Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit16c44a9

Browse files
authored
Rework limited range repeats (#628)
1 parenta2a3a6e commit16c44a9

File tree

1 file changed

+118
-124
lines changed

1 file changed

+118
-124
lines changed

‎src/pcre2_jit_compile.c‎

Lines changed: 118 additions & 124 deletions
Original file line numberDiff line numberDiff line change
@@ -10547,7 +10547,7 @@ PCRE2_UCHAR type;
1054710547
sljit_u32max=0,exact;
1054810548
sljit_s32early_fail_ptr=PRIVATE_DATA(cc+1);
1054910549
sljit_s32early_fail_type;
10550-
BOOLcharpos_enabled;
10550+
BOOLcharpos_enabled,use_tmp;
1055110551
PCRE2_UCHARcharpos_char;
1055210552
unsignedintcharpos_othercasebit;
1055310553
PCRE2_SPTRend;
@@ -10560,9 +10560,6 @@ int base = (private_data_ptr == 0) ? SLJIT_MEM1(STACK_TOP) : SLJIT_MEM1(SLJIT_SP
1056010560
intoffset0= (private_data_ptr==0) ?STACK(0) :private_data_ptr;
1056110561
intoffset1= (private_data_ptr==0) ?STACK(1) :private_data_ptr+SSIZE_OF(sw);
1056210562
inttmp_base,tmp_offset;
10563-
#if definedSUPPORT_UNICODE&&PCRE2_CODE_UNIT_WIDTH!=32
10564-
BOOLuse_tmp;
10565-
#endif
1056610563

1056710564
PUSH_BACKTRACK(sizeof(char_iterator_backtrack),cc,NULL);
1056810565

@@ -10754,6 +10751,7 @@ switch(opcode)
1075410751
charpos_char=0;
1075510752
charpos_othercasebit=0;
1075610753

10754+
SLJIT_ASSERT(tmp_base==TMP3);
1075710755
if ((type!=OP_CHAR&&type!=OP_CHARI)&& (*end==OP_CHAR||*end==OP_CHARI))
1075810756
{
1075910757
#ifdefSUPPORT_UNICODE
@@ -10786,153 +10784,146 @@ switch(opcode)
1078610784
BACKTRACK_AS(char_iterator_backtrack)->charpos_enabled= TRUE;
1078710785
BACKTRACK_AS(char_iterator_backtrack)->u.charpos.chr=charpos_char;
1078810786
BACKTRACK_AS(char_iterator_backtrack)->u.charpos.othercasebit=charpos_othercasebit;
10789-
}
10790-
}
1079110787

10792-
SLJIT_ASSERT(tmp_base==TMP3);
10793-
if (charpos_enabled)
10794-
{
10795-
if (opcode==OP_UPTO)
10796-
OP1(SLJIT_MOV,tmp_base,tmp_offset,SLJIT_IMM,max+1);
10788+
if (private_data_ptr==0)
10789+
allocate_stack(common,2);
1079710790

10798-
/* Search the first instance of charpos_char. */
10799-
if (exact==1)
10800-
{
10801-
SLJIT_ASSERT(opcode==OP_STAR);
10802-
detect_partial_match(common,&backtrack->own_backtracks);
10803-
}
10804-
else
10805-
jump=JUMP(SLJIT_JUMP);
10791+
use_tmp= (!HAS_VIRTUAL_REGISTERS&&opcode==OP_STAR);
1080610792

10807-
label=LABEL();
10808-
if (opcode==OP_UPTO)
10809-
{
10810-
OP2(SLJIT_SUB |SLJIT_SET_Z,tmp_base,tmp_offset,tmp_base,tmp_offset,SLJIT_IMM,1);
10811-
add_jump(compiler,&backtrack->own_backtracks,JUMP(SLJIT_ZERO));
10812-
}
10813-
compile_char1_matchingpath(common,type,cc,&backtrack->own_backtracks, FALSE);
10814-
if (early_fail_ptr!=0)
10815-
OP1(SLJIT_MOV,SLJIT_MEM1(SLJIT_SP),early_fail_ptr,STR_PTR,0);
10816-
10817-
if (exact!=1)
10818-
JUMPHERE(jump);
10793+
if (use_tmp)
10794+
{
10795+
OP1(SLJIT_MOV,tmp_base,tmp_offset,SLJIT_IMM,0);
10796+
OP1(SLJIT_MOV,base,offset0,tmp_base,tmp_offset);
10797+
}
10798+
else
10799+
{
10800+
OP1(SLJIT_MOV,TMP1,0,SLJIT_IMM,0);
10801+
OP1(SLJIT_MOV,base,offset0,TMP1,0);
10802+
OP1(SLJIT_MOV,base,offset1,TMP1,0);
10803+
if (opcode==OP_UPTO)
10804+
OP1(SLJIT_MOV,tmp_base,tmp_offset,SLJIT_IMM,max+1);
10805+
}
1081910806

10820-
detect_partial_match(common,&backtrack->own_backtracks);
10821-
OP1(MOV_UCHAR,TMP1,0,SLJIT_MEM1(STR_PTR),IN_UCHARS(0));
10822-
if (charpos_othercasebit!=0)
10823-
OP2(SLJIT_OR,TMP1,0,TMP1,0,SLJIT_IMM,charpos_othercasebit);
10824-
CMPTO(SLJIT_NOT_EQUAL,TMP1,0,SLJIT_IMM,charpos_char,label);
10807+
/* Search the first instance of charpos_char. */
10808+
if (exact==1)
10809+
{
10810+
SLJIT_ASSERT(opcode==OP_STAR);
10811+
detect_partial_match(common,&no_match);
10812+
}
10813+
else
10814+
jump=JUMP(SLJIT_JUMP);
1082510815

10826-
if (private_data_ptr==0)
10827-
{
10828-
if (opcode==OP_UPTO)
10829-
OP1(SLJIT_MOV,TMP2,0,tmp_base,tmp_offset);
10830-
allocate_stack(common,2);
10816+
label=LABEL();
1083110817
if (opcode==OP_UPTO)
10832-
OP1(SLJIT_MOV,tmp_base,tmp_offset,TMP2,0);
10833-
}
10834-
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
10835-
OP1(SLJIT_MOV,base,offset1,STR_PTR,0);
10818+
{
10819+
OP2(SLJIT_SUB |SLJIT_SET_Z,tmp_base,tmp_offset,tmp_base,tmp_offset,SLJIT_IMM,1);
10820+
add_jump(compiler,&no_match,JUMP(SLJIT_ZERO));
10821+
}
10822+
compile_char1_matchingpath(common,type,cc,&no_match, FALSE);
1083610823

10837-
if (opcode==OP_UPTO)
10838-
{
10839-
OP2(SLJIT_SUB |SLJIT_SET_Z,tmp_base,tmp_offset,tmp_base,tmp_offset,SLJIT_IMM,1);
10840-
add_jump(compiler,&no_match,JUMP(SLJIT_ZERO));
10841-
}
10824+
if (early_fail_ptr!=0)
10825+
OP1(SLJIT_MOV,SLJIT_MEM1(SLJIT_SP),early_fail_ptr,STR_PTR,0);
1084210826

10843-
/* Search the last instance of charpos_char. */
10844-
label=LABEL();
10845-
compile_char1_matchingpath(common,type,cc,&no_match, FALSE);
10846-
if (early_fail_ptr!=0)
10847-
OP1(SLJIT_MOV,SLJIT_MEM1(SLJIT_SP),early_fail_ptr,STR_PTR,0);
10848-
detect_partial_match(common,&no_match);
10849-
OP1(MOV_UCHAR,TMP1,0,SLJIT_MEM1(STR_PTR),IN_UCHARS(0));
10850-
if (charpos_othercasebit!=0)
10851-
OP2(SLJIT_OR,TMP1,0,TMP1,0,SLJIT_IMM,charpos_othercasebit);
10827+
if (exact!=1)
10828+
JUMPHERE(jump);
1085210829

10853-
if (opcode==OP_STAR)
10854-
{
10830+
detect_partial_match(common,&no_match);
10831+
OP1(MOV_UCHAR,TMP1,0,SLJIT_MEM1(STR_PTR),IN_UCHARS(0));
10832+
if (charpos_othercasebit!=0)
10833+
OP2(SLJIT_OR,TMP1,0,TMP1,0,SLJIT_IMM,charpos_othercasebit);
1085510834
CMPTO(SLJIT_NOT_EQUAL,TMP1,0,SLJIT_IMM,charpos_char,label);
10856-
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
10835+
10836+
if (use_tmp)
10837+
{
10838+
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
10839+
OP2U(SLJIT_SUB |SLJIT_SET_Z,tmp_base,tmp_offset,SLJIT_IMM,0);
10840+
SELECT(SLJIT_EQUAL,tmp_base,STR_PTR,0,tmp_base);
10841+
}
10842+
else
10843+
{
10844+
OP1(SLJIT_MOV,TMP2,0,base,offset1);
10845+
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
10846+
OP2U(SLJIT_SUB |SLJIT_SET_Z,TMP2,0,SLJIT_IMM,0);
10847+
SELECT(SLJIT_EQUAL,TMP2,STR_PTR,0,TMP2);
10848+
OP1(SLJIT_MOV,base,offset1,TMP2,0);
10849+
}
1085710850
JUMPTO(SLJIT_JUMP,label);
10858-
}
10859-
else
10860-
{
10861-
jump=CMP(SLJIT_NOT_EQUAL,TMP1,0,SLJIT_IMM,charpos_char);
10851+
10852+
set_jumps(no_match,LABEL());
10853+
OP1(SLJIT_MOV,STR_PTR,0,base,offset0);
10854+
if (use_tmp)
10855+
OP1(SLJIT_MOV,base,offset1,tmp_base,tmp_offset);
10856+
10857+
add_jump(compiler,&backtrack->own_backtracks,CMP(SLJIT_EQUAL,STR_PTR,0,SLJIT_IMM,0));
10858+
10859+
BACKTRACK_AS(char_iterator_backtrack)->matchingpath=LABEL();
1086210860
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
10863-
JUMPHERE(jump);
10864-
OP2(SLJIT_SUB |SLJIT_SET_Z,tmp_base,tmp_offset,tmp_base,tmp_offset,SLJIT_IMM,1);
10865-
JUMPTO(SLJIT_NOT_ZERO,label);
10861+
OP2(SLJIT_ADD,STR_PTR,0,STR_PTR,0,SLJIT_IMM,IN_UCHARS(1));
10862+
break;
1086610863
}
10867-
10868-
set_jumps(no_match,LABEL());
10869-
OP2(SLJIT_ADD,STR_PTR,0,base,offset0,SLJIT_IMM,IN_UCHARS(1));
10870-
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
1087110864
}
10872-
else
10873-
{
10874-
if (private_data_ptr==0)
10875-
allocate_stack(common,2);
1087610865

10877-
OP1(SLJIT_MOV,base,offset1,STR_PTR,0);
10866+
if (private_data_ptr==0)
10867+
allocate_stack(common,2);
10868+
10869+
OP1(SLJIT_MOV,base,offset1,STR_PTR,0);
1087810870
#if definedSUPPORT_UNICODE&&PCRE2_CODE_UNIT_WIDTH!=32
10879-
use_tmp= (!HAS_VIRTUAL_REGISTERS&&opcode==OP_STAR);
10880-
SLJIT_ASSERT(!use_tmp||tmp_base==TMP3);
10871+
use_tmp= (!HAS_VIRTUAL_REGISTERS&&opcode==OP_STAR);
10872+
SLJIT_ASSERT(!use_tmp||tmp_base==TMP3);
1088110873

10882-
if (common->utf)
10883-
OP1(SLJIT_MOV,use_tmp ?TMP3 :base,use_tmp ?0 :offset0,STR_PTR,0);
10874+
if (common->utf)
10875+
OP1(SLJIT_MOV,use_tmp ?TMP3 :base,use_tmp ?0 :offset0,STR_PTR,0);
1088410876
#endif
10885-
if (opcode==OP_UPTO)
10886-
OP1(SLJIT_MOV,tmp_base,tmp_offset,SLJIT_IMM,max);
10877+
if (opcode==OP_UPTO)
10878+
OP1(SLJIT_MOV,tmp_base,tmp_offset,SLJIT_IMM,max);
1088710879

10888-
detect_partial_match(common,&no_match);
10889-
label=LABEL();
10890-
compile_char1_matchingpath(common,type,cc,&no_char1_match, FALSE);
10880+
detect_partial_match(common,&no_match);
10881+
label=LABEL();
10882+
compile_char1_matchingpath(common,type,cc,&no_char1_match, FALSE);
1089110883
#if definedSUPPORT_UNICODE&&PCRE2_CODE_UNIT_WIDTH!=32
10892-
if (common->utf)
10893-
OP1(SLJIT_MOV,use_tmp ?TMP3 :base,use_tmp ?0 :offset0,STR_PTR,0);
10884+
if (common->utf)
10885+
OP1(SLJIT_MOV,use_tmp ?TMP3 :base,use_tmp ?0 :offset0,STR_PTR,0);
1089410886
#endif
1089510887

10896-
if (opcode==OP_UPTO)
10897-
{
10898-
OP2(SLJIT_SUB |SLJIT_SET_Z,tmp_base,tmp_offset,tmp_base,tmp_offset,SLJIT_IMM,1);
10899-
add_jump(compiler,&no_match,JUMP(SLJIT_ZERO));
10900-
}
10888+
if (opcode==OP_UPTO)
10889+
{
10890+
OP2(SLJIT_SUB |SLJIT_SET_Z,tmp_base,tmp_offset,tmp_base,tmp_offset,SLJIT_IMM,1);
10891+
add_jump(compiler,&no_match,JUMP(SLJIT_ZERO));
10892+
}
1090110893

10902-
detect_partial_match_to(common,label);
10903-
OP2(SLJIT_ADD,STR_PTR,0,STR_PTR,0,SLJIT_IMM,IN_UCHARS(1));
10894+
detect_partial_match_to(common,label);
10895+
OP2(SLJIT_ADD,STR_PTR,0,STR_PTR,0,SLJIT_IMM,IN_UCHARS(1));
1090410896

10905-
set_jumps(no_char1_match,LABEL());
10897+
set_jumps(no_char1_match,LABEL());
1090610898
#if definedSUPPORT_UNICODE&&PCRE2_CODE_UNIT_WIDTH!=32
10907-
if (common->utf)
10899+
if (common->utf)
10900+
{
10901+
set_jumps(no_match,LABEL());
10902+
if (use_tmp)
1090810903
{
10909-
set_jumps(no_match,LABEL());
10910-
if (use_tmp)
10911-
{
10912-
OP1(SLJIT_MOV,STR_PTR,0,TMP3,0);
10913-
OP1(SLJIT_MOV,base,offset0,TMP3,0);
10914-
}
10915-
else
10916-
OP1(SLJIT_MOV,STR_PTR,0,base,offset0);
10904+
OP1(SLJIT_MOV,STR_PTR,0,TMP3,0);
10905+
OP1(SLJIT_MOV,base,offset0,TMP3,0);
1091710906
}
1091810907
else
10908+
OP1(SLJIT_MOV,STR_PTR,0,base,offset0);
10909+
}
10910+
else
1091910911
#endif
10920-
{
10921-
OP2(SLJIT_SUB,STR_PTR,0,STR_PTR,0,SLJIT_IMM,IN_UCHARS(1));
10922-
set_jumps(no_match,LABEL());
10923-
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
10924-
}
10925-
10926-
if (exact==1)
10927-
{
10928-
SLJIT_ASSERT(opcode==OP_STAR);
10929-
add_jump(compiler,&BACKTRACK_AS(char_iterator_backtrack)->u.backtracks,CMP(SLJIT_EQUAL,base,offset1,STR_PTR,0));
10930-
}
10912+
{
10913+
OP2(SLJIT_SUB,STR_PTR,0,STR_PTR,0,SLJIT_IMM,IN_UCHARS(1));
10914+
set_jumps(no_match,LABEL());
10915+
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
10916+
}
1093110917

10932-
if (early_fail_ptr!=0)
10933-
OP1(SLJIT_MOV,SLJIT_MEM1(SLJIT_SP),early_fail_ptr,STR_PTR,0);
10918+
if (exact==1)
10919+
{
10920+
SLJIT_ASSERT(opcode==OP_STAR);
10921+
add_jump(compiler,&BACKTRACK_AS(char_iterator_backtrack)->u.backtracks,CMP(SLJIT_EQUAL,base,offset1,STR_PTR,0));
1093410922
}
1093510923

10924+
if (early_fail_ptr!=0)
10925+
OP1(SLJIT_MOV,SLJIT_MEM1(SLJIT_SP),early_fail_ptr,STR_PTR,0);
10926+
1093610927
BACKTRACK_AS(char_iterator_backtrack)->matchingpath=LABEL();
1093710928
break;
1093810929

@@ -11661,19 +11652,22 @@ switch(opcode)
1166111652
{
1166211653
OP1(SLJIT_MOV,STR_PTR,0,base,offset0);
1166311654
OP1(SLJIT_MOV,TMP2,0,base,offset1);
11664-
OP2(SLJIT_SUB,STR_PTR,0,STR_PTR,0,SLJIT_IMM,IN_UCHARS(1));
1166511655

1166611656
jump=CMP(SLJIT_LESS_EQUAL,STR_PTR,0,TMP2,0);
1166711657
label=LABEL();
1166811658
if (type==OP_ANYNL)
1166911659
compile_newline_move_back(common);
11670-
OP1(MOV_UCHAR,TMP1,0,SLJIT_MEM1(STR_PTR),IN_UCHARS(-1));
11671-
OP1(SLJIT_MOV,base,offset0,STR_PTR,0);
11660+
move_back(common,NULL, TRUE);
11661+
11662+
OP1(MOV_UCHAR,TMP1,0,SLJIT_MEM1(STR_PTR),IN_UCHARS(0));
1167211663
if (CURRENT_AS(char_iterator_backtrack)->u.charpos.othercasebit!=0)
1167311664
OP2(SLJIT_OR,TMP1,0,TMP1,0,SLJIT_IMM,CURRENT_AS(char_iterator_backtrack)->u.charpos.othercasebit);
1167411665
CMPTO(SLJIT_EQUAL,TMP1,0,SLJIT_IMM,CURRENT_AS(char_iterator_backtrack)->u.charpos.chr,CURRENT_AS(char_iterator_backtrack)->matchingpath);
11675-
move_back(common,NULL, TRUE);
11676-
CMPTO(SLJIT_GREATER,STR_PTR,0,TMP2,0,label);
11666+
/* The range beginning must match, no need to compare. */
11667+
JUMPTO(SLJIT_JUMP,label);
11668+
11669+
set_jumps(current->own_backtracks,LABEL());
11670+
current->own_backtracks=NULL;
1167711671
}
1167811672
else
1167911673
{

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp