Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 997 | diamond | 1 | ; in: ebx -> pattern (not empty zero-terminated string), esi -> translation table, |
| 2 | ; dword [esp+4] = encoding, byte [esp+8] non-zero for whole words only |
||
| 3 | ; out: edx and ecx -> preprocessed data |
||
| 4 | ; when search will be done, edx must be pgfree()d |
||
| 5 | search_string_pre: |
||
| 6 | ; FSM is used, number of states is limited by 256, so pattern length must be <= 255 |
||
| 7 | ; anyway, for big patterns FSM uses too many memory, so probably it is not best choice |
||
| 8 | ; get pattern length, m |
||
| 9 | or ecx, -1 |
||
| 10 | @@: |
||
| 11 | inc ecx |
||
| 12 | cmp byte [ecx+ebx], 0 |
||
| 13 | jnz @b |
||
| 14 | cmp byte [esp+8], 0 |
||
| 15 | jz @f |
||
| 16 | inc ecx |
||
| 17 | inc ecx |
||
| 18 | @@: |
||
| 19 | push ecx |
||
| 20 | ; allocate m*257 bytes for FSM and prefix function |
||
| 21 | imul ecx, 257 |
||
| 22 | call xpgalloc |
||
| 23 | pop ecx |
||
| 24 | test eax, eax |
||
| 25 | jnz @f |
||
| 26 | ret 8 |
||
| 27 | @@: |
||
| 28 | shl ecx, 8 |
||
| 29 | push eax |
||
| 30 | add eax, ecx |
||
| 31 | ; calculate prefix function |
||
| 32 | xor ecx, ecx |
||
| 33 | mov byte [eax], cl |
||
| 34 | xor edi, edi |
||
| 35 | cmp byte [esp+4+8], 0 |
||
| 36 | jnz .whole.prefixcalc |
||
| 37 | .prefixcalc: |
||
| 38 | inc edi |
||
| 39 | movzx edx, byte [ebx+edi] |
||
| 40 | mov dl, [esi+edx] |
||
| 41 | test dl, dl |
||
| 42 | jz .prefixdone |
||
| 43 | @@: |
||
| 44 | push eax |
||
| 45 | movzx eax, byte [ebx+ecx] |
||
| 46 | cmp dl, [esi+eax] |
||
| 47 | pop eax |
||
| 48 | jz @f |
||
| 49 | jecxz .prefixint |
||
| 50 | mov cl, byte [eax+ecx-1] |
||
| 51 | jmp @b |
||
| 52 | @@: |
||
| 53 | inc ecx |
||
| 54 | .prefixint: |
||
| 55 | mov [eax+edi], cl |
||
| 56 | jmp .prefixcalc |
||
| 57 | .whole.prefixcalc: |
||
| 58 | inc edi |
||
| 59 | movzx edx, byte [ebx+edi-1] |
||
| 60 | mov dl, [esi+edx] |
||
| 61 | test dl, dl |
||
| 62 | jz .whole.prefixdone |
||
| 63 | .whole.prefixloop: |
||
| 64 | jecxz .whole.testfirst |
||
| 65 | push eax |
||
| 66 | movzx eax, byte [ebx+ecx-1] |
||
| 67 | cmp dl, [esi+eax] |
||
| 68 | pop eax |
||
| 69 | jz @f |
||
| 70 | mov cl, byte [eax+ecx-1] |
||
| 71 | jmp .whole.prefixloop |
||
| 72 | .whole.testfirst: |
||
| 73 | cmp [isspace_table+edx], 0 |
||
| 74 | jz .whole.prefixint |
||
| 75 | @@: |
||
| 76 | inc ecx |
||
| 77 | .whole.prefixint: |
||
| 78 | mov [eax+edi], cl |
||
| 79 | jmp .whole.prefixcalc |
||
| 80 | .whole.prefixdone: |
||
| 81 | jecxz @f |
||
| 82 | push eax |
||
| 83 | movzx eax, byte [ebx+ecx-1] |
||
| 84 | mov al, [esi+eax] |
||
| 85 | cmp [isspace_table+eax], 0 |
||
| 86 | pop eax |
||
| 87 | jnz @f |
||
| 88 | mov cl, byte [eax+ecx-1] |
||
| 89 | jmp .whole.prefixdone |
||
| 90 | @@: |
||
| 91 | inc ecx |
||
| 92 | mov [eax+edi], cl |
||
| 93 | .prefixdone: |
||
| 94 | pop edx |
||
| 95 | ; create reverse table for encoding+translation |
||
| 96 | push ebp |
||
| 97 | movzx ebp, byte [esp+8] |
||
| 98 | cmp ebp, encodings.unicode |
||
| 99 | jb @f |
||
| 100 | xor ebp, ebp ; no translations for Unicode encodings, |
||
| 101 | ; they must be handled separately by caller |
||
| 102 | @@: |
||
| 103 | mov ecx, 256 |
||
| 104 | @@: |
||
| 105 | push 0 |
||
| 106 | loop @b |
||
| 107 | push ebx eax |
||
| 108 | mov ebx, esp |
||
| 109 | shl ebp, 7 |
||
| 110 | xor eax, eax |
||
| 111 | .createrev: |
||
| 112 | dec cl |
||
| 113 | mov al, cl |
||
| 114 | jns @f |
||
| 115 | mov al, byte [encodings.tables+ebp+ecx-80h] |
||
| 116 | @@: |
||
| 117 | mov al, [esi+eax] |
||
| 118 | pushd [ebx+8+eax*4] |
||
| 119 | pushd ecx |
||
| 120 | mov [ebx+8+eax*4], esp |
||
| 121 | jnz .createrev |
||
| 122 | @@: |
||
| 123 | dec cl |
||
| 124 | mov al, [esi+ecx] |
||
| 125 | pushd [ebx+8+eax*4] |
||
| 126 | popd [ebx+8+ecx*4] |
||
| 127 | jnz @b |
||
| 128 | ; create FSM |
||
| 129 | xor ecx, ecx |
||
| 130 | cmp byte [ebx+259*4+8], 0 |
||
| 131 | mov eax, [ebx] |
||
| 132 | mov ebx, [ebx+4] |
||
| 133 | mov edi, edx |
||
| 134 | jz .fsmcalc |
||
| 135 | mov esi, isspace_table |
||
| 136 | push 256/4 |
||
| 137 | pop ecx |
||
| 138 | rep movsd |
||
| 139 | inc ecx |
||
| 140 | .fsmcalc: |
||
| 141 | movzx esi, byte [eax+ecx] |
||
| 142 | push eax |
||
| 143 | push ecx |
||
| 144 | push 256/4 |
||
| 145 | pop ecx |
||
| 146 | dec esi |
||
| 147 | js .fsmzero |
||
| 148 | shl esi, 8 |
||
| 149 | add esi, edx |
||
| 150 | rep movsd |
||
| 151 | jmp .fsmnext |
||
| 152 | .fsmzero: |
||
| 153 | cmp byte [esp+261*4+256*8+8], 0 |
||
| 154 | jnz .whole.fsmzero |
||
| 155 | xor eax, eax |
||
| 156 | rep stosd |
||
| 157 | jmp .fsmnext |
||
| 158 | .whole.fsmzero: |
||
| 159 | mov esi, edx |
||
| 160 | rep movsd |
||
| 161 | .fsmnext: |
||
| 162 | pop ecx |
||
| 163 | movzx esi, byte [ebx] |
||
| 164 | inc ecx |
||
| 165 | mov esi, [esp+4+8*256+8+esi*4] |
||
| 166 | @@: |
||
| 167 | test esi, esi |
||
| 168 | jz @f |
||
| 169 | lodsd |
||
| 170 | mov [edi-256+eax], cl |
||
| 171 | mov esi, [esi] |
||
| 172 | jmp @b |
||
| 173 | @@: |
||
| 174 | inc ebx |
||
| 175 | pop eax |
||
| 176 | cmp byte [ebx], 0 |
||
| 177 | jnz .fsmcalc |
||
| 178 | cmp byte [esp+259*4+256*8+8], 0 |
||
| 179 | jz .nowholefin |
||
| 180 | movzx esi, byte [eax+ecx] |
||
| 181 | push ecx |
||
| 182 | mov ecx, 256 |
||
| 183 | push 256/4 |
||
| 184 | pop ecx |
||
| 185 | dec esi |
||
| 186 | shl esi, 8 |
||
| 187 | add esi, edx |
||
| 188 | rep movsd |
||
| 189 | pop ecx |
||
| 190 | inc ecx |
||
| 191 | xor eax, eax |
||
| 192 | .whole.fsmfin: |
||
| 193 | cmp [isspace_table+eax], ah |
||
| 194 | jz @f |
||
| 195 | mov byte [edi-256+eax], cl |
||
| 196 | @@: |
||
| 197 | inc al |
||
| 198 | jnz .whole.fsmfin |
||
| 199 | .nowholefin: |
||
| 200 | ; ok, now edx -> FSM, cl = final state |
||
| 201 | add esp, 8*256+8+4*256 |
||
| 202 | pop ebp |
||
| 203 | ret 8 |