Modified: tomcat/jk/trunk/native/iis/pcre/doc/pcrepattern.3 URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/doc/pcrepattern.3?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/doc/pcrepattern.3 (original) +++ tomcat/jk/trunk/native/iis/pcre/doc/pcrepattern.3 Sun Jan 17 17:23:28 2016 @@ -1,4 +1,4 @@ -.TH PCREPATTERN 3 "08 January 2014" "PCRE 8.35" +.TH PCREPATTERN 3 "14 June 2015" "PCRE 8.38" .SH NAME PCRE - Perl-compatible regular expressions .SH "PCRE REGULAR EXPRESSION DETAILS" @@ -308,7 +308,8 @@ A second use of backslash provides a way in patterns in a visible manner. There is no restriction on the appearance of non-printing characters, apart from the binary zero that terminates a pattern, but when a pattern is being prepared by text editing, it is often easier to use -one of the following escape sequences than the binary character it represents: +one of the following escape sequences than the binary character it represents. +In an ASCII or Unicode environment, these escapes are as follows: .sp \ea alarm, that is, the BEL character (hex 07) \ecx "control-x", where x is any ASCII character @@ -331,18 +332,30 @@ but \ec{ becomes hex 3B ({ is 7B), and \ data item (byte or 16-bit value) following \ec has a value greater than 127, a compile-time error occurs. This locks out non-ASCII characters in all modes. .P -The \ec facility was designed for use with ASCII characters, but with the -extension to Unicode it is even less useful than it once was. It is, however, -recognized when PCRE is compiled in EBCDIC mode, where data items are always -bytes. In this mode, all values are valid after \ec. If the next character is a -lower case letter, it is converted to upper case. Then the 0xc0 bits of the -byte are inverted. Thus \ecA becomes hex 01, as in ASCII (A is C1), but because -the EBCDIC letters are disjoint, \ecZ becomes hex 29 (Z is E9), and other -characters also generate different values. +When PCRE is compiled in EBCDIC mode, \ea, \ee, \ef, \en, \er, and \et +generate the appropriate EBCDIC code values. The \ec escape is processed +as specified for Perl in the \fBperlebcdic\fP document. The only characters +that are allowed after \ec are A-Z, a-z, or one of @, [, \e, ], ^, _, or ?. Any +other character provokes a compile-time error. The sequence \e@ encodes +character code 0; the letters (in either case) encode characters 1-26 (hex 01 +to hex 1A); [, \e, ], ^, and _ encode characters 27-31 (hex 1B to hex 1F), and +\e? becomes either 255 (hex FF) or 95 (hex 5F). +.P +Thus, apart from \e?, these escapes generate the same character code values as +they do in an ASCII environment, though the meanings of the values mostly +differ. For example, \eG always generates code value 7, which is BEL in ASCII +but DEL in EBCDIC. +.P +The sequence \e? generates DEL (127, hex 7F) in an ASCII environment, but +because 127 is not a control character in EBCDIC, Perl makes it generate the +APC character. Unfortunately, there are several variants of EBCDIC. In most of +them the APC character has the value 255 (hex FF), but in the one Perl calls +POSIX-BC its value is 95 (hex 5F). If certain other characters have POSIX-BC +values, PCRE makes \e? generate 95; otherwise it generates 255. .P After \e0 up to two further octal digits are read. If there are fewer than two -digits, just those that are present are used. Thus the sequence \e0\ex\e07 -specifies two binary zeros followed by a BEL character (code value 7). Make +digits, just those that are present are used. Thus the sequence \e0\ex\e015 +specifies two binary zeros followed by a CR character (code value 13). Make sure you supply two digits after the initial zero if the pattern character that follows is itself an octal digit. .P @@ -3283,6 +3296,6 @@ Cambridge CB2 3QH, England. .rs .sp .nf -Last updated: 08 January 2014 -Copyright (c) 1997-2014 University of Cambridge. +Last updated: 14 June 2015 +Copyright (c) 1997-2015 University of Cambridge. .fi
Modified: tomcat/jk/trunk/native/iis/pcre/pcre.h.generic URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcre.h.generic?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcre.h.generic (original) +++ tomcat/jk/trunk/native/iis/pcre/pcre.h.generic Sun Jan 17 17:23:28 2016 @@ -42,9 +42,9 @@ POSSIBILITY OF SUCH DAMAGE. /* The current PCRE version information. */ #define PCRE_MAJOR 8 -#define PCRE_MINOR 37 +#define PCRE_MINOR 38 #define PCRE_PRERELEASE -#define PCRE_DATE 2015-04-28 +#define PCRE_DATE 2015-11-23 /* When an application links to a PCRE DLL in Windows, the symbols that are imported have to be identified as such. When building PCRE, the appropriate Modified: tomcat/jk/trunk/native/iis/pcre/pcre_compile.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcre_compile.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcre_compile.c (original) +++ tomcat/jk/trunk/native/iis/pcre/pcre_compile.c Sun Jan 17 17:23:28 2016 @@ -174,7 +174,7 @@ static const short int escapes[] = { -ESC_Z, CHAR_LEFT_SQUARE_BRACKET, CHAR_BACKSLASH, CHAR_RIGHT_SQUARE_BRACKET, CHAR_CIRCUMFLEX_ACCENT, CHAR_UNDERSCORE, - CHAR_GRAVE_ACCENT, 7, + CHAR_GRAVE_ACCENT, ESC_a, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, @@ -202,9 +202,9 @@ static const short int escapes[] = { /* 68 */ 0, 0, '|', ',', '%', '_', '>', '?', /* 70 */ 0, 0, 0, 0, 0, 0, 0, 0, /* 78 */ 0, '`', ':', '#', '@', '\'', '=', '"', -/* 80 */ 0, 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, +/* 80 */ 0, ESC_a, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, /* 88 */-ESC_h, 0, 0, '{', 0, 0, 0, 0, -/* 90 */ 0, 0, -ESC_k, 'l', 0, ESC_n, 0, -ESC_p, +/* 90 */ 0, 0, -ESC_k, 0, 0, ESC_n, 0, -ESC_p, /* 98 */ 0, ESC_r, 0, '}', 0, 0, 0, 0, /* A0 */ 0, '~', -ESC_s, ESC_tee, 0,-ESC_v, -ESC_w, 0, /* A8 */ 0,-ESC_z, 0, 0, 0, '[', 0, 0, @@ -219,6 +219,12 @@ static const short int escapes[] = { /* F0 */ 0, 0, 0, 0, 0, 0, 0, 0, /* F8 */ 0, 0, 0, 0, 0, 0, 0, 0 }; + +/* We also need a table of characters that may follow \c in an EBCDIC +environment for characters 0-31. */ + +static unsigned char ebcdic_escape_c[] = "@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_"; + #endif @@ -458,7 +464,7 @@ static const char error_texts[] = "range out of order in character class\0" "nothing to repeat\0" /* 10 */ - "operand of unlimited repeat could match the empty string\0" /** DEAD **/ + "internal error: invalid forward reference offset\0" "internal error: unexpected repeat\0" "unrecognized character after (? or (?-\0" "POSIX named classes are supported only within a class\0" @@ -527,7 +533,11 @@ static const char error_texts[] = "different names for subpatterns of the same number are not allowed\0" "(*MARK) must have an argument\0" "this version of PCRE is not compiled with Unicode property support\0" +#ifndef EBCDIC "\\c must be followed by an ASCII character\0" +#else + "\\c must be followed by a letter or one of [\\]^_?\0" +#endif "\\k is not followed by a braced, angle-bracketed, or quoted name\0" /* 70 */ "internal error: unknown opcode in find_fixedlength()\0" @@ -1425,7 +1435,16 @@ else c ^= 0x40; #else /* EBCDIC coding */ if (c >= CHAR_a && c <= CHAR_z) c += 64; - c ^= 0xC0; + if (c == CHAR_QUESTION_MARK) + c = ('\\' == 188 && '`' == 74)? 0x5f : 0xff; + else + { + for (i = 0; i < 32; i++) + { + if (c == ebcdic_escape_c[i]) break; + } + if (i < 32) c = i; else *errorcodeptr = ERR68; + } #endif break; @@ -1799,7 +1818,7 @@ for (;;) case OP_ASSERTBACK: case OP_ASSERTBACK_NOT: do cc += GET(cc, 1); while (*cc == OP_ALT); - cc += PRIV(OP_lengths)[*cc]; + cc += 1 + LINK_SIZE; break; /* Skip over things that don't match chars */ @@ -2487,7 +2506,7 @@ for (code = first_significant_code(code if (c == OP_BRA || c == OP_BRAPOS || c == OP_CBRA || c == OP_CBRAPOS || c == OP_ONCE || c == OP_ONCE_NC || - c == OP_COND) + c == OP_COND || c == OP_SCOND) { BOOL empty_branch; if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */ @@ -3886,11 +3905,11 @@ didn't consider this to be a POSIX class The problem in trying to be exactly like Perl is in the handling of escapes. We have to be sure that [abc[:x\]pqr] is *not* treated as containing a POSIX class, but [abc[:x\]pqr:]] is (so that an error can be generated). The code -below handles the special case of \], but does not try to do any other escape -processing. This makes it different from Perl for cases such as [:l\ower:] -where Perl recognizes it as the POSIX class "lower" but PCRE does not recognize -"l\ower". This is a lesser evil than not diagnosing bad classes when Perl does, -I think. +below handles the special cases \\ and \], but does not try to do any other +escape processing. This makes it different from Perl for cases such as +[:l\ower:] where Perl recognizes it as the POSIX class "lower" but PCRE does +not recognize "l\ower". This is a lesser evil than not diagnosing bad classes +when Perl does, I think. A user pointed out that PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the appearance of a nested POSIX class supersedes an apparent @@ -3917,21 +3936,16 @@ pcre_uchar terminator; /* Don't terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */ for (++ptr; *ptr != CHAR_NULL; ptr++) { - if (*ptr == CHAR_BACKSLASH && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) + if (*ptr == CHAR_BACKSLASH && + (ptr[1] == CHAR_RIGHT_SQUARE_BRACKET || + ptr[1] == CHAR_BACKSLASH)) ptr++; - else if (*ptr == CHAR_RIGHT_SQUARE_BRACKET) return FALSE; - else + else if ((*ptr == CHAR_LEFT_SQUARE_BRACKET && ptr[1] == terminator) || + *ptr == CHAR_RIGHT_SQUARE_BRACKET) return FALSE; + else if (*ptr == terminator && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) { - if (*ptr == terminator && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) - { - *endptr = ptr; - return TRUE; - } - if (*ptr == CHAR_LEFT_SQUARE_BRACKET && - (ptr[1] == CHAR_COLON || ptr[1] == CHAR_DOT || - ptr[1] == CHAR_EQUALS_SIGN) && - check_posix_syntax(ptr, endptr)) - return FALSE; + *endptr = ptr; + return TRUE; } } return FALSE; @@ -3985,11 +3999,12 @@ have their offsets adjusted. That one of is called, the partially compiled regex must be temporarily terminated with OP_END. -This function has been extended with the possibility of forward references for -recursions and subroutine calls. It must also check the list of such references -for the group we are dealing with. If it finds that one of the recursions in -the current group is on this list, it adjusts the offset in the list, not the -value in the reference (which is a group number). +This function has been extended to cope with forward references for recursions +and subroutine calls. It must check the list of such references for the +group we are dealing with. If it finds that one of the recursions in the +current group is on this list, it does not adjust the value in the reference +(which is a group number). After the group has been scanned, all the offsets in +the forward reference list for the group are adjusted. Arguments: group points to the start of the group @@ -4005,29 +4020,21 @@ static void adjust_recurse(pcre_uchar *group, int adjust, BOOL utf, compile_data *cd, size_t save_hwm_offset) { +int offset; +pcre_uchar *hc; pcre_uchar *ptr = group; while ((ptr = (pcre_uchar *)find_recurse(ptr, utf)) != NULL) { - int offset; - pcre_uchar *hc; - - /* See if this recursion is on the forward reference list. If so, adjust the - reference. */ - for (hc = (pcre_uchar *)cd->start_workspace + save_hwm_offset; hc < cd->hwm; hc += LINK_SIZE) { offset = (int)GET(hc, 0); - if (cd->start_code + offset == ptr + 1) - { - PUT(hc, 0, offset + adjust); - break; - } + if (cd->start_code + offset == ptr + 1) break; } - /* Otherwise, adjust the recursion offset if it's after the start of this - group. */ + /* If we have not found this recursion on the forward reference list, adjust + the recursion's offset if it's after the start of this group. */ if (hc >= cd->hwm) { @@ -4037,6 +4044,15 @@ while ((ptr = (pcre_uchar *)find_recurse ptr += 1 + LINK_SIZE; } + +/* Now adjust all forward reference offsets for the group. */ + +for (hc = (pcre_uchar *)cd->start_workspace + save_hwm_offset; hc < cd->hwm; + hc += LINK_SIZE) + { + offset = (int)GET(hc, 0); + PUT(hc, 0, offset + adjust); + } } @@ -4465,7 +4481,7 @@ const pcre_uchar *tempptr; const pcre_uchar *nestptr = NULL; pcre_uchar *previous = NULL; pcre_uchar *previous_callout = NULL; -size_t save_hwm_offset = 0; +size_t item_hwm_offset = 0; pcre_uint8 classbits[32]; /* We can fish out the UTF-8 setting once and for all into a BOOL, but we @@ -4623,8 +4639,7 @@ for (;; ptr++) /* In the real compile phase, just check the workspace used by the forward reference list. */ - else if (cd->hwm > cd->start_workspace + cd->workspace_size - - WORK_SIZE_SAFETY_MARGIN) + else if (cd->hwm > cd->start_workspace + cd->workspace_size) { *errorcodeptr = ERR52; goto FAILED; @@ -4767,6 +4782,7 @@ for (;; ptr++) zeroreqchar = reqchar; zeroreqcharflags = reqcharflags; previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; *code++ = ((options & PCRE_DOTALL) != 0)? OP_ALLANY: OP_ANY; break; @@ -4818,6 +4834,7 @@ for (;; ptr++) /* Handle a real character class. */ previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; /* PCRE supports POSIX class stuff inside a class. Perl gives an error if they are encountered at the top level, so we'll do that too. */ @@ -4923,9 +4940,10 @@ for (;; ptr++) (which is on the stack). We have to remember that there was XCLASS data, however. */ + if (class_uchardata > class_uchardata_base) xclass = TRUE; + if (lengthptr != NULL && class_uchardata > class_uchardata_base) { - xclass = TRUE; *lengthptr += (int)(class_uchardata - class_uchardata_base); class_uchardata = class_uchardata_base; } @@ -5028,10 +5046,26 @@ for (;; ptr++) ptr = tempptr + 1; continue; - /* For all other POSIX classes, no special action is taken in UCP - mode. Fall through to the non_UCP case. */ + /* For the other POSIX classes (ascii, xdigit) we are going to fall + through to the non-UCP case and build a bit map for characters with + code points less than 256. If we are in a negated POSIX class + within a non-negated overall class, characters with code points + greater than 255 must all match. In the special case where we have + not yet generated any xclass data, and this is the final item in + the overall class, we need do nothing: later on, the opcode + OP_NCLASS will be used to indicate that characters greater than 255 + are acceptable. If we have already seen an xclass item or one may + follow (we have to assume that it might if this is not the end of + the class), explicitly match all wide codepoints. */ default: + if (!negate_class && local_negate && + (xclass || tempptr[2] != CHAR_RIGHT_SQUARE_BRACKET)) + { + *class_uchardata++ = XCL_RANGE; + class_uchardata += PRIV(ord2utf)(0x100, class_uchardata); + class_uchardata += PRIV(ord2utf)(0x10ffff, class_uchardata); + } break; } } @@ -5195,9 +5229,9 @@ for (;; ptr++) cd, PRIV(vspace_list)); continue; -#ifdef SUPPORT_UCP case ESC_p: case ESC_P: +#ifdef SUPPORT_UCP { BOOL negated; unsigned int ptype = 0, pdata = 0; @@ -5211,6 +5245,9 @@ for (;; ptr++) class_has_8bitchar--; /* Undo! */ continue; } +#else + *errorcodeptr = ERR45; + goto FAILED; #endif /* Unrecognized escapes are faulted if PCRE is running in its strict mode. By default, for compatibility with Perl, they are @@ -5367,16 +5404,20 @@ for (;; ptr++) CLASS_SINGLE_CHARACTER: if (class_one_char < 2) class_one_char++; - /* If class_one_char is 1, we have the first single character in the - class, and there have been no prior ranges, or XCLASS items generated by - escapes. If this is the final character in the class, we can optimize by - turning the item into a 1-character OP_CHAR[I] if it's positive, or - OP_NOT[I] if it's negative. In the positive case, it can cause firstchar - to be set. Otherwise, there can be no first char if this item is first, - whatever repeat count may follow. In the case of reqchar, save the - previous value for reinstating. */ + /* If xclass_has_prop is false and class_one_char is 1, we have the first + single character in the class, and there have been no prior ranges, or + XCLASS items generated by escapes. If this is the final character in the + class, we can optimize by turning the item into a 1-character OP_CHAR[I] + if it's positive, or OP_NOT[I] if it's negative. In the positive case, it + can cause firstchar to be set. Otherwise, there can be no first char if + this item is first, whatever repeat count may follow. In the case of + reqchar, save the previous value for reinstating. */ - if (!inescq && class_one_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) + if (!inescq && +#ifdef SUPPORT_UCP + !xclass_has_prop && +#endif + class_one_char == 1 && ptr[1] == CHAR_RIGHT_SQUARE_BRACKET) { ptr++; zeroreqchar = reqchar; @@ -5492,9 +5533,10 @@ for (;; ptr++) actual compiled code. */ #ifdef SUPPORT_UTF - if (xclass && (!should_flip_negation || (options & PCRE_UCP) != 0)) + if (xclass && (xclass_has_prop || !should_flip_negation || + (options & PCRE_UCP) != 0)) #elif !defined COMPILE_PCRE8 - if (xclass && !should_flip_negation) + if (xclass && (xclass_has_prop || !should_flip_negation)) #endif #if defined SUPPORT_UTF || !defined COMPILE_PCRE8 { @@ -5930,7 +5972,7 @@ for (;; ptr++) { register int i; int len = (int)(code - previous); - size_t base_hwm_offset = save_hwm_offset; + size_t base_hwm_offset = item_hwm_offset; pcre_uchar *bralink = NULL; pcre_uchar *brazeroptr = NULL; @@ -5985,7 +6027,7 @@ for (;; ptr++) if (repeat_max <= 1) /* Covers 0, 1, and unlimited */ { *code = OP_END; - adjust_recurse(previous, 1, utf, cd, save_hwm_offset); + adjust_recurse(previous, 1, utf, cd, item_hwm_offset); memmove(previous + 1, previous, IN_UCHARS(len)); code++; if (repeat_max == 0) @@ -6009,7 +6051,7 @@ for (;; ptr++) { int offset; *code = OP_END; - adjust_recurse(previous, 2 + LINK_SIZE, utf, cd, save_hwm_offset); + adjust_recurse(previous, 2 + LINK_SIZE, utf, cd, item_hwm_offset); memmove(previous + 2 + LINK_SIZE, previous, IN_UCHARS(len)); code += 2 + LINK_SIZE; *previous++ = OP_BRAZERO + repeat_type; @@ -6254,6 +6296,12 @@ for (;; ptr++) while (*scode == OP_ALT); } + /* A conditional group with only one branch has an implicit empty + alternative branch. */ + + if (*bracode == OP_COND && bracode[GET(bracode,1)] != OP_ALT) + *bracode = OP_SCOND; + /* Handle possessive quantifiers. */ if (possessive_quantifier) @@ -6267,11 +6315,11 @@ for (;; ptr++) { int nlen = (int)(code - bracode); *code = OP_END; - adjust_recurse(bracode, 1 + LINK_SIZE, utf, cd, save_hwm_offset); + adjust_recurse(bracode, 1 + LINK_SIZE, utf, cd, item_hwm_offset); memmove(bracode + 1 + LINK_SIZE, bracode, IN_UCHARS(nlen)); code += 1 + LINK_SIZE; nlen += 1 + LINK_SIZE; - *bracode = OP_BRAPOS; + *bracode = (*bracode == OP_COND)? OP_BRAPOS : OP_SBRAPOS; *code++ = OP_KETRPOS; PUTINC(code, 0, nlen); PUT(bracode, 1, nlen); @@ -6401,7 +6449,7 @@ for (;; ptr++) else { *code = OP_END; - adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, save_hwm_offset); + adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, item_hwm_offset); memmove(tempcode + 1 + LINK_SIZE, tempcode, IN_UCHARS(len)); code += 1 + LINK_SIZE; len += 1 + LINK_SIZE; @@ -6450,7 +6498,7 @@ for (;; ptr++) default: *code = OP_END; - adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, save_hwm_offset); + adjust_recurse(tempcode, 1 + LINK_SIZE, utf, cd, item_hwm_offset); memmove(tempcode + 1 + LINK_SIZE, tempcode, IN_UCHARS(len)); code += 1 + LINK_SIZE; len += 1 + LINK_SIZE; @@ -6586,9 +6634,17 @@ for (;; ptr++) goto FAILED; } setverb = *code++ = verbs[i].op_arg; - *code++ = arglen; - memcpy(code, arg, IN_UCHARS(arglen)); - code += arglen; + if (lengthptr != NULL) /* In pass 1 just add in the length */ + { /* to avoid potential workspace */ + *lengthptr += arglen; /* overflow. */ + *code++ = 0; + } + else + { + *code++ = arglen; + memcpy(code, arg, IN_UCHARS(arglen)); + code += arglen; + } *code++ = 0; } @@ -6623,7 +6679,7 @@ for (;; ptr++) newoptions = options; skipbytes = 0; bravalue = OP_CBRA; - save_hwm_offset = cd->hwm - cd->start_workspace; + item_hwm_offset = cd->hwm - cd->start_workspace; reset_bracount = FALSE; /* Deal with the extended parentheses; all are introduced by '?', and the @@ -6641,6 +6697,7 @@ for (;; ptr++) /* ------------------------------------------------------------ */ case CHAR_VERTICAL_LINE: /* Reset capture count for each branch */ reset_bracount = TRUE; + cd->dupgroups = TRUE; /* Record (?| encountered */ /* Fall through */ /* ------------------------------------------------------------ */ @@ -6741,6 +6798,12 @@ for (;; ptr++) { while (IS_DIGIT(*ptr)) { + if (recno > INT_MAX / 10 - 1) /* Integer overflow */ + { + while (IS_DIGIT(*ptr)) ptr++; + *errorcodeptr = ERR61; + goto FAILED; + } recno = recno * 10 + (int)(*ptr - CHAR_0); ptr++; } @@ -6769,7 +6832,7 @@ for (;; ptr++) ptr++; } namelen = (int)(ptr - name); - if (lengthptr != NULL) *lengthptr += IMM2_SIZE; + if (lengthptr != NULL) skipbytes += IMM2_SIZE; } /* Check the terminator */ @@ -6875,6 +6938,11 @@ for (;; ptr++) *errorcodeptr = ERR15; goto FAILED; } + if (recno > INT_MAX / 10 - 1) /* Integer overflow */ + { + *errorcodeptr = ERR61; + goto FAILED; + } recno = recno * 10 + name[i] - CHAR_0; } if (recno == 0) recno = RREF_ANY; @@ -7151,6 +7219,7 @@ for (;; ptr++) if (lengthptr != NULL) { named_group *ng; + recno = 0; if (namelen == 0) { @@ -7168,20 +7237,6 @@ for (;; ptr++) goto FAILED; } - /* The name table does not exist in the first pass; instead we must - scan the list of names encountered so far in order to get the - number. If the name is not found, set the value to 0 for a forward - reference. */ - - ng = cd->named_groups; - for (i = 0; i < cd->names_found; i++, ng++) - { - if (namelen == ng->length && - STRNCMP_UC_UC(name, ng->name, namelen) == 0) - break; - } - recno = (i < cd->names_found)? ng->number : 0; - /* Count named back references. */ if (!is_recurse) cd->namedrefcount++; @@ -7191,6 +7246,56 @@ for (;; ptr++) 16-bit data item. */ *lengthptr += IMM2_SIZE; + + /* If this is a forward reference and we are within a (?|...) group, + the reference may end up as the number of a group which we are + currently inside, that is, it could be a recursive reference. In the + real compile this will be picked up and the reference wrapped with + OP_ONCE to make it atomic, so we must space in case this occurs. */ + + /* In fact, this can happen for a non-forward reference because + another group with the same number might be created later. This + issue is fixed "properly" in PCRE2. As PCRE1 is now in maintenance + only mode, we finesse the bug by allowing more memory always. */ + + *lengthptr += 2 + 2*LINK_SIZE; + + /* It is even worse than that. The current reference may be to an + existing named group with a different number (so apparently not + recursive) but which later on is also attached to a group with the + current number. This can only happen if $(| has been previous + encountered. In that case, we allow yet more memory, just in case. + (Again, this is fixed "properly" in PCRE2. */ + + if (cd->dupgroups) *lengthptr += 4 + 4*LINK_SIZE; + + /* Otherwise, check for recursion here. The name table does not exist + in the first pass; instead we must scan the list of names encountered + so far in order to get the number. If the name is not found, leave + the value of recno as 0 for a forward reference. */ + + else + { + ng = cd->named_groups; + for (i = 0; i < cd->names_found; i++, ng++) + { + if (namelen == ng->length && + STRNCMP_UC_UC(name, ng->name, namelen) == 0) + { + open_capitem *oc; + recno = ng->number; + if (is_recurse) break; + for (oc = cd->open_caps; oc != NULL; oc = oc->next) + { + if (oc->number == recno) + { + oc->flag = TRUE; + break; + } + } + } + } + } } /* In the real compile, search the name table. We check the name @@ -7237,8 +7342,6 @@ for (;; ptr++) for (i++; i < cd->names_found; i++) { if (STRCMP_UC_UC(slot + IMM2_SIZE, cslot + IMM2_SIZE) != 0) break; - - count++; cslot += cd->name_entry_size; } @@ -7247,6 +7350,7 @@ for (;; ptr++) { if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; *code++ = ((options & PCRE_CASELESS) != 0)? OP_DNREFI : OP_DNREF; PUT2INC(code, 0, index); PUT2INC(code, 0, count); @@ -7284,9 +7388,14 @@ for (;; ptr++) /* ------------------------------------------------------------ */ - case CHAR_R: /* Recursion */ - ptr++; /* Same as (?0) */ - /* Fall through */ + case CHAR_R: /* Recursion, same as (?0) */ + recno = 0; + if (*(++ptr) != CHAR_RIGHT_PARENTHESIS) + { + *errorcodeptr = ERR29; + goto FAILED; + } + goto HANDLE_RECURSION; /* ------------------------------------------------------------ */ @@ -7323,7 +7432,15 @@ for (;; ptr++) recno = 0; while(IS_DIGIT(*ptr)) + { + if (recno > INT_MAX / 10 - 1) /* Integer overflow */ + { + while (IS_DIGIT(*ptr)) ptr++; + *errorcodeptr = ERR61; + goto FAILED; + } recno = recno * 10 + *ptr++ - CHAR_0; + } if (*ptr != (pcre_uchar)terminator) { @@ -7360,6 +7477,7 @@ for (;; ptr++) HANDLE_RECURSION: previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; called = cd->start_code; /* When we are actually compiling, find the bracket that is being @@ -7561,7 +7679,11 @@ for (;; ptr++) previous = NULL; cd->iscondassert = FALSE; } - else previous = code; + else + { + previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; + } *code = bravalue; tempcode = code; @@ -7809,7 +7931,7 @@ for (;; ptr++) const pcre_uchar *p; pcre_uint32 cf; - save_hwm_offset = cd->hwm - cd->start_workspace; /* Normally this is set when '(' is read */ + item_hwm_offset = cd->hwm - cd->start_workspace; /* Normally this is set when '(' is read */ terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)? CHAR_GREATER_THAN_SIGN : CHAR_APOSTROPHE; @@ -7838,7 +7960,7 @@ for (;; ptr++) if (*p != (pcre_uchar)terminator) { *errorcodeptr = ERR57; - break; + goto FAILED; } ptr++; goto HANDLE_NUMERICAL_RECURSION; @@ -7853,7 +7975,7 @@ for (;; ptr++) ptr[1] != CHAR_APOSTROPHE && ptr[1] != CHAR_LEFT_CURLY_BRACKET)) { *errorcodeptr = ERR69; - break; + goto FAILED; } is_recurse = FALSE; terminator = (*(++ptr) == CHAR_LESS_THAN_SIGN)? @@ -7877,6 +7999,7 @@ for (;; ptr++) HANDLE_REFERENCE: if (firstcharflags == REQ_UNSET) firstcharflags = REQ_NONE; previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; *code++ = ((options & PCRE_CASELESS) != 0)? OP_REFI : OP_REF; PUT2INC(code, 0, recno); cd->backref_map |= (recno < 32)? (1 << recno) : 1; @@ -7906,6 +8029,7 @@ for (;; ptr++) if (!get_ucp(&ptr, &negated, &ptype, &pdata, errorcodeptr)) goto FAILED; previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; *code++ = ((escape == ESC_p) != negated)? OP_PROP : OP_NOTPROP; *code++ = ptype; *code++ = pdata; @@ -7946,6 +8070,7 @@ for (;; ptr++) { previous = (escape > ESC_b && escape < ESC_Z)? code : NULL; + item_hwm_offset = cd->hwm - cd->start_workspace; *code++ = (!utf && escape == ESC_C)? OP_ALLANY : escape; } } @@ -7989,6 +8114,7 @@ for (;; ptr++) ONE_CHAR: previous = code; + item_hwm_offset = cd->hwm - cd->start_workspace; /* For caseless UTF-8 mode when UCP support is available, check whether this character has more than one other case. If so, generate a special @@ -9164,6 +9290,7 @@ cd->names_found = 0; cd->name_entry_size = 0; cd->name_table = NULL; cd->dupnames = FALSE; +cd->dupgroups = FALSE; cd->namedrefcount = 0; cd->start_code = cworkspace; cd->hwm = cworkspace; @@ -9336,6 +9463,16 @@ if (cd->hwm > cd->start_workspace) int offset, recno; cd->hwm -= LINK_SIZE; offset = GET(cd->hwm, 0); + + /* Check that the hwm handling hasn't gone wrong. This whole area is + rewritten in PCRE2 because there are some obscure cases. */ + + if (offset == 0 || codestart[offset-1] != OP_RECURSE) + { + errorcode = ERR10; + break; + } + recno = GET(codestart, offset); if (recno != prev_recno) { @@ -9366,7 +9503,7 @@ used in this code because at least one c "const" attribute if the cast (pcre_uchar *)codestart is used directly in the function call. */ -if ((options & PCRE_NO_AUTO_POSSESS) == 0) +if (errorcode == 0 && (options & PCRE_NO_AUTO_POSSESS) == 0) { pcre_uchar *temp = (pcre_uchar *)codestart; auto_possessify(temp, utf, cd); @@ -9380,7 +9517,7 @@ OP_RECURSE that are not fixed length get exceptional ones forgo this. We scan the pattern to check that they are fixed length, and set their lengths. */ -if (cd->check_lookbehind) +if (errorcode == 0 && cd->check_lookbehind) { pcre_uchar *cc = (pcre_uchar *)codestart; @@ -9593,4 +9730,3 @@ return (pcre32 *)re; } /* End of pcre_compile.c */ - Modified: tomcat/jk/trunk/native/iis/pcre/pcre_exec.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcre_exec.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcre_exec.c (original) +++ tomcat/jk/trunk/native/iis/pcre/pcre_exec.c Sun Jan 17 17:23:28 2016 @@ -6685,7 +6685,8 @@ if (md->offset_vector != NULL) register int *iend = iptr - re->top_bracket; if (iend < md->offset_vector + 2) iend = md->offset_vector + 2; while (--iptr >= iend) *iptr = -1; - md->offset_vector[0] = md->offset_vector[1] = -1; + if (offsetcount > 0) md->offset_vector[0] = -1; + if (offsetcount > 1) md->offset_vector[1] = -1; } /* Set up the first character to match, if available. The first_char value is Modified: tomcat/jk/trunk/native/iis/pcre/pcre_internal.h URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcre_internal.h?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcre_internal.h (original) +++ tomcat/jk/trunk/native/iis/pcre/pcre_internal.h Sun Jan 17 17:23:28 2016 @@ -984,7 +984,7 @@ other. NOTE: The values also appear in p #ifndef EBCDIC #define HSPACE_LIST \ - CHAR_HT, CHAR_SPACE, 0xa0, \ + CHAR_HT, CHAR_SPACE, CHAR_NBSP, \ 0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \ 0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \ NOTACHAR @@ -1010,7 +1010,7 @@ other. NOTE: The values also appear in p #define HSPACE_BYTE_CASES \ case CHAR_HT: \ case CHAR_SPACE: \ - case 0xa0 /* NBSP */ + case CHAR_NBSP #define HSPACE_CASES \ HSPACE_BYTE_CASES: \ @@ -1037,11 +1037,12 @@ other. NOTE: The values also appear in p /* ------ EBCDIC environments ------ */ #else -#define HSPACE_LIST CHAR_HT, CHAR_SPACE +#define HSPACE_LIST CHAR_HT, CHAR_SPACE, CHAR_NBSP, NOTACHAR #define HSPACE_BYTE_CASES \ case CHAR_HT: \ - case CHAR_SPACE + case CHAR_SPACE: \ + case CHAR_NBSP #define HSPACE_CASES HSPACE_BYTE_CASES @@ -1215,6 +1216,7 @@ same code point. */ #define CHAR_ESC '\047' #define CHAR_DEL '\007' +#define CHAR_NBSP '\x41' #define STR_ESC "\047" #define STR_DEL "\007" @@ -1229,6 +1231,7 @@ a positive value. */ #define CHAR_NEL ((unsigned char)'\x85') #define CHAR_ESC '\033' #define CHAR_DEL '\177' +#define CHAR_NBSP ((unsigned char)'\xa0') #define STR_LF "\n" #define STR_NL STR_LF @@ -1606,6 +1609,7 @@ only. */ #define CHAR_VERTICAL_LINE '\174' #define CHAR_RIGHT_CURLY_BRACKET '\175' #define CHAR_TILDE '\176' +#define CHAR_NBSP ((unsigned char)'\xa0') #define STR_HT "\011" #define STR_VT "\013" @@ -1762,6 +1766,10 @@ only. */ /* Escape items that are just an encoding of a particular data value. */ +#ifndef ESC_a +#define ESC_a CHAR_BEL +#endif + #ifndef ESC_e #define ESC_e CHAR_ESC #endif @@ -2446,6 +2454,7 @@ typedef struct compile_data { BOOL had_pruneorskip; /* (*PRUNE) or (*SKIP) encountered */ BOOL check_lookbehind; /* Lookbehinds need later checking */ BOOL dupnames; /* Duplicate names exist */ + BOOL dupgroups; /* Duplicate groups exist: (?| found */ BOOL iscondassert; /* Next assert is a condition */ int nltype; /* Newline type */ int nllen; /* Newline string length */ Modified: tomcat/jk/trunk/native/iis/pcre/pcre_jit_compile.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcre_jit_compile.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcre_jit_compile.c (original) +++ tomcat/jk/trunk/native/iis/pcre/pcre_jit_compile.c Sun Jan 17 17:23:28 2016 @@ -1064,6 +1064,7 @@ pcre_uchar *alternative; pcre_uchar *end = NULL; int private_data_ptr = *private_data_start; int space, size, bracketlen; +BOOL repeat_check = TRUE; while (cc < ccend) { @@ -1071,9 +1072,10 @@ while (cc < ccend) size = 0; bracketlen = 0; if (private_data_ptr > SLJIT_MAX_LOCAL_SIZE) - return; + break; - if (*cc == OP_ONCE || *cc == OP_ONCE_NC || *cc == OP_BRA || *cc == OP_CBRA || *cc == OP_COND) + if (repeat_check && (*cc == OP_ONCE || *cc == OP_ONCE_NC || *cc == OP_BRA || *cc == OP_CBRA || *cc == OP_COND)) + { if (detect_repeat(common, cc)) { /* These brackets are converted to repeats, so no global @@ -1081,6 +1083,8 @@ while (cc < ccend) if (cc >= end) end = bracketend(cc); } + } + repeat_check = TRUE; switch(*cc) { @@ -1136,6 +1140,13 @@ while (cc < ccend) bracketlen = 1 + LINK_SIZE + IMM2_SIZE; break; + case OP_BRAZERO: + case OP_BRAMINZERO: + case OP_BRAPOSZERO: + repeat_check = FALSE; + size = 1; + break; + CASE_ITERATOR_PRIVATE_DATA_1 space = 1; size = -2; @@ -1162,12 +1173,17 @@ while (cc < ccend) size = 1; break; - CASE_ITERATOR_TYPE_PRIVATE_DATA_2B + case OP_TYPEUPTO: if (cc[1 + IMM2_SIZE] != OP_ANYNL && cc[1 + IMM2_SIZE] != OP_EXTUNI) space = 2; size = 1 + IMM2_SIZE; break; + case OP_TYPEMINUPTO: + space = 2; + size = 1 + IMM2_SIZE; + break; + case OP_CLASS: case OP_NCLASS: size += 1 + 32 / sizeof(pcre_uchar); @@ -1316,6 +1332,13 @@ while (cc < ccend) cc += 1 + LINK_SIZE + IMM2_SIZE; break; + case OP_THEN: + stack_restore = TRUE; + if (common->control_head_ptr != 0) + *needs_control_head = TRUE; + cc ++; + break; + default: stack_restore = TRUE; /* Fall through. */ @@ -2220,6 +2243,7 @@ while (current != NULL) SLJIT_ASSERT_STOP(); break; } + SLJIT_ASSERT(current > (sljit_sw*)current[-1]); current = (sljit_sw*)current[-1]; } return -1; @@ -3209,7 +3233,7 @@ bytes[len] = byte; bytes[0] = len; } -static int scan_prefix(compiler_common *common, pcre_uchar *cc, pcre_uint32 *chars, pcre_uint8 *bytes, int max_chars) +static int scan_prefix(compiler_common *common, pcre_uchar *cc, pcre_uint32 *chars, pcre_uint8 *bytes, int max_chars, pcre_uint32 *rec_count) { /* Recursive function, which scans prefix literals. */ BOOL last, any, caseless; @@ -3227,9 +3251,14 @@ pcre_uchar othercase[1]; repeat = 1; while (TRUE) { + if (*rec_count == 0) + return 0; + (*rec_count)--; + last = TRUE; any = FALSE; caseless = FALSE; + switch (*cc) { case OP_CHARI: @@ -3291,7 +3320,7 @@ while (TRUE) #ifdef SUPPORT_UTF if (common->utf && HAS_EXTRALEN(*cc)) len += GET_EXTRALEN(*cc); #endif - max_chars = scan_prefix(common, cc + len, chars, bytes, max_chars); + max_chars = scan_prefix(common, cc + len, chars, bytes, max_chars, rec_count); if (max_chars == 0) return consumed; last = FALSE; @@ -3314,7 +3343,7 @@ while (TRUE) alternative = cc + GET(cc, 1); while (*alternative == OP_ALT) { - max_chars = scan_prefix(common, alternative + 1 + LINK_SIZE, chars, bytes, max_chars); + max_chars = scan_prefix(common, alternative + 1 + LINK_SIZE, chars, bytes, max_chars, rec_count); if (max_chars == 0) return consumed; alternative += GET(alternative, 1); @@ -3556,6 +3585,7 @@ int i, max, from; int range_right = -1, range_len = 3 - 1; sljit_ub *update_table = NULL; BOOL in_range; +pcre_uint32 rec_count; for (i = 0; i < MAX_N_CHARS; i++) { @@ -3564,7 +3594,8 @@ for (i = 0; i < MAX_N_CHARS; i++) bytes[i * MAX_N_BYTES] = 0; } -max = scan_prefix(common, common->start, chars, bytes, MAX_N_CHARS); +rec_count = 10000; +max = scan_prefix(common, common->start, chars, bytes, MAX_N_CHARS, &rec_count); if (max <= 1) return FALSE; @@ -4311,8 +4342,10 @@ switch(length) case 4: if ((ranges[1] - ranges[0]) == (ranges[3] - ranges[2]) && (ranges[0] | (ranges[2] - ranges[0])) == ranges[2] + && (ranges[1] & (ranges[2] - ranges[0])) == 0 && is_powerof2(ranges[2] - ranges[0])) { + SLJIT_ASSERT((ranges[0] & (ranges[2] - ranges[0])) == 0 && (ranges[2] & ranges[3] & (ranges[2] - ranges[0])) != 0); OP2(SLJIT_OR, TMP1, 0, TMP1, 0, SLJIT_IMM, ranges[2] - ranges[0]); if (ranges[2] + 1 != ranges[3]) { @@ -4900,9 +4933,10 @@ else if ((cc[-1] & XCL_MAP) != 0) if (!check_class_ranges(common, (const pcre_uint8 *)cc, FALSE, TRUE, list)) { #ifdef COMPILE_PCRE8 - SLJIT_ASSERT(common->utf); + jump = NULL; + if (common->utf) #endif - jump = CMP(SLJIT_GREATER, TMP1, 0, SLJIT_IMM, 255); + jump = CMP(SLJIT_GREATER, TMP1, 0, SLJIT_IMM, 255); OP2(SLJIT_AND, TMP2, 0, TMP1, 0, SLJIT_IMM, 0x7); OP2(SLJIT_LSHR, TMP1, 0, TMP1, 0, SLJIT_IMM, 3); @@ -4911,7 +4945,10 @@ else if ((cc[-1] & XCL_MAP) != 0) OP2(SLJIT_AND | SLJIT_SET_E, SLJIT_UNUSED, 0, TMP1, 0, TMP2, 0); add_jump(compiler, list, JUMP(SLJIT_NOT_ZERO)); - JUMPHERE(jump); +#ifdef COMPILE_PCRE8 + if (common->utf) +#endif + JUMPHERE(jump); } OP1(SLJIT_MOV, TMP1, 0, TMP3, 0); @@ -5219,7 +5256,7 @@ while (*cc != XCL_END) OP_FLAGS(SLJIT_MOV, TMP2, 0, SLJIT_UNUSED, 0, SLJIT_LESS_EQUAL); SET_CHAR_OFFSET(0); - OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0xff); + OP2(SLJIT_SUB | SLJIT_SET_U, SLJIT_UNUSED, 0, TMP1, 0, SLJIT_IMM, 0x7f); OP_FLAGS(SLJIT_AND, TMP2, 0, TMP2, 0, SLJIT_LESS_EQUAL); SET_TYPE_OFFSET(ucp_Pc); @@ -7665,6 +7702,10 @@ while (*cc != OP_KETRPOS) OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(0), STR_PTR, 0); } + /* Even if the match is empty, we need to reset the control head. */ + if (needs_control_head) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_SP), common->control_head_ptr, SLJIT_MEM1(STACK_TOP), STACK(stack)); + if (opcode == OP_SBRAPOS || opcode == OP_SCBRAPOS) add_jump(compiler, &emptymatch, CMP(SLJIT_EQUAL, TMP1, 0, STR_PTR, 0)); @@ -7692,6 +7733,10 @@ while (*cc != OP_KETRPOS) OP1(SLJIT_MOV, SLJIT_MEM1(TMP2), (framesize + 1) * sizeof(sljit_sw), STR_PTR, 0); } + /* Even if the match is empty, we need to reset the control head. */ + if (needs_control_head) + OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_SP), common->control_head_ptr, SLJIT_MEM1(STACK_TOP), STACK(stack)); + if (opcode == OP_SBRAPOS || opcode == OP_SCBRAPOS) add_jump(compiler, &emptymatch, CMP(SLJIT_EQUAL, TMP1, 0, STR_PTR, 0)); @@ -7704,9 +7749,6 @@ while (*cc != OP_KETRPOS) } } - if (needs_control_head) - OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_SP), common->control_head_ptr, SLJIT_MEM1(STACK_TOP), STACK(stack)); - JUMPTO(SLJIT_JUMP, loop); flush_stubs(common); @@ -8441,8 +8483,7 @@ while (cc < ccend) OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(1), STR_PTR, 0); } BACKTRACK_AS(braminzero_backtrack)->matchingpath = LABEL(); - if (cc[1] > OP_ASSERTBACK_NOT) - count_match(common); + count_match(common); break; case OP_ONCE: @@ -9624,7 +9665,7 @@ static SLJIT_INLINE void compile_recurse DEFINE_COMPILER; pcre_uchar *cc = common->start + common->currententry->start; pcre_uchar *ccbegin = cc + 1 + LINK_SIZE + (*cc == OP_BRA ? 0 : IMM2_SIZE); -pcre_uchar *ccend = bracketend(cc); +pcre_uchar *ccend = bracketend(cc) - (1 + LINK_SIZE); BOOL needs_control_head; int framesize = get_framesize(common, cc, NULL, TRUE, &needs_control_head); int private_data_size = get_private_data_copy_length(common, ccbegin, ccend, needs_control_head); @@ -9648,6 +9689,7 @@ set_jumps(common->currententry->calls, c sljit_emit_fast_enter(compiler, TMP2, 0); allocate_stack(common, private_data_size + framesize + alternativesize); +count_match(common); OP1(SLJIT_MOV, SLJIT_MEM1(STACK_TOP), STACK(private_data_size + framesize + alternativesize - 1), TMP2, 0); copy_private_data(common, ccbegin, ccend, TRUE, private_data_size + framesize + alternativesize, framesize + alternativesize, needs_control_head); if (needs_control_head) @@ -9992,6 +10034,7 @@ OP1(SLJIT_MOV, TMP2, 0, SLJIT_MEM1(TMP1) OP1(SLJIT_MOV_UI, TMP1, 0, SLJIT_MEM1(TMP1), SLJIT_OFFSETOF(jit_arguments, limit_match)); OP1(SLJIT_MOV, STACK_TOP, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(struct sljit_stack, base)); OP1(SLJIT_MOV, STACK_LIMIT, 0, SLJIT_MEM1(TMP2), SLJIT_OFFSETOF(struct sljit_stack, limit)); +OP2(SLJIT_ADD, TMP1, 0, TMP1, 0, SLJIT_IMM, 1); OP1(SLJIT_MOV, SLJIT_MEM1(SLJIT_SP), LIMIT_MATCH, TMP1, 0); if (mode == JIT_PARTIAL_SOFT_COMPILE) Modified: tomcat/jk/trunk/native/iis/pcre/pcre_jit_test.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcre_jit_test.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcre_jit_test.c (original) +++ tomcat/jk/trunk/native/iis/pcre/pcre_jit_test.c Sun Jan 17 17:23:28 2016 @@ -182,6 +182,7 @@ static struct regression_test_case regre { CMUAP, 0, "\xf0\x90\x90\x80{2}", "\xf0\x90\x90\x80#\xf0\x90\x90\xa8\xf0\x90\x90\x80" }, { CMUAP, 0, "\xf0\x90\x90\xa8{2}", "\xf0\x90\x90\x80#\xf0\x90\x90\xa8\xf0\x90\x90\x80" }, { CMUAP, 0, "\xe1\xbd\xb8\xe1\xbf\xb8", "\xe1\xbf\xb8\xe1\xbd\xb8" }, + { MA, 0, "[3-57-9]", "5" }, /* Assertions. */ { MUA, 0, "\\b[^A]", "A_B#" }, Modified: tomcat/jk/trunk/native/iis/pcre/pcre_study.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcre_study.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcre_study.c (original) +++ tomcat/jk/trunk/native/iis/pcre/pcre_study.c Sun Jan 17 17:23:28 2016 @@ -71,6 +71,7 @@ Arguments: startcode pointer to start of the whole pattern's code options the compiling options recurses chain of recurse_check to catch mutual recursion + countptr pointer to call count (to catch over complexity) Returns: the minimum length -1 if \C in UTF-8 mode or (*ACCEPT) was encountered @@ -80,7 +81,8 @@ Returns: the minimum length static int find_minlength(const REAL_PCRE *re, const pcre_uchar *code, - const pcre_uchar *startcode, int options, recurse_check *recurses) + const pcre_uchar *startcode, int options, recurse_check *recurses, + int *countptr) { int length = -1; /* PCRE_UTF16 has the same value as PCRE_UTF8. */ @@ -90,6 +92,8 @@ recurse_check this_recurse; register int branchlength = 0; register pcre_uchar *cc = (pcre_uchar *)code + 1 + LINK_SIZE; +if ((*countptr)++ > 1000) return -1; /* too complex */ + if (*code == OP_CBRA || *code == OP_SCBRA || *code == OP_CBRAPOS || *code == OP_SCBRAPOS) cc += IMM2_SIZE; @@ -131,7 +135,7 @@ for (;;) case OP_SBRAPOS: case OP_ONCE: case OP_ONCE_NC: - d = find_minlength(re, cc, startcode, options, recurses); + d = find_minlength(re, cc, startcode, options, recurses, countptr); if (d < 0) return d; branchlength += d; do cc += GET(cc, 1); while (*cc == OP_ALT); @@ -415,7 +419,8 @@ for (;;) int dd; this_recurse.prev = recurses; this_recurse.group = cs; - dd = find_minlength(re, cs, startcode, options, &this_recurse); + dd = find_minlength(re, cs, startcode, options, &this_recurse, + countptr); if (dd < d) d = dd; } } @@ -451,7 +456,8 @@ for (;;) { this_recurse.prev = recurses; this_recurse.group = cs; - d = find_minlength(re, cs, startcode, options, &this_recurse); + d = find_minlength(re, cs, startcode, options, &this_recurse, + countptr); } } } @@ -514,7 +520,7 @@ for (;;) this_recurse.prev = recurses; this_recurse.group = cs; branchlength += find_minlength(re, cs, startcode, options, - &this_recurse); + &this_recurse, countptr); } } cc += 1 + LINK_SIZE; @@ -1453,6 +1459,7 @@ pcre32_study(const pcre32 *external_re, #endif { int min; +int count = 0; BOOL bits_set = FALSE; pcre_uint8 start_bits[32]; PUBL(extra) *extra = NULL; @@ -1539,7 +1546,7 @@ if ((re->options & PCRE_ANCHORED) == 0 & /* Find the minimum length of subject string. */ -switch(min = find_minlength(re, code, code, re->options, NULL)) +switch(min = find_minlength(re, code, code, re->options, NULL, &count)) { case -2: *errorptr = "internal error: missing capturing bracket"; return NULL; case -3: *errorptr = "internal error: opcode not recognized"; return NULL; Modified: tomcat/jk/trunk/native/iis/pcre/pcre_xclass.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcre_xclass.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcre_xclass.c (original) +++ tomcat/jk/trunk/native/iis/pcre/pcre_xclass.c Sun Jan 17 17:23:28 2016 @@ -246,7 +246,7 @@ while ((t = *data++) != XCL_END) case PT_PXPUNCT: if ((PRIV(ucp_gentype)[prop->chartype] == ucp_P || - (c < 256 && PRIV(ucp_gentype)[prop->chartype] == ucp_S)) == isprop) + (c < 128 && PRIV(ucp_gentype)[prop->chartype] == ucp_S)) == isprop) return !negated; break; Modified: tomcat/jk/trunk/native/iis/pcre/pcregrep.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcregrep.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcregrep.c (original) +++ tomcat/jk/trunk/native/iis/pcre/pcregrep.c Sun Jan 17 17:23:28 2016 @@ -1692,9 +1692,13 @@ while (ptr < endptr) if (filenames == FN_NOMATCH_ONLY) return 1; + /* If all we want is a yes/no answer, stop now. */ + + if (quiet) return 0; + /* Just count if just counting is wanted. */ - if (count_only) count++; + else if (count_only) count++; /* When handling a binary file and binary-files==binary, the "binary" variable will be set true (it's false in all other cases). In this @@ -1715,10 +1719,6 @@ while (ptr < endptr) return 0; } - /* Likewise, if all we want is a yes/no answer. */ - - else if (quiet) return 0; - /* The --only-matching option prints just the substring that matched, and/or one or more captured portions of it, as long as these strings are not empty. The --file-offsets and --line-offsets options output offsets for @@ -2089,7 +2089,7 @@ if (filenames == FN_NOMATCH_ONLY) /* Print the match count if wanted */ -if (count_only) +if (count_only && !quiet) { if (count > 0 || !omit_zero_count) { Modified: tomcat/jk/trunk/native/iis/pcre/pcretest.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/pcretest.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/pcretest.c (original) +++ tomcat/jk/trunk/native/iis/pcre/pcretest.c Sun Jan 17 17:23:28 2016 @@ -4621,9 +4621,9 @@ while (!done) else switch ((c = *p++)) { - case 'a': c = 7; break; + case 'a': c = CHAR_BEL; break; case 'b': c = '\b'; break; - case 'e': c = 27; break; + case 'e': c = CHAR_ESC; break; case 'f': c = '\f'; break; case 'n': c = '\n'; break; case 'r': c = '\r'; break; Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitConfig.h URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitConfig.h?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitConfig.h (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitConfig.h Sun Jan 17 17:23:28 2016 @@ -96,6 +96,15 @@ #define SLJIT_EXECUTABLE_ALLOCATOR 1 #endif +/* Force cdecl calling convention even if a better calling + convention (e.g. fastcall) is supported by the C compiler. + If this option is enabled, C functions without + SLJIT_CALL can also be called from JIT code. */ +#ifndef SLJIT_USE_CDECL_CALLING_CONVENTION +/* Disabled by default */ +#define SLJIT_USE_CDECL_CALLING_CONVENTION 0 +#endif + /* Return with error when an invalid argument is passed. */ #ifndef SLJIT_ARGUMENT_CHECKS /* Disabled by default */ Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitConfigInternal.h URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitConfigInternal.h?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitConfigInternal.h (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitConfigInternal.h Sun Jan 17 17:23:28 2016 @@ -468,7 +468,12 @@ typedef double sljit_d; #ifndef SLJIT_CALL -#if (defined SLJIT_CONFIG_X86_32 && SLJIT_CONFIG_X86_32) +#if (defined SLJIT_USE_CDECL_CALLING_CONVENTION && SLJIT_USE_CDECL_CALLING_CONVENTION) + +/* Force cdecl. */ +#define SLJIT_CALL + +#elif (defined SLJIT_CONFIG_X86_32 && SLJIT_CONFIG_X86_32) #if defined(__GNUC__) && !defined(__APPLE__) @@ -608,6 +613,12 @@ SLJIT_API_FUNC_ATTRIBUTE void sljit_free #define SLJIT_LOCALS_OFFSET_BASE ((23 + 1) * sizeof(sljit_sw)) #endif +#elif (defined SLJIT_CONFIG_TILEGX && SLJIT_CONFIG_TILEGX) + +#define SLJIT_NUMBER_OF_REGISTERS 10 +#define SLJIT_NUMBER_OF_SAVED_REGISTERS 5 +#define SLJIT_LOCALS_OFFSET_BASE 0 + #elif (defined SLJIT_CONFIG_UNSUPPORTED && SLJIT_CONFIG_UNSUPPORTED) #define SLJIT_NUMBER_OF_REGISTERS 0 Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitLir.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitLir.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitLir.c (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitLir.c Sun Jan 17 17:23:28 2016 @@ -845,8 +845,8 @@ SLJIT_API_FUNC_ATTRIBUTE void sljit_comp } static SLJIT_CONST char* op0_names[] = { - (char*)"breakpoint", (char*)"nop", - (char*)"lumul", (char*)"lsmul", (char*)"ludiv", (char*)"lsdiv", + (char*)"breakpoint", (char*)"nop", (char*)"lumul", (char*)"lsmul", + (char*)"udivmod", (char*)"sdivmod", (char*)"udivi", (char*)"sdivi" }; static SLJIT_CONST char* op1_names[] = { @@ -1036,7 +1036,7 @@ static SLJIT_INLINE CHECK_RETURN_TYPE ch { #if (defined SLJIT_ARGUMENT_CHECKS && SLJIT_ARGUMENT_CHECKS) CHECK_ARGUMENT((op >= SLJIT_BREAKPOINT && op <= SLJIT_LSMUL) - || ((op & ~SLJIT_INT_OP) >= SLJIT_LUDIV && (op & ~SLJIT_INT_OP) <= SLJIT_LSDIV)); + || ((op & ~SLJIT_INT_OP) >= SLJIT_UDIVMOD && (op & ~SLJIT_INT_OP) <= SLJIT_SDIVI)); CHECK_ARGUMENT(op < SLJIT_LUMUL || compiler->scratches >= 2); #endif #if (defined SLJIT_VERBOSE && SLJIT_VERBOSE) @@ -1447,6 +1447,8 @@ static SLJIT_INLINE CHECK_RETURN_TYPE ch static SLJIT_INLINE CHECK_RETURN_TYPE check_sljit_get_local_base(struct sljit_compiler *compiler, sljit_si dst, sljit_sw dstw, sljit_sw offset) { + SLJIT_UNUSED_ARG(offset); + #if (defined SLJIT_ARGUMENT_CHECKS && SLJIT_ARGUMENT_CHECKS) FUNCTION_CHECK_DST(dst, dstw); #endif @@ -1462,6 +1464,8 @@ static SLJIT_INLINE CHECK_RETURN_TYPE ch static SLJIT_INLINE CHECK_RETURN_TYPE check_sljit_emit_const(struct sljit_compiler *compiler, sljit_si dst, sljit_sw dstw, sljit_sw init_value) { + SLJIT_UNUSED_ARG(init_value); + #if (defined SLJIT_ARGUMENT_CHECKS && SLJIT_ARGUMENT_CHECKS) FUNCTION_CHECK_DST(dst, dstw); #endif Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitLir.h URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitLir.h?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitLir.h (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitLir.h Sun Jan 17 17:23:28 2016 @@ -687,7 +687,7 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ #define SLJIT_OP0_BASE 0 /* Flags: - (never set any flags) - Note: breakpoint instruction is not supported by all architectures (namely ppc) + Note: breakpoint instruction is not supported by all architectures (e.g. ppc) It falls back to SLJIT_NOP in those cases. */ #define SLJIT_BREAKPOINT (SLJIT_OP0_BASE + 0) /* Flags: - (never set any flags) @@ -696,24 +696,42 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ #define SLJIT_NOP (SLJIT_OP0_BASE + 1) /* Flags: - (may destroy flags) Unsigned multiplication of SLJIT_R0 and SLJIT_R1. - Result goes to SLJIT_R1:SLJIT_R0 (high:low) word */ + Result is placed into SLJIT_R1:SLJIT_R0 (high:low) word */ #define SLJIT_LUMUL (SLJIT_OP0_BASE + 2) /* Flags: - (may destroy flags) Signed multiplication of SLJIT_R0 and SLJIT_R1. - Result goes to SLJIT_R1:SLJIT_R0 (high:low) word */ + Result is placed into SLJIT_R1:SLJIT_R0 (high:low) word */ #define SLJIT_LSMUL (SLJIT_OP0_BASE + 3) /* Flags: I - (may destroy flags) Unsigned divide of the value in SLJIT_R0 by the value in SLJIT_R1. - The result is placed in SLJIT_R0 and the remainder goes to SLJIT_R1. - Note: if SLJIT_R1 contains 0, the behaviour is undefined. */ -#define SLJIT_LUDIV (SLJIT_OP0_BASE + 4) -#define SLJIT_ILUDIV (SLJIT_LUDIV | SLJIT_INT_OP) + The result is placed into SLJIT_R0 and the remainder into SLJIT_R1. + Note: if SLJIT_R1 is 0, the behaviour is undefined. */ +#define SLJIT_UDIVMOD (SLJIT_OP0_BASE + 4) +#define SLJIT_IUDIVMOD (SLJIT_UDIVMOD | SLJIT_INT_OP) /* Flags: I - (may destroy flags) Signed divide of the value in SLJIT_R0 by the value in SLJIT_R1. - The result is placed in SLJIT_R0 and the remainder goes to SLJIT_R1. - Note: if SLJIT_R1 contains 0, the behaviour is undefined. */ -#define SLJIT_LSDIV (SLJIT_OP0_BASE + 5) -#define SLJIT_ILSDIV (SLJIT_LSDIV | SLJIT_INT_OP) + The result is placed into SLJIT_R0 and the remainder into SLJIT_R1. + Note: if SLJIT_R1 is 0, the behaviour is undefined. + Note: if SLJIT_R1 is -1 and SLJIT_R0 is integer min (0x800..00), + the behaviour is undefined. */ +#define SLJIT_SDIVMOD (SLJIT_OP0_BASE + 5) +#define SLJIT_ISDIVMOD (SLJIT_SDIVMOD | SLJIT_INT_OP) +/* Flags: I - (may destroy flags) + Unsigned divide of the value in SLJIT_R0 by the value in SLJIT_R1. + The result is placed into SLJIT_R0. SLJIT_R1 preserves its value. + Note: if SLJIT_R1 is 0, the behaviour is undefined. + Note: SLJIT_SDIV is single precision divide. */ +#define SLJIT_UDIVI (SLJIT_OP0_BASE + 6) +#define SLJIT_IUDIVI (SLJIT_UDIVI | SLJIT_INT_OP) +/* Flags: I - (may destroy flags) + Signed divide of the value in SLJIT_R0 by the value in SLJIT_R1. + The result is placed into SLJIT_R0. SLJIT_R1 preserves its value. + Note: if SLJIT_R1 is 0, the behaviour is undefined. + Note: if SLJIT_R1 is -1 and SLJIT_R0 is integer min (0x800..00), + the behaviour is undefined. + Note: SLJIT_SDIV is single precision divide. */ +#define SLJIT_SDIVI (SLJIT_OP0_BASE + 7) +#define SLJIT_ISDIVI (SLJIT_SDIVI | SLJIT_INT_OP) SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_emit_op0(struct sljit_compiler *compiler, sljit_si op); @@ -851,34 +869,6 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ sljit_si src1, sljit_sw src1w, sljit_si src2, sljit_sw src2w); -/* The following function is a helper function for sljit_emit_op_custom. - It returns with the real machine register index ( >=0 ) of any SLJIT_R, - SLJIT_S and SLJIT_SP registers. - - Note: it returns with -1 for virtual registers (only on x86-32). */ - -SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_get_register_index(sljit_si reg); - -/* The following function is a helper function for sljit_emit_op_custom. - It returns with the real machine register index of any SLJIT_FLOAT register. - - Note: the index is always an even number on ARM (except ARM-64), MIPS, and SPARC. */ - -SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_get_float_register_index(sljit_si reg); - -/* Any instruction can be inserted into the instruction stream by - sljit_emit_op_custom. It has a similar purpose as inline assembly. - The size parameter must match to the instruction size of the target - architecture: - - x86: 0 < size <= 15. The instruction argument can be byte aligned. - Thumb2: if size == 2, the instruction argument must be 2 byte aligned. - if size == 4, the instruction argument must be 4 byte aligned. - Otherwise: size must be 4 and instruction argument must be 4 byte aligned. */ - -SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_emit_op_custom(struct sljit_compiler *compiler, - void *instruction, sljit_si size); - /* Returns with non-zero if fpu is available. */ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_is_fpu_available(void); @@ -1196,4 +1186,64 @@ SLJIT_API_FUNC_ATTRIBUTE void sljit_set_ #endif /* !(defined SLJIT_INDIRECT_CALL && SLJIT_INDIRECT_CALL) */ +/* --------------------------------------------------------------------- */ +/* CPU specific functions */ +/* --------------------------------------------------------------------- */ + +/* The following function is a helper function for sljit_emit_op_custom. + It returns with the real machine register index ( >=0 ) of any SLJIT_R, + SLJIT_S and SLJIT_SP registers. + + Note: it returns with -1 for virtual registers (only on x86-32). */ + +SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_get_register_index(sljit_si reg); + +/* The following function is a helper function for sljit_emit_op_custom. + It returns with the real machine register index of any SLJIT_FLOAT register. + + Note: the index is always an even number on ARM (except ARM-64), MIPS, and SPARC. */ + +SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_get_float_register_index(sljit_si reg); + +/* Any instruction can be inserted into the instruction stream by + sljit_emit_op_custom. It has a similar purpose as inline assembly. + The size parameter must match to the instruction size of the target + architecture: + + x86: 0 < size <= 15. The instruction argument can be byte aligned. + Thumb2: if size == 2, the instruction argument must be 2 byte aligned. + if size == 4, the instruction argument must be 4 byte aligned. + Otherwise: size must be 4 and instruction argument must be 4 byte aligned. */ + +SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_emit_op_custom(struct sljit_compiler *compiler, + void *instruction, sljit_si size); + +#if (defined SLJIT_CONFIG_X86 && SLJIT_CONFIG_X86) + +/* Returns with non-zero if sse2 is available. */ + +SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_x86_is_sse2_available(void); + +/* Returns with non-zero if cmov instruction is available. */ + +SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_x86_is_cmov_available(void); + +/* Emit a conditional mov instruction on x86 CPUs. This instruction + moves src to destination, if the condition is satisfied. Unlike + other arithmetic instructions, destination must be a register. + Before such instructions are emitted, cmov support should be + checked by sljit_x86_is_cmov_available function. + type must be between SLJIT_EQUAL and SLJIT_S_ORDERED + dst_reg must be a valid register and it can be combined + with SLJIT_INT_OP to perform 32 bit arithmetic + Flags: I - (never set any flags) + */ + +SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_x86_emit_cmov(struct sljit_compiler *compiler, + sljit_si type, + sljit_si dst_reg, + sljit_si src, sljit_sw srcw); + +#endif + #endif /* _SLJIT_LIR_H_ */ Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_32.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_32.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_32.c (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_32.c Sun Jan 17 17:23:28 2016 @@ -1833,18 +1833,33 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ | (reg_map[SLJIT_R0] << 8) | reg_map[TMP_REG1]); #endif - case SLJIT_LUDIV: - case SLJIT_LSDIV: - if (compiler->scratches >= 3) + case SLJIT_UDIVMOD: + case SLJIT_SDIVMOD: + case SLJIT_UDIVI: + case SLJIT_SDIVI: + SLJIT_COMPILE_ASSERT((SLJIT_UDIVMOD & 0x2) == 0 && SLJIT_UDIVI - 0x2 == SLJIT_UDIVMOD, bad_div_opcode_assignments); + SLJIT_COMPILE_ASSERT(reg_map[2] == 1 && reg_map[3] == 2, bad_register_mapping); + + if ((op >= SLJIT_UDIVI) && (compiler->scratches >= 3)) { FAIL_IF(push_inst(compiler, 0xe52d2008 /* str r2, [sp, #-8]! */)); + FAIL_IF(push_inst(compiler, 0xe58d1004 /* str r1, [sp, #4] */)); + } + else if ((op >= SLJIT_UDIVI) || (compiler->scratches >= 3)) + FAIL_IF(push_inst(compiler, 0xe52d0008 | (op >= SLJIT_UDIVI ? 0x1000 : 0x2000) /* str r1/r2, [sp, #-8]! */)); + #if defined(__GNUC__) FAIL_IF(sljit_emit_ijump(compiler, SLJIT_FAST_CALL, SLJIT_IMM, - (op == SLJIT_LUDIV ? SLJIT_FUNC_OFFSET(__aeabi_uidivmod) : SLJIT_FUNC_OFFSET(__aeabi_idivmod)))); + ((op | 0x2) == SLJIT_UDIVI ? SLJIT_FUNC_OFFSET(__aeabi_uidivmod) : SLJIT_FUNC_OFFSET(__aeabi_idivmod)))); #else #error "Software divmod functions are needed" #endif - if (compiler->scratches >= 3) - return push_inst(compiler, 0xe49d2008 /* ldr r2, [sp], #8 */); + + if ((op >= SLJIT_UDIVI) && (compiler->scratches >= 3)) { + FAIL_IF(push_inst(compiler, 0xe59d1004 /* ldr r1, [sp, #4] */)); + FAIL_IF(push_inst(compiler, 0xe49d2008 /* ldr r2, [sp], #8 */)); + } + else if ((op >= SLJIT_UDIVI) || (compiler->scratches >= 3)) + return push_inst(compiler, 0xe49d0008 | (op >= SLJIT_UDIVI ? 0x1000 : 0x2000) /* ldr r1/r2, [sp], #8 */); return SLJIT_SUCCESS; } Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_64.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_64.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_64.c (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_64.c Sun Jan 17 17:23:28 2016 @@ -1087,14 +1087,20 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ saved_regs_size += sizeof(sljit_sw); } local_size -= saved_regs_size + SLJIT_LOCALS_OFFSET; - FAIL_IF(push_inst(compiler, SUBI | RD(TMP_SP) | RN(TMP_SP) | (saved_regs_size << 10))); + if (saved_regs_size > 0) + FAIL_IF(push_inst(compiler, SUBI | RD(TMP_SP) | RN(TMP_SP) | (saved_regs_size << 10))); } tmp = saveds < SLJIT_NUMBER_OF_SAVED_REGISTERS ? (SLJIT_S0 + 1 - saveds) : SLJIT_FIRST_SAVED_REG; prev = -1; for (i = SLJIT_S0; i >= tmp; i--) { if (prev == -1) { - prev = i; + if (!(offs & (1 << 15))) { + prev = i; + continue; + } + FAIL_IF(push_inst(compiler, STRI | RT(i) | RN(TMP_SP) | (offs >> 5))); + offs += 1 << 15; continue; } FAIL_IF(push_inst(compiler, STP | RT(prev) | RT2(i) | RN(TMP_SP) | offs)); @@ -1104,7 +1110,12 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ for (i = scratches; i >= SLJIT_FIRST_SAVED_REG; i--) { if (prev == -1) { - prev = i; + if (!(offs & (1 << 15))) { + prev = i; + continue; + } + FAIL_IF(push_inst(compiler, STRI | RT(i) | RN(TMP_SP) | (offs >> 5))); + offs += 1 << 15; continue; } FAIL_IF(push_inst(compiler, STP | RT(prev) | RT2(i) | RN(TMP_SP) | offs)); @@ -1112,8 +1123,7 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ prev = -1; } - if (prev != -1) - FAIL_IF(push_inst(compiler, STRI | RT(prev) | RN(TMP_SP) | (offs >> 5))); + SLJIT_ASSERT(prev == -1); if (compiler->local_size > (63 * sizeof(sljit_sw))) { /* The local_size is already adjusted by the saved registers. */ @@ -1188,7 +1198,12 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ prev = -1; for (i = SLJIT_S0; i >= tmp; i--) { if (prev == -1) { - prev = i; + if (!(offs & (1 << 15))) { + prev = i; + continue; + } + FAIL_IF(push_inst(compiler, LDRI | RT(i) | RN(TMP_SP) | (offs >> 5))); + offs += 1 << 15; continue; } FAIL_IF(push_inst(compiler, LDP | RT(prev) | RT2(i) | RN(TMP_SP) | offs)); @@ -1198,7 +1213,12 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ for (i = compiler->scratches; i >= SLJIT_FIRST_SAVED_REG; i--) { if (prev == -1) { - prev = i; + if (!(offs & (1 << 15))) { + prev = i; + continue; + } + FAIL_IF(push_inst(compiler, LDRI | RT(i) | RN(TMP_SP) | (offs >> 5))); + offs += 1 << 15; continue; } FAIL_IF(push_inst(compiler, LDP | RT(prev) | RT2(i) | RN(TMP_SP) | offs)); @@ -1206,13 +1226,12 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ prev = -1; } - if (prev != -1) - FAIL_IF(push_inst(compiler, LDRI | RT(prev) | RN(TMP_SP) | (offs >> 5))); + SLJIT_ASSERT(prev == -1); if (compiler->local_size <= (63 * sizeof(sljit_sw))) { FAIL_IF(push_inst(compiler, LDP_PST | 29 | RT2(TMP_LR) | RN(TMP_SP) | (((local_size >> 3) & 0x7f) << 15))); - } else { + } else if (saved_regs_size > 0) { FAIL_IF(push_inst(compiler, ADDI | RD(TMP_SP) | RN(TMP_SP) | (saved_regs_size << 10))); } @@ -1242,12 +1261,15 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ FAIL_IF(push_inst(compiler, ORR | RD(TMP_REG1) | RN(TMP_ZERO) | RM(SLJIT_R0))); FAIL_IF(push_inst(compiler, MADD | RD(SLJIT_R0) | RN(SLJIT_R0) | RM(SLJIT_R1) | RT2(TMP_ZERO))); return push_inst(compiler, (op == SLJIT_LUMUL ? UMULH : SMULH) | RD(SLJIT_R1) | RN(TMP_REG1) | RM(SLJIT_R1)); - case SLJIT_LUDIV: - case SLJIT_LSDIV: + case SLJIT_UDIVMOD: + case SLJIT_SDIVMOD: FAIL_IF(push_inst(compiler, (ORR ^ inv_bits) | RD(TMP_REG1) | RN(TMP_ZERO) | RM(SLJIT_R0))); - FAIL_IF(push_inst(compiler, ((op == SLJIT_LUDIV ? UDIV : SDIV) ^ inv_bits) | RD(SLJIT_R0) | RN(SLJIT_R0) | RM(SLJIT_R1))); + FAIL_IF(push_inst(compiler, ((op == SLJIT_UDIVMOD ? UDIV : SDIV) ^ inv_bits) | RD(SLJIT_R0) | RN(SLJIT_R0) | RM(SLJIT_R1))); FAIL_IF(push_inst(compiler, (MADD ^ inv_bits) | RD(SLJIT_R1) | RN(SLJIT_R0) | RM(SLJIT_R1) | RT2(TMP_ZERO))); return push_inst(compiler, (SUB ^ inv_bits) | RD(SLJIT_R1) | RN(TMP_REG1) | RM(SLJIT_R1)); + case SLJIT_UDIVI: + case SLJIT_SDIVI: + return push_inst(compiler, ((op == SLJIT_UDIVI ? UDIV : SDIV) ^ inv_bits) | RD(SLJIT_R0) | RN(SLJIT_R0) | RM(SLJIT_R1)); } return SLJIT_SUCCESS; Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_T2_32.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_T2_32.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_T2_32.c (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeARM_T2_32.c Sun Jan 17 17:23:28 2016 @@ -1239,6 +1239,9 @@ extern int __aeabi_idivmod(int numerator SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_emit_op0(struct sljit_compiler *compiler, sljit_si op) { + sljit_sw saved_reg_list[3]; + sljit_sw saved_reg_count; + CHECK_ERROR(); CHECK(check_sljit_emit_op0(compiler, op)); @@ -1255,24 +1258,53 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ | (reg_map[SLJIT_R0] << 12) | (reg_map[SLJIT_R0] << 16) | reg_map[SLJIT_R1]); - case SLJIT_LUDIV: - case SLJIT_LSDIV: - if (compiler->scratches >= 4) { - FAIL_IF(push_inst32(compiler, 0xf84d2d04 /* str r2, [sp, #-4]! */)); - FAIL_IF(push_inst32(compiler, 0xf84dcd04 /* str ip, [sp, #-4]! */)); - } else if (compiler->scratches >= 3) - FAIL_IF(push_inst32(compiler, 0xf84d2d08 /* str r2, [sp, #-8]! */)); + case SLJIT_UDIVMOD: + case SLJIT_SDIVMOD: + case SLJIT_UDIVI: + case SLJIT_SDIVI: + SLJIT_COMPILE_ASSERT((SLJIT_UDIVMOD & 0x2) == 0 && SLJIT_UDIVI - 0x2 == SLJIT_UDIVMOD, bad_div_opcode_assignments); + SLJIT_COMPILE_ASSERT(reg_map[2] == 1 && reg_map[3] == 2 && reg_map[4] == 12, bad_register_mapping); + + saved_reg_count = 0; + if (compiler->scratches >= 4) + saved_reg_list[saved_reg_count++] = 12; + if (compiler->scratches >= 3) + saved_reg_list[saved_reg_count++] = 2; + if (op >= SLJIT_UDIVI) + saved_reg_list[saved_reg_count++] = 1; + + if (saved_reg_count > 0) { + FAIL_IF(push_inst32(compiler, 0xf84d0d00 | (saved_reg_count >= 3 ? 16 : 8) + | (saved_reg_list[0] << 12) /* str rX, [sp, #-8/-16]! */)); + if (saved_reg_count >= 2) { + SLJIT_ASSERT(saved_reg_list[1] < 8); + FAIL_IF(push_inst16(compiler, 0x9001 | (saved_reg_list[1] << 8) /* str rX, [sp, #4] */)); + } + if (saved_reg_count >= 3) { + SLJIT_ASSERT(saved_reg_list[2] < 8); + FAIL_IF(push_inst16(compiler, 0x9002 | (saved_reg_list[2] << 8) /* str rX, [sp, #8] */)); + } + } + #if defined(__GNUC__) FAIL_IF(sljit_emit_ijump(compiler, SLJIT_FAST_CALL, SLJIT_IMM, - (op == SLJIT_LUDIV ? SLJIT_FUNC_OFFSET(__aeabi_uidivmod) : SLJIT_FUNC_OFFSET(__aeabi_idivmod)))); + ((op | 0x2) == SLJIT_UDIVI ? SLJIT_FUNC_OFFSET(__aeabi_uidivmod) : SLJIT_FUNC_OFFSET(__aeabi_idivmod)))); #else #error "Software divmod functions are needed" #endif - if (compiler->scratches >= 4) { - FAIL_IF(push_inst32(compiler, 0xf85dcb04 /* ldr ip, [sp], #4 */)); - return push_inst32(compiler, 0xf85d2b04 /* ldr r2, [sp], #4 */); - } else if (compiler->scratches >= 3) - return push_inst32(compiler, 0xf85d2b08 /* ldr r2, [sp], #8 */); + + if (saved_reg_count > 0) { + if (saved_reg_count >= 3) { + SLJIT_ASSERT(saved_reg_list[2] < 8); + FAIL_IF(push_inst16(compiler, 0x9802 | (saved_reg_list[2] << 8) /* ldr rX, [sp, #8] */)); + } + if (saved_reg_count >= 2) { + SLJIT_ASSERT(saved_reg_list[1] < 8); + FAIL_IF(push_inst16(compiler, 0x9801 | (saved_reg_list[1] << 8) /* ldr rX, [sp, #4] */)); + } + return push_inst32(compiler, 0xf85d0b00 | (saved_reg_count >= 3 ? 16 : 8) + | (saved_reg_list[0] << 12) /* ldr rX, [sp], #8/16 */); + } return SLJIT_SUCCESS; } Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeMIPS_common.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeMIPS_common.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeMIPS_common.c (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeMIPS_common.c Sun Jan 17 17:23:28 2016 @@ -1053,8 +1053,11 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ #endif FAIL_IF(push_inst(compiler, MFLO | D(SLJIT_R0), DR(SLJIT_R0))); return push_inst(compiler, MFHI | D(SLJIT_R1), DR(SLJIT_R1)); - case SLJIT_LUDIV: - case SLJIT_LSDIV: + case SLJIT_UDIVMOD: + case SLJIT_SDIVMOD: + case SLJIT_UDIVI: + case SLJIT_SDIVI: + SLJIT_COMPILE_ASSERT((SLJIT_UDIVMOD & 0x2) == 0 && SLJIT_UDIVI - 0x2 == SLJIT_UDIVMOD, bad_div_opcode_assignments); #if !(defined SLJIT_MIPS_R1 && SLJIT_MIPS_R1) FAIL_IF(push_inst(compiler, NOP, UNMOVABLE_INS)); FAIL_IF(push_inst(compiler, NOP, UNMOVABLE_INS)); @@ -1062,15 +1065,15 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ #if (defined SLJIT_CONFIG_MIPS_64 && SLJIT_CONFIG_MIPS_64) if (int_op) - FAIL_IF(push_inst(compiler, (op == SLJIT_LUDIV ? DIVU : DIV) | S(SLJIT_R0) | T(SLJIT_R1), MOVABLE_INS)); + FAIL_IF(push_inst(compiler, ((op | 0x2) == SLJIT_UDIVI ? DIVU : DIV) | S(SLJIT_R0) | T(SLJIT_R1), MOVABLE_INS)); else - FAIL_IF(push_inst(compiler, (op == SLJIT_LUDIV ? DDIVU : DDIV) | S(SLJIT_R0) | T(SLJIT_R1), MOVABLE_INS)); + FAIL_IF(push_inst(compiler, ((op | 0x2) == SLJIT_UDIVI ? DDIVU : DDIV) | S(SLJIT_R0) | T(SLJIT_R1), MOVABLE_INS)); #else - FAIL_IF(push_inst(compiler, (op == SLJIT_LUDIV ? DIVU : DIV) | S(SLJIT_R0) | T(SLJIT_R1), MOVABLE_INS)); + FAIL_IF(push_inst(compiler, ((op | 0x2) == SLJIT_UDIVI ? DIVU : DIV) | S(SLJIT_R0) | T(SLJIT_R1), MOVABLE_INS)); #endif FAIL_IF(push_inst(compiler, MFLO | D(SLJIT_R0), DR(SLJIT_R0))); - return push_inst(compiler, MFHI | D(SLJIT_R1), DR(SLJIT_R1)); + return (op >= SLJIT_UDIVI) ? SLJIT_SUCCESS : push_inst(compiler, MFHI | D(SLJIT_R1), DR(SLJIT_R1)); } return SLJIT_SUCCESS; Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativePPC_common.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativePPC_common.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativePPC_common.c (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativePPC_common.c Sun Jan 17 17:23:28 2016 @@ -1267,22 +1267,23 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ FAIL_IF(push_inst(compiler, MULLW | D(SLJIT_R0) | A(TMP_REG1) | B(SLJIT_R1))); return push_inst(compiler, (op == SLJIT_LUMUL ? MULHWU : MULHW) | D(SLJIT_R1) | A(TMP_REG1) | B(SLJIT_R1)); #endif - case SLJIT_LUDIV: - case SLJIT_LSDIV: + case SLJIT_UDIVMOD: + case SLJIT_SDIVMOD: FAIL_IF(push_inst(compiler, OR | S(SLJIT_R0) | A(TMP_REG1) | B(SLJIT_R0))); #if (defined SLJIT_CONFIG_PPC_64 && SLJIT_CONFIG_PPC_64) - if (int_op) { - FAIL_IF(push_inst(compiler, (op == SLJIT_LUDIV ? DIVWU : DIVW) | D(SLJIT_R0) | A(TMP_REG1) | B(SLJIT_R1))); - FAIL_IF(push_inst(compiler, MULLW | D(SLJIT_R1) | A(SLJIT_R0) | B(SLJIT_R1))); - } else { - FAIL_IF(push_inst(compiler, (op == SLJIT_LUDIV ? DIVDU : DIVD) | D(SLJIT_R0) | A(TMP_REG1) | B(SLJIT_R1))); - FAIL_IF(push_inst(compiler, MULLD | D(SLJIT_R1) | A(SLJIT_R0) | B(SLJIT_R1))); - } - return push_inst(compiler, SUBF | D(SLJIT_R1) | A(SLJIT_R1) | B(TMP_REG1)); + FAIL_IF(push_inst(compiler, (int_op ? (op == SLJIT_UDIVMOD ? DIVWU : DIVW) : (op == SLJIT_UDIVMOD ? DIVDU : DIVD)) | D(SLJIT_R0) | A(SLJIT_R0) | B(SLJIT_R1))); + FAIL_IF(push_inst(compiler, (int_op ? MULLW : MULLD) | D(SLJIT_R1) | A(SLJIT_R0) | B(SLJIT_R1))); #else - FAIL_IF(push_inst(compiler, (op == SLJIT_LUDIV ? DIVWU : DIVW) | D(SLJIT_R0) | A(TMP_REG1) | B(SLJIT_R1))); + FAIL_IF(push_inst(compiler, (op == SLJIT_UDIVMOD ? DIVWU : DIVW) | D(SLJIT_R0) | A(SLJIT_R0) | B(SLJIT_R1))); FAIL_IF(push_inst(compiler, MULLW | D(SLJIT_R1) | A(SLJIT_R0) | B(SLJIT_R1))); +#endif return push_inst(compiler, SUBF | D(SLJIT_R1) | A(SLJIT_R1) | B(TMP_REG1)); + case SLJIT_UDIVI: + case SLJIT_SDIVI: +#if (defined SLJIT_CONFIG_PPC_64 && SLJIT_CONFIG_PPC_64) + return push_inst(compiler, (int_op ? (op == SLJIT_UDIVI ? DIVWU : DIVW) : (op == SLJIT_UDIVI ? DIVDU : DIVD)) | D(SLJIT_R0) | A(SLJIT_R0) | B(SLJIT_R1)); +#else + return push_inst(compiler, (op == SLJIT_UDIVI ? DIVWU : DIVW) | D(SLJIT_R0) | A(SLJIT_R0) | B(SLJIT_R1)); #endif } Modified: tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeSPARC_common.c URL: http://svn.apache.org/viewvc/tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeSPARC_common.c?rev=1725105&r1=1725104&r2=1725105&view=diff ============================================================================== --- tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeSPARC_common.c (original) +++ tomcat/jk/trunk/native/iis/pcre/sljit/sljitNativeSPARC_common.c Sun Jan 17 17:23:28 2016 @@ -777,20 +777,25 @@ SLJIT_API_FUNC_ATTRIBUTE sljit_si sljit_ #else #error "Implementation required" #endif - case SLJIT_LUDIV: - case SLJIT_LSDIV: + case SLJIT_UDIVMOD: + case SLJIT_SDIVMOD: + case SLJIT_UDIVI: + case SLJIT_SDIVI: + SLJIT_COMPILE_ASSERT((SLJIT_UDIVMOD & 0x2) == 0 && SLJIT_UDIVI - 0x2 == SLJIT_UDIVMOD, bad_div_opcode_assignments); #if (defined SLJIT_CONFIG_SPARC_32 && SLJIT_CONFIG_SPARC_32) - if (op == SLJIT_LUDIV) + if ((op | 0x2) == SLJIT_UDIVI) FAIL_IF(push_inst(compiler, WRY | S1(0), MOVABLE_INS)); else { FAIL_IF(push_inst(compiler, SRA | D(TMP_REG1) | S1(SLJIT_R0) | IMM(31), DR(TMP_REG1))); FAIL_IF(push_inst(compiler, WRY | S1(TMP_REG1), MOVABLE_INS)); } - FAIL_IF(push_inst(compiler, OR | D(TMP_REG2) | S1(0) | S2(SLJIT_R0), DR(TMP_REG2))); - FAIL_IF(push_inst(compiler, (op == SLJIT_LUDIV ? UDIV : SDIV) | D(SLJIT_R0) | S1(SLJIT_R0) | S2(SLJIT_R1), DR(SLJIT_R0))); + if (op <= SLJIT_SDIVMOD) + FAIL_IF(push_inst(compiler, OR | D(TMP_REG2) | S1(0) | S2(SLJIT_R0), DR(TMP_REG2))); + FAIL_IF(push_inst(compiler, ((op | 0x2) == SLJIT_UDIVI ? UDIV : SDIV) | D(SLJIT_R0) | S1(SLJIT_R0) | S2(SLJIT_R1), DR(SLJIT_R0))); + if (op >= SLJIT_UDIVI) + return SLJIT_SUCCESS; FAIL_IF(push_inst(compiler, SMUL | D(SLJIT_R1) | S1(SLJIT_R0) | S2(SLJIT_R1), DR(SLJIT_R1))); - FAIL_IF(push_inst(compiler, SUB | D(SLJIT_R1) | S1(TMP_REG2) | S2(SLJIT_R1), DR(SLJIT_R1))); - return SLJIT_SUCCESS; + return push_inst(compiler, SUB | D(SLJIT_R1) | S1(TMP_REG2) | S2(SLJIT_R1), DR(SLJIT_R1)); #else #error "Implementation required" #endif --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
