584.4.1
by Monty Taylor
Split out DTCollation. |
1 |
/* -*- mode: c++; c-basic-offset: 2; indent-tabs-mode: nil; -*-
|
2 |
* vim:expandtab:shiftwidth=2:tabstop=2:smarttab:
|
|
3 |
*
|
|
4 |
* Copyright (C) 2008 Sun Microsystems
|
|
5 |
*
|
|
6 |
* This program is free software; you can redistribute it and/or modify
|
|
7 |
* it under the terms of the GNU General Public License as published by
|
|
8 |
* the Free Software Foundation; version 2 of the License.
|
|
9 |
*
|
|
10 |
* This program is distributed in the hope that it will be useful,
|
|
11 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
12 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
13 |
* GNU General Public License for more details.
|
|
14 |
*
|
|
15 |
* You should have received a copy of the GNU General Public License
|
|
16 |
* along with this program; if not, write to the Free Software
|
|
17 |
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
|
|
18 |
*/
|
|
19 |
||
20 |
#include <drizzled/server_includes.h> |
|
584.4.7
by Monty Taylor
Removed a big bank of includes from item.h. |
21 |
#include <drizzled/dtcollation.h> |
22 |
||
584.4.1
by Monty Taylor
Split out DTCollation. |
23 |
#include <drizzled/definitions.h> |
24 |
#include <mysys/my_sys.h> |
|
25 |
#include <mystrings/m_ctype.h> |
|
26 |
#include <drizzled/error.h> |
|
670.1.20
by Monty Taylor
Renamed functions to function... everything else is singular. |
27 |
#include <drizzled/function/str/conv_charset.h> |
584.1.15
by Monty Taylor
The mega-patch from hell. Renamed sql_class to session (since that's what it is) and removed it and field and table from common_includes. |
28 |
#include <drizzled/session.h> |
584.4.1
by Monty Taylor
Split out DTCollation. |
29 |
|
30 |
||
31 |
DTCollation::DTCollation() |
|
32 |
{
|
|
33 |
collation= &my_charset_bin; |
|
34 |
derivation= DERIVATION_NONE; |
|
35 |
}
|
|
36 |
||
37 |
||
38 |
DTCollation::DTCollation(const CHARSET_INFO * const collation_arg, |
|
39 |
Derivation derivation_arg) |
|
40 |
{
|
|
41 |
collation= collation_arg; |
|
42 |
derivation= derivation_arg; |
|
43 |
}
|
|
44 |
||
45 |
||
46 |
void DTCollation::set(DTCollation &dt) |
|
47 |
{
|
|
48 |
collation= dt.collation; |
|
49 |
derivation= dt.derivation; |
|
50 |
}
|
|
51 |
||
52 |
||
53 |
void DTCollation::set(const CHARSET_INFO * const collation_arg, |
|
54 |
Derivation derivation_arg) |
|
55 |
{
|
|
56 |
collation= collation_arg; |
|
57 |
derivation= derivation_arg; |
|
58 |
}
|
|
59 |
||
60 |
||
61 |
void DTCollation::set(const CHARSET_INFO * const collation_arg) |
|
62 |
{
|
|
63 |
collation= collation_arg; |
|
64 |
}
|
|
65 |
||
66 |
||
67 |
void DTCollation::set(Derivation derivation_arg) |
|
68 |
{
|
|
69 |
derivation= derivation_arg; |
|
70 |
}
|
|
71 |
||
72 |
||
73 |
bool DTCollation::aggregate(DTCollation &dt, uint32_t flags) |
|
74 |
{
|
|
75 |
if (!my_charset_same(collation, dt.collation)) |
|
76 |
{
|
|
77 |
/*
|
|
78 |
We do allow to use binary strings (like BLOBS)
|
|
79 |
together with character strings.
|
|
80 |
Binaries have more precedence than a character
|
|
81 |
string of the same derivation.
|
|
82 |
*/
|
|
83 |
if (collation == &my_charset_bin) |
|
84 |
{
|
|
85 |
if (derivation <= dt.derivation) |
|
86 |
; // Do nothing |
|
87 |
else
|
|
88 |
{
|
|
89 |
set(dt); |
|
90 |
}
|
|
91 |
}
|
|
92 |
else if (dt.collation == &my_charset_bin) |
|
93 |
{
|
|
94 |
if (dt.derivation <= derivation) |
|
95 |
{
|
|
96 |
set(dt); |
|
97 |
}
|
|
98 |
else
|
|
99 |
{
|
|
100 |
// Do nothing
|
|
101 |
}
|
|
102 |
}
|
|
103 |
else if ((flags & MY_COLL_ALLOW_SUPERSET_CONV) && |
|
1022.1.3
by Brian Aker
Force UTF8 (remove the bits for looking for ascii). |
104 |
collation->state & MY_CS_UNICODE && |
105 |
(derivation < dt.derivation || |
|
106 |
(derivation == dt.derivation && |
|
107 |
!(dt.collation->state & MY_CS_UNICODE)))) |
|
584.4.1
by Monty Taylor
Split out DTCollation. |
108 |
{
|
109 |
// Do nothing
|
|
110 |
}
|
|
111 |
else if ((flags & MY_COLL_ALLOW_SUPERSET_CONV) && |
|
1022.1.3
by Brian Aker
Force UTF8 (remove the bits for looking for ascii). |
112 |
dt.collation->state & MY_CS_UNICODE && |
113 |
(dt.derivation < derivation || |
|
114 |
(dt.derivation == derivation && |
|
115 |
!(collation->state & MY_CS_UNICODE)))) |
|
584.4.1
by Monty Taylor
Split out DTCollation. |
116 |
{
|
117 |
set(dt); |
|
118 |
}
|
|
119 |
else if ((flags & MY_COLL_ALLOW_COERCIBLE_CONV) && |
|
120 |
derivation < dt.derivation && |
|
121 |
dt.derivation >= DERIVATION_SYSCONST) |
|
122 |
{
|
|
123 |
// Do nothing;
|
|
124 |
}
|
|
125 |
else if ((flags & MY_COLL_ALLOW_COERCIBLE_CONV) && |
|
126 |
dt.derivation < derivation && |
|
127 |
derivation >= DERIVATION_SYSCONST) |
|
128 |
{
|
|
129 |
set(dt); |
|
130 |
}
|
|
131 |
else
|
|
132 |
{
|
|
133 |
// Cannot apply conversion
|
|
1022.1.3
by Brian Aker
Force UTF8 (remove the bits for looking for ascii). |
134 |
set(0, DERIVATION_NONE); |
135 |
return true; |
|
584.4.1
by Monty Taylor
Split out DTCollation. |
136 |
}
|
137 |
}
|
|
138 |
else if (derivation < dt.derivation) |
|
139 |
{
|
|
140 |
// Do nothing
|
|
141 |
}
|
|
142 |
else if (dt.derivation < derivation) |
|
143 |
{
|
|
144 |
set(dt); |
|
145 |
}
|
|
146 |
else
|
|
147 |
{
|
|
148 |
if (collation == dt.collation) |
|
149 |
{
|
|
150 |
// Do nothing
|
|
151 |
}
|
|
152 |
else
|
|
153 |
{
|
|
154 |
if (derivation == DERIVATION_EXPLICIT) |
|
155 |
{
|
|
1022.1.3
by Brian Aker
Force UTF8 (remove the bits for looking for ascii). |
156 |
set(0, DERIVATION_NONE); |
157 |
return true; |
|
584.4.1
by Monty Taylor
Split out DTCollation. |
158 |
}
|
159 |
if (collation->state & MY_CS_BINSORT) |
|
1022.1.3
by Brian Aker
Force UTF8 (remove the bits for looking for ascii). |
160 |
return false; |
584.4.1
by Monty Taylor
Split out DTCollation. |
161 |
if (dt.collation->state & MY_CS_BINSORT) |
162 |
{
|
|
163 |
set(dt); |
|
1022.1.3
by Brian Aker
Force UTF8 (remove the bits for looking for ascii). |
164 |
return false; |
584.4.1
by Monty Taylor
Split out DTCollation. |
165 |
}
|
862
by Brian Aker
Remove charset directory code. |
166 |
const CHARSET_INFO * const bin= get_charset_by_csname(collation->csname, MY_CS_BINSORT); |
584.4.1
by Monty Taylor
Split out DTCollation. |
167 |
set(bin, DERIVATION_NONE); |
168 |
}
|
|
169 |
}
|
|
1022.1.3
by Brian Aker
Force UTF8 (remove the bits for looking for ascii). |
170 |
|
171 |
return false; |
|
584.4.1
by Monty Taylor
Split out DTCollation. |
172 |
}
|
173 |
||
174 |
||
175 |
bool DTCollation::set(DTCollation &dt1, DTCollation &dt2, uint32_t flags) |
|
176 |
{ set(dt1); return aggregate(dt2, flags); } |
|
177 |
||
178 |
||
179 |
const char *DTCollation::derivation_name() const |
|
180 |
{
|
|
181 |
switch(derivation) |
|
182 |
{
|
|
183 |
case DERIVATION_IGNORABLE: return "IGNORABLE"; |
|
184 |
case DERIVATION_COERCIBLE: return "COERCIBLE"; |
|
185 |
case DERIVATION_IMPLICIT: return "IMPLICIT"; |
|
186 |
case DERIVATION_SYSCONST: return "SYSCONST"; |
|
187 |
case DERIVATION_EXPLICIT: return "EXPLICIT"; |
|
188 |
case DERIVATION_NONE: return "NONE"; |
|
189 |
default: return "UNKNOWN"; |
|
190 |
}
|
|
191 |
}
|
|
192 |
||
193 |
||
194 |
bool agg_item_collations(DTCollation &c, const char *fname, |
|
195 |
Item **av, uint32_t count, |
|
196 |
uint32_t flags, int item_sep) |
|
197 |
{
|
|
198 |
uint32_t i; |
|
199 |
Item **arg; |
|
200 |
c.set(av[0]->collation); |
|
201 |
for (i= 1, arg= &av[item_sep]; i < count; i++, arg++) |
|
202 |
{
|
|
203 |
if (c.aggregate((*arg)->collation, flags)) |
|
204 |
{
|
|
205 |
my_coll_agg_error(av, count, fname, item_sep); |
|
206 |
return true; |
|
207 |
}
|
|
208 |
}
|
|
209 |
if ((flags & MY_COLL_DISALLOW_NONE) && |
|
210 |
c.derivation == DERIVATION_NONE) |
|
211 |
{
|
|
212 |
my_coll_agg_error(av, count, fname, item_sep); |
|
213 |
return true; |
|
214 |
}
|
|
215 |
return false; |
|
216 |
}
|
|
217 |
||
218 |
||
219 |
bool agg_item_collations_for_comparison(DTCollation &c, const char *fname, |
|
220 |
Item **av, uint32_t count, |
|
221 |
uint32_t flags) |
|
222 |
{
|
|
223 |
return (agg_item_collations(c, fname, av, count, |
|
224 |
flags | MY_COLL_DISALLOW_NONE, 1)); |
|
225 |
}
|
|
226 |
||
227 |
||
228 |
bool agg_item_charsets(DTCollation &coll, const char *fname, |
|
229 |
Item **args, uint32_t nargs, uint32_t flags, |
|
230 |
int item_sep) |
|
231 |
{
|
|
232 |
Item **arg, *safe_args[2]; |
|
233 |
||
234 |
memset(safe_args, 0, sizeof(safe_args)); |
|
235 |
||
236 |
if (agg_item_collations(coll, fname, args, nargs, flags, item_sep)) |
|
237 |
return true; |
|
238 |
||
239 |
/*
|
|
240 |
For better error reporting: save the first and the second argument.
|
|
241 |
We need this only if the the number of args is 3 or 2:
|
|
242 |
- for a longer argument list, "Illegal mix of collations"
|
|
243 |
doesn't display each argument's characteristics.
|
|
244 |
- if nargs is 1, then this error cannot happen.
|
|
245 |
*/
|
|
246 |
if (nargs >=2 && nargs <= 3) |
|
247 |
{
|
|
248 |
safe_args[0]= args[0]; |
|
249 |
safe_args[1]= args[item_sep]; |
|
250 |
}
|
|
251 |
||
252 |
Session *session= current_session; |
|
253 |
bool res= false; |
|
254 |
uint32_t i; |
|
255 |
||
256 |
for (i= 0, arg= args; i < nargs; i++, arg+= item_sep) |
|
257 |
{
|
|
258 |
Item* conv; |
|
259 |
uint32_t dummy_offset; |
|
260 |
if (!String::needs_conversion(0, (*arg)->collation.collation, |
|
261 |
coll.collation, |
|
262 |
&dummy_offset)) |
|
263 |
continue; |
|
264 |
||
1022.1.3
by Brian Aker
Force UTF8 (remove the bits for looking for ascii). |
265 |
if (!(conv= (*arg)->safe_charset_converter(coll.collation))) |
584.4.1
by Monty Taylor
Split out DTCollation. |
266 |
{
|
267 |
if (nargs >=2 && nargs <= 3) |
|
268 |
{
|
|
269 |
/* restore the original arguments for better error message */
|
|
270 |
args[0]= safe_args[0]; |
|
271 |
args[item_sep]= safe_args[1]; |
|
272 |
}
|
|
273 |
my_coll_agg_error(args, nargs, fname, item_sep); |
|
274 |
res= true; |
|
275 |
break; // we cannot return here, we need to restore "arena". |
|
276 |
}
|
|
277 |
if ((*arg)->type() == Item::FIELD_ITEM) |
|
278 |
((Item_field *)(*arg))->no_const_subst= 1; |
|
279 |
/*
|
|
280 |
If in statement prepare, then we create a converter for two
|
|
281 |
constant items, do it once and then reuse it.
|
|
282 |
If we're in execution of a prepared statement, arena is NULL,
|
|
283 |
and the conv was created in runtime memory. This can be
|
|
284 |
the case only if the argument is a parameter marker ('?'),
|
|
285 |
because for all true constants the charset converter has already
|
|
286 |
been created in prepare. In this case register the change for
|
|
287 |
rollback.
|
|
288 |
*/
|
|
289 |
session->change_item_tree(arg, conv); |
|
290 |
/*
|
|
291 |
We do not check conv->fixed, because Item_func_conv_charset which can
|
|
292 |
be return by safe_charset_converter can't be fixed at creation
|
|
293 |
*/
|
|
294 |
conv->fix_fields(session, arg); |
|
295 |
}
|
|
296 |
||
297 |
return res; |
|
298 |
}
|
|
299 |
||
300 |
||
301 |
void my_coll_agg_error(DTCollation &c1, |
|
302 |
DTCollation &c2, const char *fname) |
|
303 |
{
|
|
304 |
my_error(ER_CANT_AGGREGATE_2COLLATIONS,MYF(0), |
|
305 |
c1.collation->name,c1.derivation_name(), |
|
306 |
c2.collation->name,c2.derivation_name(), |
|
307 |
fname); |
|
308 |
}
|
|
309 |
||
310 |
||
311 |
void my_coll_agg_error(DTCollation &c1, |
|
312 |
DTCollation &c2, |
|
313 |
DTCollation &c3, |
|
314 |
const char *fname) |
|
315 |
{
|
|
316 |
my_error(ER_CANT_AGGREGATE_3COLLATIONS,MYF(0), |
|
317 |
c1.collation->name,c1.derivation_name(), |
|
318 |
c2.collation->name,c2.derivation_name(), |
|
319 |
c3.collation->name,c3.derivation_name(), |
|
320 |
fname); |
|
321 |
}
|
|
322 |
||
323 |
||
324 |
void my_coll_agg_error(Item** args, uint32_t count, const char *fname, |
|
325 |
int item_sep) |
|
326 |
{
|
|
327 |
if (count == 2) |
|
328 |
my_coll_agg_error(args[0]->collation, args[item_sep]->collation, fname); |
|
329 |
else if (count == 3) |
|
330 |
my_coll_agg_error(args[0]->collation, args[item_sep]->collation, |
|
331 |
args[2*item_sep]->collation, fname); |
|
332 |
else
|
|
333 |
my_error(ER_CANT_AGGREGATE_NCOLLATIONS,MYF(0),fname); |
|
334 |
}
|
|
335 |