aker
uint32 -> uin32_t |
462 |
int String::strstr(const String &s,uint32_t offset) |
1
by brian
clean slate |
463 |
{
|
464 |
if (s.length()+offset <= str_length) |
|
465 |
{
|
|
466 |
if (!s.length()) |
|
467 |
return ((int) offset); // Empty string is always found |
|
468 |
||
469 |
register const char *str = Ptr+offset; |
|
470 |
register const char *search=s.ptr(); |
|
471 |
const char *end=Ptr+str_length-s.length()+1; |
|
472 |
const char *search_end=s.ptr()+s.length(); |
|
473 |
skip: |
|
474 |
while (str != end) |
|
475 |
{
|
|
476 |
if (*str++ == *search) |
|
477 |
{
|
|
478 |
register char *i,*j; |
|
479 |
i=(char*) str; j=(char*) search+1; |
|
480 |
while (j != search_end) |
|
481 |
if (*i++ != *j++) goto skip; |
|
482 |
return (int) (str-Ptr) -1; |
|
483 |
}
|
|
484 |
}
|
|
485 |
}
|
|
486 |
return -1; |
|
487 |
}
|
|
488 |
||
489 |
/*
|
|
490 |
** Search string from end. Offset is offset to the end of string
|
|
491 |
*/
|
|
492 |
||
205
by Brian Aker
uint32 -> uin32_t |
493 |
int String::strrstr(const String &s,uint32_t offset) |
1
by brian
clean slate |
494 |
{
|
495 |
if (s.length() <= offset && offset <= str_length) |
|
496 |
{
|
|
497 |
if (!s.length()) |
|
498 |
return offset; // Empty string is always found |
|
499 |
register const char *str = Ptr+offset-1; |
|
500 |
register const char *search=s.ptr()+s.length()-1; |
|
501 |
||
502 |
const char *end=Ptr+s.length()-2; |
|
503 |
const char *search_end=s.ptr()-1; |
|
504 |
skip: |
|
505 |
while (str != end) |
|
506 |
{
|
|
507 |
if (*str-- == *search) |
|
508 |
{
|
|
509 |
register char *i,*j; |
|
510 |
i=(char*) str; j=(char*) search-1; |
|
511 |
while (j != search_end) |
|
512 |
if (*i-- != *j--) goto skip; |
|
513 |
return (int) (i-Ptr) +1; |
|
514 |
}
|
|
515 |
}
|
|
516 |
}
|
|
517 |
return -1; |
|
518 |
}
|
|
519 |
||
520 |
/*
|
|
521 |
Replace substring with string
|
|
522 |
If wrong parameter or not enough memory, do nothing
|
|
523 |
*/
|
|
524 |
||
205
by Brian Aker
uint32 -> uin32_t |
525 |
bool String::replace(uint32_t offset,uint32_t arg_length,const String &to) |
1
by brian
clean slate |
526 |
{
|
527 |
return replace(offset,arg_length,to.ptr(),to.length()); |
|
528 |
}
|
|
529 |
||
205
by Brian Aker
uint32 -> uin32_t |
530 |
bool String::replace(uint32_t offset,uint32_t arg_length, |
531 |
const char *to, uint32_t to_length) |
|
1
by brian
clean slate |
532 |
{
|
533 |
long diff = (long) to_length-(long) arg_length; |
|
534 |
if (offset+arg_length <= str_length) |
|
535 |
{
|
|
536 |
if (diff < 0) |
|
537 |
{
|
|
538 |
if (to_length) |
|
539 |
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 |
540 |
memmove(Ptr+offset+to_length, Ptr+offset+arg_length, |
541 |
str_length-offset-arg_length); |
|
1
by brian
clean slate |
542 |
}
|
543 |
else
|
|
544 |
{
|
|
545 |
if (diff) |
|
546 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
547 |
if (realloc(str_length+(uint32_t) diff)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
548 |
return true; |
481
by Brian Aker
Remove all of uchar. |
549 |
bmove_upp((unsigned char*) Ptr+str_length+diff, (unsigned char*) Ptr+str_length, |
1
by brian
clean slate |
550 |
str_length-offset-arg_length); |
551 |
}
|
|
552 |
if (to_length) |
|
553 |
memcpy(Ptr+offset,to,to_length); |
|
554 |
}
|
|
205
by Brian Aker
uint32 -> uin32_t |
555 |
str_length+=(uint32_t) diff; |
1
by brian
clean slate |
556 |
}
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
557 |
return false; |
1
by brian
clean slate |
558 |
}
|
559 |
||
560 |
||
561 |
// added by Holyfoot for "geometry" needs
|
|
205
by Brian Aker
uint32 -> uin32_t |
562 |
int String::reserve(uint32_t space_needed, uint32_t grow_by) |
1
by brian
clean slate |
563 |
{
|
564 |
if (Alloced_length < str_length + space_needed) |
|
565 |
{
|
|
398.1.4
by Monty Taylor
Renamed max/min. |
566 |
if (realloc(Alloced_length + cmax(space_needed, grow_by) - 1)) |
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
567 |
return true; |
1
by brian
clean slate |
568 |
}
|
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
569 |
return false; |
1
by brian
clean slate |
570 |
}
|
571 |
||
205
by Brian Aker
uint32 -> uin32_t |
572 |
void String::qs_append(const char *str, uint32_t len) |
1
by brian
clean slate |
573 |
{
|
574 |
memcpy(Ptr + str_length, str, len + 1); |
|
575 |
str_length += len; |
|
576 |
}
|
|
577 |
||
578 |
void String::qs_append(double d) |
|
579 |
{
|
|
580 |
char *buff = Ptr + str_length; |
|
581 |
str_length+= my_gcvt(d, MY_GCVT_ARG_DOUBLE, FLOATING_POINT_BUFFER - 1, buff, NULL); |
|
582 |
}
|
|
583 |
||
584 |
void String::qs_append(double *d) |
|
585 |
{
|
|
586 |
double ld; |
|
587 |
float8get(ld, (char*) d); |
|
588 |
qs_append(ld); |
|
589 |
}
|
|
590 |
||
591 |
void String::qs_append(int i) |
|
592 |
{
|
|
593 |
char *buff= Ptr + str_length; |
|
594 |
char *end= int10_to_str(i, buff, -10); |
|
595 |
str_length+= (int) (end-buff); |
|
596 |
}
|
|
597 |
||
482
by Brian Aker
Remove uint. |
598 |
void String::qs_append(uint32_t i) |
1
by brian
clean slate |
599 |
{
|
600 |
char *buff= Ptr + str_length; |
|
601 |
char *end= int10_to_str(i, buff, 10); |
|
602 |
str_length+= (int) (end-buff); |
|
603 |
}
|
|
604 |
||
605 |
/*
|
|
606 |
Compare strings according to collation, without end space.
|
|
607 |
||
608 |
SYNOPSIS
|
|
609 |
sortcmp()
|
|
610 |
s First string
|
|
611 |
t Second string
|
|
612 |
cs Collation
|
|
613 |
||
614 |
NOTE:
|
|
615 |
Normally this is case sensitive comparison
|
|
616 |
||
617 |
RETURN
|
|
618 |
< 0 s < t
|
|
619 |
0 s == t
|
|
620 |
> 0 s > t
|
|
621 |
*/
|
|
622 |
||
623 |
||
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
624 |
int sortcmp(const String *s,const String *t, const CHARSET_INFO * const cs) |
1
by brian
clean slate |
625 |
{
|
626 |
return cs->coll->strnncollsp(cs, |
|
481
by Brian Aker
Remove all of uchar. |
627 |
(unsigned char *) s->ptr(),s->length(), |
628 |
(unsigned char *) t->ptr(),t->length(), 0); |
|
1
by brian
clean slate |
629 |
}
|
630 |
||
631 |
||
632 |
/*
|
|
633 |
Compare strings byte by byte. End spaces are also compared.
|
|
634 |
||
635 |
SYNOPSIS
|
|
636 |
stringcmp()
|
|
637 |
s First string
|
|
638 |
t Second string
|
|
639 |
||
640 |
NOTE:
|
|
481
by Brian Aker
Remove all of uchar. |
641 |
Strings are compared as a stream of unsigned chars
|
1
by brian
clean slate |
642 |
|
643 |
RETURN
|
|
644 |
< 0 s < t
|
|
645 |
0 s == t
|
|
646 |
> 0 s > t
|
|
647 |
*/
|
|
648 |
||
649 |
||
650 |
int stringcmp(const String *s,const String *t) |
|
651 |
{
|
|
398.1.4
by Monty Taylor
Renamed max/min. |
652 |
uint32_t s_len=s->length(),t_len=t->length(),len=cmin(s_len,t_len); |
1
by brian
clean slate |
653 |
int cmp= memcmp(s->ptr(), t->ptr(), len); |
654 |
return (cmp) ? cmp : (int) (s_len - t_len); |
|
655 |
}
|
|
656 |
||
657 |
||
205
by Brian Aker
uint32 -> uin32_t |
658 |
String *copy_if_not_alloced(String *to,String *from,uint32_t from_length) |
1
by brian
clean slate |
659 |
{
|
660 |
if (from->Alloced_length >= from_length) |
|
661 |
return from; |
|
662 |
if (from->alloced || !to || from == to) |
|
663 |
{
|
|
664 |
(void) from->realloc(from_length); |
|
665 |
return from; |
|
666 |
}
|
|
667 |
if (to->realloc(from_length)) |
|
668 |
return from; // Actually an error |
|
398.1.4
by Monty Taylor
Renamed max/min. |
669 |
if ((to->str_length=cmin(from->str_length,from_length))) |
1
by brian
clean slate |
670 |
memcpy(to->Ptr,from->Ptr,to->str_length); |
671 |
to->str_charset=from->str_charset; |
|
672 |
return to; |
|
673 |
}
|
|
674 |
||
675 |
||
676 |
/****************************************************************************
|
|
677 |
Help functions
|
|
678 |
****************************************************************************/
|
|
679 |
||
680 |
/*
|
|
681 |
copy a string from one character set to another
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
682 |
|
1
by brian
clean slate |
683 |
SYNOPSIS
|
684 |
copy_and_convert()
|
|
685 |
to Store result here
|
|
686 |
to_cs Character set of result string
|
|
687 |
from Copy from here
|
|
688 |
from_length Length of from string
|
|
689 |
from_cs From character set
|
|
690 |
||
691 |
NOTES
|
|
692 |
'to' must be big enough as form_length * to_cs->mbmaxlen
|
|
693 |
||
694 |
RETURN
|
|
695 |
length of bytes copied to 'to'
|
|
696 |
*/
|
|
697 |
||
698 |
||
205
by Brian Aker
uint32 -> uin32_t |
699 |
static uint32_t |
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
700 |
copy_and_convert_extended(char *to, uint32_t to_length, |
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
701 |
const CHARSET_INFO * const to_cs, |
205
by Brian Aker
uint32 -> uin32_t |
702 |
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. |
703 |
const CHARSET_INFO * const from_cs, |
482
by Brian Aker
Remove uint. |
704 |
uint32_t *errors) |
1
by brian
clean slate |
705 |
{
|
706 |
int cnvres; |
|
707 |
my_wc_t wc; |
|
481
by Brian Aker
Remove all of uchar. |
708 |
const unsigned char *from_end= (const unsigned char*) from+from_length; |
1
by brian
clean slate |
709 |
char *to_start= to; |
481
by Brian Aker
Remove all of uchar. |
710 |
unsigned char *to_end= (unsigned char*) to+to_length; |
1
by brian
clean slate |
711 |
my_charset_conv_mb_wc mb_wc= from_cs->cset->mb_wc; |
712 |
my_charset_conv_wc_mb wc_mb= to_cs->cset->wc_mb; |
|
482
by Brian Aker
Remove uint. |
713 |
uint32_t error_count= 0; |
1
by brian
clean slate |
714 |
|
715 |
while (1) |
|
716 |
{
|
|
481
by Brian Aker
Remove all of uchar. |
717 |
if ((cnvres= (*mb_wc)(from_cs, &wc, (unsigned char*) from, |
1
by brian
clean slate |
718 |
from_end)) > 0) |
719 |
from+= cnvres; |
|
720 |
else if (cnvres == MY_CS_ILSEQ) |
|
721 |
{
|
|
722 |
error_count++; |
|
723 |
from++; |
|
724 |
wc= '?'; |
|
725 |
}
|
|
726 |
else if (cnvres > MY_CS_TOOSMALL) |
|
727 |
{
|
|
728 |
/*
|
|
729 |
A correct multibyte sequence detected
|
|
730 |
But it doesn't have Unicode mapping.
|
|
731 |
*/
|
|
732 |
error_count++; |
|
733 |
from+= (-cnvres); |
|
734 |
wc= '?'; |
|
735 |
}
|
|
736 |
else
|
|
737 |
break; // Not enough characters |
|
738 |
||
739 |
outp: |
|
481
by Brian Aker
Remove all of uchar. |
740 |
if ((cnvres= (*wc_mb)(to_cs, wc, (unsigned char*) to, to_end)) > 0) |
1
by brian
clean slate |
741 |
to+= cnvres; |
742 |
else if (cnvres == MY_CS_ILUNI && wc != '?') |
|
743 |
{
|
|
744 |
error_count++; |
|
745 |
wc= '?'; |
|
746 |
goto outp; |
|
747 |
}
|
|
748 |
else
|
|
749 |
break; |
|
750 |
}
|
|
751 |
*errors= error_count; |
|
205
by Brian Aker
uint32 -> uin32_t |
752 |
return (uint32_t) (to - to_start); |
1
by brian
clean slate |
753 |
}
|
754 |
||
755 |
||
756 |
/*
|
|
757 |
Optimized for quick copying of ASCII characters in the range 0x00..0x7F.
|
|
758 |
*/
|
|
205
by Brian Aker
uint32 -> uin32_t |
759 |
uint32_t
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
760 |
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. |
761 |
const char *from, uint32_t from_length, |
482
by Brian Aker
Remove uint. |
762 |
const CHARSET_INFO * const from_cs, uint32_t *errors) |
1
by brian
clean slate |
763 |
{
|
764 |
/*
|
|
765 |
If any of the character sets is not ASCII compatible,
|
|
766 |
immediately switch to slow mb_wc->wc_mb method.
|
|
767 |
*/
|
|
768 |
if ((to_cs->state | from_cs->state) & MY_CS_NONASCII) |
|
769 |
return copy_and_convert_extended(to, to_length, to_cs, |
|
770 |
from, from_length, from_cs, errors); |
|
771 |
||
398.1.4
by Monty Taylor
Renamed max/min. |
772 |
uint32_t length= cmin(to_length, from_length), length2= length; |
1
by brian
clean slate |
773 |
|
774 |
#if defined(__i386__)
|
|
775 |
/*
|
|
776 |
Special loop for i386, it allows to refer to a
|
|
777 |
non-aligned memory block as UINT32, which makes
|
|
778 |
it possible to copy four bytes at once. This
|
|
779 |
gives about 10% performance improvement comparing
|
|
780 |
to byte-by-byte loop.
|
|
781 |
*/
|
|
782 |
for ( ; length >= 4; length-= 4, from+= 4, to+= 4) |
|
783 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
784 |
if ((*(uint32_t*)from) & 0x80808080) |
1
by brian
clean slate |
785 |
break; |
205
by Brian Aker
uint32 -> uin32_t |
786 |
*((uint32_t*) to)= *((const uint32_t*) from); |
1
by brian
clean slate |
787 |
}
|
788 |
#endif
|
|
789 |
||
790 |
for (; ; *to++= *from++, length--) |
|
791 |
{
|
|
792 |
if (!length) |
|
793 |
{
|
|
794 |
*errors= 0; |
|
795 |
return length2; |
|
796 |
}
|
|
797 |
if (*((unsigned char*) from) > 0x7F) /* A non-ASCII character */ |
|
798 |
{
|
|
205
by Brian Aker
uint32 -> uin32_t |
799 |
uint32_t copied_length= length2 - length; |
1
by brian
clean slate |
800 |
to_length-= copied_length; |
801 |
from_length-= copied_length; |
|
802 |
return copied_length + copy_and_convert_extended(to, to_length, |
|
803 |
to_cs, |
|
804 |
from, from_length, |
|
805 |
from_cs, |
|
806 |
errors); |
|
807 |
}
|
|
808 |
}
|
|
809 |
||
51.1.74
by Jay Pipes
Standardized TRUE/FALSE, removed/replaced DBUG symbols |
810 |
assert(false); // Should never get to here |
1
by brian
clean slate |
811 |
return 0; // Make compiler happy |
812 |
}
|
|
813 |
||
814 |
||
815 |
/**
|
|
816 |
Copy string with HEX-encoding of "bad" characters.
|
|
817 |
||
818 |
@details This functions copies the string pointed by "src"
|
|
819 |
to the string pointed by "dst". Not more than "srclen" bytes
|
|
820 |
are read from "src". Any sequences of bytes representing
|
|
821 |
a not-well-formed substring (according to cs) are hex-encoded,
|
|
822 |
and all well-formed substrings (according to cs) are copied as is.
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
823 |
Not more than "dstlen" bytes are written to "dst". The number
|
1
by brian
clean slate |
824 |
of bytes written to "dst" is returned.
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
825 |
|
1
by brian
clean slate |
826 |
@param cs character set pointer of the destination string
|
827 |
@param[out] dst destination string
|
|
828 |
@param dstlen size of dst
|
|
829 |
@param src source string
|
|
830 |
@param srclen length of src
|
|
831 |
||
832 |
@retval result length
|
|
833 |
*/
|
|
834 |
||
835 |
size_t
|
|
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
836 |
my_copy_with_hex_escaping(const CHARSET_INFO * const cs, |
1
by brian
clean slate |
837 |
char *dst, size_t dstlen, |
838 |
const char *src, size_t srclen) |
|
839 |
{
|
|
840 |
const char *srcend= src + srclen; |
|
841 |
char *dst0= dst; |
|
842 |
||
843 |
for ( ; src < srcend ; ) |
|
844 |
{
|
|
845 |
size_t chlen; |
|
846 |
if ((chlen= my_ismbchar(cs, src, srcend))) |
|
847 |
{
|
|
848 |
if (dstlen < chlen) |
|
849 |
break; /* purecov: inspected */ |
|
850 |
memcpy(dst, src, chlen); |
|
851 |
src+= chlen; |
|
852 |
dst+= chlen; |
|
853 |
dstlen-= chlen; |
|
854 |
}
|
|
855 |
else if (*src & 0x80) |
|
856 |
{
|
|
857 |
if (dstlen < 4) |
|
858 |
break; /* purecov: inspected */ |
|
859 |
*dst++= '\\'; |
|
860 |
*dst++= 'x'; |
|
861 |
*dst++= _dig_vec_upper[((unsigned char) *src) >> 4]; |
|
862 |
*dst++= _dig_vec_upper[((unsigned char) *src) & 15]; |
|
863 |
src++; |
|
864 |
dstlen-= 4; |
|
865 |
}
|
|
866 |
else
|
|
867 |
{
|
|
868 |
if (dstlen < 1) |
|
869 |
break; /* purecov: inspected */ |
|
870 |
*dst++= *src++; |
|
871 |
dstlen--; |
|
872 |
}
|
|
873 |
}
|
|
874 |
return dst - dst0; |
|
875 |
}
|
|
876 |
||
877 |
/*
|
|
878 |
copy a string,
|
|
879 |
with optional character set conversion,
|
|
880 |
with optional left padding (for binary -> UCS2 conversion)
|
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
881 |
|
1
by brian
clean slate |
882 |
SYNOPSIS
|
883 |
well_formed_copy_nchars()
|
|
884 |
to Store result here
|
|
885 |
to_length Maxinum length of "to" string
|
|
886 |
to_cs Character set of "to" string
|
|
887 |
from Copy from here
|
|
888 |
from_length Length of from string
|
|
889 |
from_cs From character set
|
|
890 |
nchars Copy not more that nchars characters
|
|
891 |
well_formed_error_pos Return position when "from" is not well formed
|
|
892 |
or NULL otherwise.
|
|
893 |
cannot_convert_error_pos Return position where a not convertable
|
|
894 |
character met, or NULL otherwise.
|
|
895 |
from_end_pos Return position where scanning of "from"
|
|
896 |
string stopped.
|
|
897 |
NOTES
|
|
898 |
||
899 |
RETURN
|
|
900 |
length of bytes copied to 'to'
|
|
901 |
*/
|
|
902 |
||
903 |
||
205
by Brian Aker
uint32 -> uin32_t |
904 |
uint32_t
|
264.2.6
by Andrey Hristov
Constify the usage of CHARSET_INFO almost to the last place in the code. |
905 |
well_formed_copy_nchars(const CHARSET_INFO * const to_cs, |
482
by Brian Aker
Remove uint. |
906 |
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. |
907 |
const CHARSET_INFO * const from_cs, |
482
by Brian Aker
Remove uint. |
908 |
const char *from, uint32_t from_length, |
909 |
uint32_t nchars, |
|
1
by brian
clean slate |
910 |
const char **well_formed_error_pos, |
911 |
const char **cannot_convert_error_pos, |
|
912 |
const char **from_end_pos) |
|
913 |
{
|
|
482
by Brian Aker
Remove uint. |
914 |
uint32_t res; |
1
by brian
clean slate |
915 |
|
660.1.3
by Eric Herman
removed trailing whitespace with simple script: |
916 |
if ((to_cs == &my_charset_bin) || |
1
by brian
clean slate |
917 |
(from_cs == &my_charset_bin) || |
918 |
(to_cs == from_cs) || |
|
919 |
my_charset_same(from_cs, to_cs)) |
|
920 |
{
|
|
921 |
if (to_length < to_cs->mbminlen || !nchars) |
|
922 |
{
|
|
923 |
*from_end_pos= from; |
|
924 |
*cannot_convert_error_pos= NULL; |
|
925 |
*well_formed_error_pos= NULL; |
|
926 |
return 0; |
|
927 |
}
|
|
928 |
||
929 |
if (to_cs == &my_charset_bin) |
|
930 |
{
|
|
398.1.4
by Monty Taylor
Renamed max/min. |
931 |
res= cmin(cmin(nchars, to_length), from_length); |
1
by brian
clean slate |
932 |
memmove(to, from, res); |
933 |
*from_end_pos= from + res; |
|
934 |
*well_formed_error_pos= NULL; |
|
935 |
*cannot_convert_error_pos= NULL; |
|
936 |
}
|
|
937 |
else
|
|
938 |
{
|
|
939 |
int well_formed_error; |
|
482
by Brian Aker
Remove uint. |
940 |
uint32_t from_offset; |
1
by brian
clean slate |
941 |
|
942 |
if ((from_offset= (from_length % to_cs->mbminlen)) && |
|
943 |
(from_cs == &my_charset_bin)) |
|
944 |
{
|
|
945 |
/*
|
|
946 |
Copying from BINARY to UCS2 needs to prepend zeros sometimes:
|
|
947 |
INSERT INTO t1 (ucs2_column) VALUES (0x01);
|
|
948 |
0x01 -> 0x0001
|
|
949 |
*/
|
|
482
by Brian Aker
Remove uint. |
950 |
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. |
951 |
memset(to, 0, pad_length); |
1
by brian
clean slate |
952 |
memmove(to + pad_length, from, from_offset); |
953 |
nchars--; |
|
954 |
from+= from_offset; |
|
955 |
from_length-= from_offset; |
|
956 |
to+= to_cs->mbminlen; |
|
957 |
to_length-= to_cs->mbminlen; |
|
958 |
}
|
|
959 |
||
960 |
set_if_smaller(from_length, to_length); |
|
961 |
res= to_cs->cset->well_formed_len(to_cs, from, from + from_length, |
|
962 |
nchars, &well_formed_error); |
|
963 |
memmove(to, from, res); |
|
964 |
*from_end_pos= from + res; |
|
965 |
*well_formed_error_pos= well_formed_error ? from + res : NULL; |
|
966 |
*cannot_convert_error_pos= NULL; |
|
967 |
if (from_offset) |
|
968 |
res+= to_cs->mbminlen; |
|
969 |
}
|
|
970 |
}
|
|
971 |
else
|
|
972 |
{
|
|
973 |
int cnvres; |
|
974 |
my_wc_t wc; |
|
975 |
my_charset_conv_mb_wc mb_wc= from_cs->cset->mb_wc; |
|
976 |
my_charset_conv_wc_mb wc_mb= to_cs->cset->wc_mb; |
|
481
by Brian Aker
Remove all of uchar. |
977 |
const unsigned char *from_end= (const unsigned char*) from + from_length; |
978 |
unsigned char *to_end= (unsigned char*) to + to_length; |
|
1
by brian
clean slate |
979 |
char *to_start= to; |
980 |
*well_formed_error_pos= NULL; |
|
981 |
*cannot_convert_error_pos= NULL; |
|
982 |
||
983 |
for ( ; nchars; nchars--) |
|
984 |
{
|
|
985 |
const char *from_prev= from; |
|
481
by Brian Aker
Remove all of uchar. |
986 |
if ((cnvres= (*mb_wc)(from_cs, &wc, (unsigned char*) from, from_end)) > 0) |
1
by brian
clean slate |
987 |
from+= cnvres; |
988 |
else if (cnvres == MY_CS_ILSEQ) |
|
989 |
{
|
|
990 |
if (!*well_formed_error_pos) |
|
991 |
*well_formed_error_pos= from; |
|
992 |
from++; |
|
993 |
wc= '?'; |
|
994 |
}
|
|
995 |
else if (cnvres > MY_CS_TOOSMALL) |
|
996 |
{
|
|
997 |
/*
|
|
998 |
A correct multibyte sequence detected
|
|
999 |
But it doesn't have Unicode mapping.
|
|
1000 |
*/
|
|
1001 |
if (!*cannot_convert_error_pos) |
|
1002 |
*cannot_convert_error_pos= from; |
|
1003 |
from+= (-cnvres); |
|
1004 |
wc= '?'; |
|
1005 |
}
|
|
1006 |
else
|
|
1007 |
break; // Not enough characters |
|
1008 |
||
1009 |
outp: |
|
481
by Brian Aker
Remove all of uchar. |
1010 |
if ((cnvres= (*wc_mb)(to_cs, wc, (unsigned char*) to, to_end)) > 0) |
1
by brian
clean slate |
1011 |
to+= cnvres; |
1012 |
else if (cnvres == MY_CS_ILUNI && wc != '?') |
|
1013 |
{
|
|
1014 |
if (!*cannot_convert_error_pos) |
|
1015 |
*cannot_convert_error_pos= from_prev; |
|
1016 |
wc= '?'; |
|
1017 |
goto outp; |
|
1018 |
}
|
|
1019 |
else
|
|
1020 |
{
|
|
1021 |
from= from_prev; |
|
1022 |
break; |
|
1023 |
}
|
|
1024 |
}
|
|
1025 |
*from_end_pos= from; |
|
1026 |
res= to - to_start; |
|
1027 |
}
|
|
205
by Brian Aker
uint32 -> uin32_t |
1028 |
return (uint32_t) res; |
1
by brian
clean slate |
1029 |
}
|
1030 |
||
1031 |
||
1032 |
||
1033 |
||
1034 |
void String::print(String *str) |
|
1035 |
{
|
|
1036 |
char *st= (char*)Ptr, *end= st+str_length; |
|
1037 |
for (; st < end; st++) |
|
1038 |
{
|
|
481
by Brian Aker
Remove all of uchar. |
1039 |
unsigned char c= *st; |
1
by brian
clean slate |
1040 |
switch (c) |
1041 |
{
|
|
1042 |
case '\\': |
|
520.4.32
by Monty Taylor
Fixed oops. |
1043 |
str->append("\\\\", sizeof("\\\\")-1); |
1
by brian
clean slate |
1044 |
break; |
1045 |
case '\0': |
|
520.4.32
by Monty Taylor
Fixed oops. |
1046 |
str->append("\\0", sizeof("\\0")-1); |
1
by brian
clean slate |
1047 |
break; |
1048 |
case '\'': |
|
520.4.32
by Monty Taylor
Fixed oops. |
1049 |
str->append("\\'", sizeof("\\'")-1); |
1
by brian
clean slate |
1050 |
break; |
1051 |
case '\n': |
|
520.4.32
by Monty Taylor
Fixed oops. |
1052 |
str->append("\\n", sizeof("\\n")-1); |
1
by brian
clean slate |
1053 |
break; |
1054 |
case '\r': |
|
520.4.32
by Monty Taylor
Fixed oops. |
1055 |
str->append("\\r", sizeof("\\r")-1); |
1
by brian
clean slate |
1056 |
break; |
1057 |
case '\032': // Ctrl-Z |
|
520.4.32
by Monty Taylor
Fixed oops. |
1058 |
str->append("\\Z", sizeof("\\Z")-1); |
1
by brian
clean slate |
1059 |
break; |
1060 |
default: |
|
1061 |
str->append(c); |
|
1062 |
}
|
|
1063 |
}
|
|
1064 |
}
|
|
1065 |
||
1066 |
||
1067 |
/*
|
|
1068 |
Exchange state of this object and argument.
|
|
1069 |
||
1070 |
SYNOPSIS
|
|
1071 |
String::swap()
|
|
1072 |
||
1073 |
RETURN
|
|
1074 |
Target string will contain state of this object and vice versa.
|
|
1075 |
*/
|
|
1076 |
||
1077 |
void String::swap(String &s) |
|
1078 |
{
|
|
322.2.2
by Mats Kindahl
Hiding THD::proc_info field and providing a setter and getter. |
1079 |
std::swap(Ptr, s.Ptr); |
1080 |
std::swap(str_length, s.str_length); |
|
1081 |
std::swap(Alloced_length, s.Alloced_length); |
|
1082 |
std::swap(alloced, s.alloced); |
|
1083 |
std::swap(str_charset, s.str_charset); |
|
1
by brian
clean slate |
1084 |
}
|
598.1.1
by Super-User
Fixed solaris build crap. |
1085 |
|
1086 |
||
1087 |
bool operator==(const String &s1, const String &s2) |
|
1088 |
{
|
|
1089 |
return stringcmp(&s1,&s2) == 0; |
|
1090 |
}
|
|
1091 |
||
1092 |
bool operator!=(const String &s1, const String &s2) |
|
1093 |
{
|
|
1094 |
return !(s1 == s2); |
|
1095 |
}
|
|
1096 |