9 |
|
|
10 |
Written by: Philip Hazel <ph10@cam.ac.uk> |
Written by: Philip Hazel <ph10@cam.ac.uk> |
11 |
|
|
12 |
Copyright (c) 1998 University of Cambridge |
Copyright (c) 1997-1999 University of Cambridge |
13 |
|
|
14 |
----------------------------------------------------------------------------- |
----------------------------------------------------------------------------- |
15 |
Permission is granted to anyone to use this software for any purpose on any |
Permission is granted to anyone to use this software for any purpose on any |
47 |
start_bits points to the bit map |
start_bits points to the bit map |
48 |
c is the character |
c is the character |
49 |
caseless the caseless flag |
caseless the caseless flag |
50 |
|
cd the block with char table pointers |
51 |
|
|
52 |
Returns: nothing |
Returns: nothing |
53 |
*/ |
*/ |
54 |
|
|
55 |
static void |
static void |
56 |
set_bit(uschar *start_bits, int c, BOOL caseless) |
set_bit(uschar *start_bits, int c, BOOL caseless, compile_data *cd) |
57 |
{ |
{ |
58 |
start_bits[c/8] |= (1 << (c&7)); |
start_bits[c/8] |= (1 << (c&7)); |
59 |
if (caseless && (pcre_ctypes[c] & ctype_letter) != 0) |
if (caseless && (cd->ctypes[c] & ctype_letter) != 0) |
60 |
start_bits[pcre_fcc[c]/8] |= (1 << (pcre_fcc[c]&7)); |
start_bits[cd->fcc[c]/8] |= (1 << (cd->fcc[c]&7)); |
61 |
} |
} |
62 |
|
|
63 |
|
|
74 |
code points to an expression |
code points to an expression |
75 |
start_bits points to a 32-byte table, initialized to 0 |
start_bits points to a 32-byte table, initialized to 0 |
76 |
caseless the current state of the caseless flag |
caseless the current state of the caseless flag |
77 |
|
cd the block with char table pointers |
78 |
|
|
79 |
Returns: TRUE if table built, FALSE otherwise |
Returns: TRUE if table built, FALSE otherwise |
80 |
*/ |
*/ |
81 |
|
|
82 |
static BOOL |
static BOOL |
83 |
set_start_bits(const uschar *code, uschar *start_bits, BOOL caseless) |
set_start_bits(const uschar *code, uschar *start_bits, BOOL caseless, |
84 |
|
compile_data *cd) |
85 |
{ |
{ |
86 |
register int c; |
register int c; |
87 |
|
|
88 |
|
/* This next statement and the later reference to dummy are here in order to |
89 |
|
trick the optimizer of the IBM C compiler for OS/2 into generating correct |
90 |
|
code. Apparently IBM isn't going to fix the problem, and we would rather not |
91 |
|
disable optimization (in this module it actually makes a big difference, and |
92 |
|
the pcre module can use all the optimization it can get). */ |
93 |
|
|
94 |
|
volatile int dummy; |
95 |
|
|
96 |
do |
do |
97 |
{ |
{ |
98 |
const uschar *tcode = code + 3; |
const uschar *tcode = code + 3; |
107 |
|
|
108 |
if ((int)*tcode >= OP_BRA || *tcode == OP_ASSERT) |
if ((int)*tcode >= OP_BRA || *tcode == OP_ASSERT) |
109 |
{ |
{ |
110 |
if (!set_start_bits(tcode, start_bits, caseless)) return FALSE; |
if (!set_start_bits(tcode, start_bits, caseless, cd)) |
111 |
|
return FALSE; |
112 |
} |
} |
113 |
|
|
114 |
else switch(*tcode) |
else switch(*tcode) |
138 |
|
|
139 |
case OP_BRAZERO: |
case OP_BRAZERO: |
140 |
case OP_BRAMINZERO: |
case OP_BRAMINZERO: |
141 |
if (!set_start_bits(++tcode, start_bits, caseless)) return FALSE; |
if (!set_start_bits(++tcode, start_bits, caseless, cd)) |
142 |
|
return FALSE; |
143 |
|
dummy = 1; |
144 |
do tcode += (tcode[1] << 8) + tcode[2]; while (*tcode == OP_ALT); |
do tcode += (tcode[1] << 8) + tcode[2]; while (*tcode == OP_ALT); |
145 |
tcode += 3; |
tcode += 3; |
146 |
try_next = TRUE; |
try_next = TRUE; |
152 |
case OP_MINSTAR: |
case OP_MINSTAR: |
153 |
case OP_QUERY: |
case OP_QUERY: |
154 |
case OP_MINQUERY: |
case OP_MINQUERY: |
155 |
set_bit(start_bits, tcode[1], caseless); |
set_bit(start_bits, tcode[1], caseless, cd); |
156 |
tcode += 2; |
tcode += 2; |
157 |
try_next = TRUE; |
try_next = TRUE; |
158 |
break; |
break; |
161 |
|
|
162 |
case OP_UPTO: |
case OP_UPTO: |
163 |
case OP_MINUPTO: |
case OP_MINUPTO: |
164 |
set_bit(start_bits, tcode[3], caseless); |
set_bit(start_bits, tcode[3], caseless, cd); |
165 |
tcode += 4; |
tcode += 4; |
166 |
try_next = TRUE; |
try_next = TRUE; |
167 |
break; |
break; |
176 |
|
|
177 |
case OP_PLUS: |
case OP_PLUS: |
178 |
case OP_MINPLUS: |
case OP_MINPLUS: |
179 |
set_bit(start_bits, tcode[1], caseless); |
set_bit(start_bits, tcode[1], caseless, cd); |
180 |
break; |
break; |
181 |
|
|
182 |
/* Single character type sets the bits and stops */ |
/* Single character type sets the bits and stops */ |
183 |
|
|
184 |
case OP_NOT_DIGIT: |
case OP_NOT_DIGIT: |
185 |
for (c = 0; c < 32; c++) start_bits[c] |= ~pcre_cbits[c+cbit_digit]; |
for (c = 0; c < 32; c++) |
186 |
|
start_bits[c] |= ~cd->cbits[c+cbit_digit]; |
187 |
break; |
break; |
188 |
|
|
189 |
case OP_DIGIT: |
case OP_DIGIT: |
190 |
for (c = 0; c < 32; c++) start_bits[c] |= pcre_cbits[c+cbit_digit]; |
for (c = 0; c < 32; c++) |
191 |
|
start_bits[c] |= cd->cbits[c+cbit_digit]; |
192 |
break; |
break; |
193 |
|
|
194 |
case OP_NOT_WHITESPACE: |
case OP_NOT_WHITESPACE: |
195 |
for (c = 0; c < 32; c++) start_bits[c] |= ~pcre_cbits[c+cbit_space]; |
for (c = 0; c < 32; c++) |
196 |
|
start_bits[c] |= ~cd->cbits[c+cbit_space]; |
197 |
break; |
break; |
198 |
|
|
199 |
case OP_WHITESPACE: |
case OP_WHITESPACE: |
200 |
for (c = 0; c < 32; c++) start_bits[c] |= pcre_cbits[c+cbit_space]; |
for (c = 0; c < 32; c++) |
201 |
|
start_bits[c] |= cd->cbits[c+cbit_space]; |
202 |
break; |
break; |
203 |
|
|
204 |
case OP_NOT_WORDCHAR: |
case OP_NOT_WORDCHAR: |
205 |
for (c = 0; c < 32; c++) |
for (c = 0; c < 32; c++) |
206 |
start_bits[c] |= ~(pcre_cbits[c] | pcre_cbits[c+cbit_word]); |
start_bits[c] |= ~(cd->cbits[c] | cd->cbits[c+cbit_word]); |
207 |
break; |
break; |
208 |
|
|
209 |
case OP_WORDCHAR: |
case OP_WORDCHAR: |
210 |
for (c = 0; c < 32; c++) |
for (c = 0; c < 32; c++) |
211 |
start_bits[c] |= (pcre_cbits[c] | pcre_cbits[c+cbit_word]); |
start_bits[c] |= (cd->cbits[c] | cd->cbits[c+cbit_word]); |
212 |
break; |
break; |
213 |
|
|
214 |
/* One or more character type fudges the pointer and restarts, knowing |
/* One or more character type fudges the pointer and restarts, knowing |
239 |
switch(tcode[1]) |
switch(tcode[1]) |
240 |
{ |
{ |
241 |
case OP_NOT_DIGIT: |
case OP_NOT_DIGIT: |
242 |
for (c = 0; c < 32; c++) start_bits[c] |= ~pcre_cbits[c+cbit_digit]; |
for (c = 0; c < 32; c++) |
243 |
|
start_bits[c] |= ~cd->cbits[c+cbit_digit]; |
244 |
break; |
break; |
245 |
|
|
246 |
case OP_DIGIT: |
case OP_DIGIT: |
247 |
for (c = 0; c < 32; c++) start_bits[c] |= pcre_cbits[c+cbit_digit]; |
for (c = 0; c < 32; c++) |
248 |
|
start_bits[c] |= cd->cbits[c+cbit_digit]; |
249 |
break; |
break; |
250 |
|
|
251 |
case OP_NOT_WHITESPACE: |
case OP_NOT_WHITESPACE: |
252 |
for (c = 0; c < 32; c++) start_bits[c] |= ~pcre_cbits[c+cbit_space]; |
for (c = 0; c < 32; c++) |
253 |
|
start_bits[c] |= ~cd->cbits[c+cbit_space]; |
254 |
break; |
break; |
255 |
|
|
256 |
case OP_WHITESPACE: |
case OP_WHITESPACE: |
257 |
for (c = 0; c < 32; c++) start_bits[c] |= pcre_cbits[c+cbit_space]; |
for (c = 0; c < 32; c++) |
258 |
|
start_bits[c] |= cd->cbits[c+cbit_space]; |
259 |
break; |
break; |
260 |
|
|
261 |
case OP_NOT_WORDCHAR: |
case OP_NOT_WORDCHAR: |
262 |
for (c = 0; c < 32; c++) |
for (c = 0; c < 32; c++) |
263 |
start_bits[c] |= ~(pcre_cbits[c] | pcre_cbits[c+cbit_word]); |
start_bits[c] |= ~(cd->cbits[c] | cd->cbits[c+cbit_word]); |
264 |
break; |
break; |
265 |
|
|
266 |
case OP_WORDCHAR: |
case OP_WORDCHAR: |
267 |
for (c = 0; c < 32; c++) |
for (c = 0; c < 32; c++) |
268 |
start_bits[c] |= (pcre_cbits[c] | pcre_cbits[c+cbit_word]); |
start_bits[c] |= (cd->cbits[c] | cd->cbits[c+cbit_word]); |
269 |
break; |
break; |
270 |
} |
} |
271 |
|
|
338 |
uschar start_bits[32]; |
uschar start_bits[32]; |
339 |
real_pcre_extra *extra; |
real_pcre_extra *extra; |
340 |
const real_pcre *re = (const real_pcre *)external_re; |
const real_pcre *re = (const real_pcre *)external_re; |
341 |
|
compile_data compile_block; |
342 |
|
|
343 |
*errorptr = NULL; |
*errorptr = NULL; |
344 |
|
|
361 |
if ((re->options & (PCRE_ANCHORED|PCRE_FIRSTSET|PCRE_STARTLINE)) != 0) |
if ((re->options & (PCRE_ANCHORED|PCRE_FIRSTSET|PCRE_STARTLINE)) != 0) |
362 |
return NULL; |
return NULL; |
363 |
|
|
364 |
|
/* Set the character tables in the block which is passed around */ |
365 |
|
|
366 |
|
compile_block.lcc = re->tables + lcc_offset; |
367 |
|
compile_block.fcc = re->tables + fcc_offset; |
368 |
|
compile_block.cbits = re->tables + cbits_offset; |
369 |
|
compile_block.ctypes = re->tables + ctypes_offset; |
370 |
|
|
371 |
/* See if we can find a fixed set of initial characters for the pattern. */ |
/* See if we can find a fixed set of initial characters for the pattern. */ |
372 |
|
|
373 |
memset(start_bits, 0, 32 * sizeof(uschar)); |
memset(start_bits, 0, 32 * sizeof(uschar)); |
374 |
if (!set_start_bits(re->code, start_bits, (re->options & PCRE_CASELESS) != 0)) |
if (!set_start_bits(re->code, start_bits, (re->options & PCRE_CASELESS) != 0, |
375 |
return NULL; |
&compile_block)) return NULL; |
376 |
|
|
377 |
/* Get an "extra" block and put the information therein. */ |
/* Get an "extra" block and put the information therein. */ |
378 |
|
|