fixes for 10.47 (#819)

* update required autoconf version

Since abc2ae6 (Update AX_PTHREAD (#694), 2025-02-12), when using
autoconf 2.60 with autogen, shows:

  error: possibly undefined macro: AS_ECHO

Use 2.62 instead, which include that macro, as the new required version.

* jit: enable x86 SIMD helper if SSE2 is available

The current implementation needs SSE2 support (which is part of the
base amd64 specification and available in most CPUs from this century).

Without this change, SIMD will be only enabled if SSE4.1 is detected
which is not true for older AMD cpus, even if 64bit capable.

While at it, update related documentation and other minor tweaks.

Author: carenas

NON-AUTOTOOLS-BUILD

@@ -171,7 +171,7 @@ example.
      src/config.h) . Compile src/pcre2test.c; don't forget -DHAVE_CONFIG_H if
      necessary, but do NOT define PCRE2_CODE_UNIT_WIDTH. Then link with the
      appropriate library/ies. If you compiled an 8-bit library, pcre2test also
-     needs the pcre2posix wrapper library.
+     needs the pcre2posix wrapper library when linking.
 
  (9) Run pcre2test on the testinput files in the testdata directory, and check
      that the output matches the corresponding testoutput files. There are

configure.ac

@@ -22,7 +22,7 @@ m4_define(libpcre2_posix_version, [3:6:0])
 # NOTE: The CMakeLists.txt file searches for the above variables in the first
 # 50 lines of this file. Please update that if the variables above are moved.
 
-AC_PREREQ([2.60])
+AC_PREREQ([2.62])
 AC_INIT([PCRE2],pcre2_major.pcre2_minor[]pcre2_prerelease,[],[pcre2])
 AC_CONFIG_SRCDIR([src/pcre2.h.in])
 AM_INIT_AUTOMAKE([dist-bzip2 dist-zip foreign])

doc/pcre2api.3

@@ -1,4 +1,4 @@
-.TH PCRE2API 3 "19 July 2025" "PCRE2 10.47-DEV"
+.TH PCRE2API 3 "05 October 2025" "PCRE2 10.47-DEV"
 .SH NAME
 PCRE2 - Perl-compatible regular expressions (revised API)
 .sp
@@ -1296,7 +1296,7 @@ documentation for more details.
 .sp
   PCRE2_CONFIG_JITTARGET
 .sp
-The \fIwhere\fP argument should point to a buffer that is at least 48 code
+The \fIwhere\fP argument should point to a buffer that is at least 64 code
 units long. (The exact length required can be found by calling
 \fBpcre2_config()\fP with \fBwhere\fP set to NULL.) The buffer is filled with a
 string that contains the name of the architecture for which the JIT compiler is
@@ -4598,6 +4598,6 @@ Cambridge, England.
 .rs
 .sp
 .nf
-Last updated: 19 July 2025
+Last updated: 05 October 2025
 Copyright (c) 1997-2024 University of Cambridge.
 .fi

doc/pcre2build.3

@@ -134,8 +134,8 @@ UTF support allows the libraries to process character code points up to
 0x10ffff in the strings that they handle. Unicode support also gives access to
 the Unicode properties of characters, using pattern escapes such as \eP, \ep,
 and \eX. Only the general category properties such as \fILu\fP and \fINd\fP,
-script names, and some bi-directional properties are supported. Details are
-given in the
+script names, and some bi-directional and binary properties are supported.
+Details are given in the
 .\" HREF
 \fBpcre2pattern\fP
 .\"
@@ -152,8 +152,8 @@ request this by starting with (*UCP).
 .sp
 The \eC escape sequence, which matches a single code unit, even in a UTF mode,
 can cause unpredictable behaviour because it may leave the current matching
-point in the middle of a multi-code-unit character. The application can lock it
-out by setting the PCRE2_NEVER_BACKSLASH_C option when calling
+point in the middle of a multi-code-unit character. The application can lock
+it out by setting the PCRE2_NEVER_BACKSLASH_C option when calling
 \fBpcre2_compile()\fP. There is also a build-time option
 .sp
   --enable-never-backslash-C
@@ -517,7 +517,7 @@ use), some extra configuration may be necessary. The INSTALL file for
 If your environment has not been set up so that an appropriate library is
 automatically included, you may need to add something like
 .sp
-  LIBS="-ncurses"
+  LIBS="-lncurses"
 .sp
 immediately before the \fBconfigure\fP command.
 .

src/pcre2_jit_compile.c

@@ -6888,7 +6888,7 @@ if (JIT_HAS_FAST_FORWARD_CHAR_SIMD && (common->nltype == NLTYPE_FIXED || common-
     OP2(SLJIT_ADD, STR_PTR, 0, STR_PTR, 0, SLJIT_IMM, IN_UCHARS(1));
     quit = CMP(SLJIT_NOT_EQUAL, TMP1, 0, SLJIT_IMM, CHAR_CR);
     }
-   else
+  else
     {
     fast_forward_char_simd(common, common->newline, common->newline, 0);
 

src/pcre2_jit_simd_inc.h

@@ -101,7 +101,7 @@ return CMP(SLJIT_NOT_EQUAL, reg, 0, SLJIT_IMM, 0xdc00);
 }
 #endif
 
-#endif /* SLJIT_CONFIG_X86 || SLJIT_CONFIG_S390X */
+#endif /* SLJIT_CONFIG_X86 || SLJIT_CONFIG_ARM_64 || SLJIT_CONFIG_S390X || SLJIT_CONFIG_LOONGARCH_64 */
 
 #if (defined SLJIT_CONFIG_X86 && SLJIT_CONFIG_X86)
 
@@ -229,7 +229,14 @@ switch (step)
   }
 }
 
+/* The AVX2 code path is currently disabled.
 #define JIT_HAS_FAST_FORWARD_CHAR_SIMD (sljit_has_cpu_feature(SLJIT_HAS_SIMD))
+*/
+#if defined(SLJIT_CONFIG_X86_64) && SLJIT_CONFIG_X86_64
+#define JIT_HAS_FAST_FORWARD_CHAR_SIMD 1
+#else
+#define JIT_HAS_FAST_FORWARD_CHAR_SIMD (sljit_has_cpu_feature(SLJIT_HAS_FPU))
+#endif
 
 static void fast_forward_char_simd(compiler_common *common, PCRE2_UCHAR char1, PCRE2_UCHAR char2, sljit_s32 offset)
 {
@@ -247,10 +254,10 @@ struct sljit_jump *quit;
 struct sljit_jump *partial_quit[2];
 vector_compare_type compare_type = vector_compare_match1;
 sljit_s32 tmp1_reg_ind = sljit_get_register_index(SLJIT_GP_REGISTER, TMP1);
-sljit_s32 data_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR0);
-sljit_s32 cmp1_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR1);
-sljit_s32 cmp2_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR2);
-sljit_s32 tmp_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR3);
+sljit_s32 data_ind = sljit_get_register_index(reg_type, SLJIT_VR0);
+sljit_s32 cmp1_ind = sljit_get_register_index(reg_type, SLJIT_VR1);
+sljit_s32 cmp2_ind = sljit_get_register_index(reg_type, SLJIT_VR2);
+sljit_s32 tmp_ind = sljit_get_register_index(reg_type, SLJIT_VR3);
 sljit_u32 bit = 0;
 int i;
 
@@ -366,7 +373,14 @@ if (common->utf && offset > 0)
 #endif
 }
 
+/* The AVX2 code path is currently disabled.
 #define JIT_HAS_FAST_REQUESTED_CHAR_SIMD (sljit_has_cpu_feature(SLJIT_HAS_SIMD))
+*/
+#if defined(SLJIT_CONFIG_X86_64) && SLJIT_CONFIG_X86_64
+#define JIT_HAS_FAST_REQUESTED_CHAR_SIMD 1
+#else
+#define JIT_HAS_FAST_REQUESTED_CHAR_SIMD (sljit_has_cpu_feature(SLJIT_HAS_FPU))
+#endif
 
 static jump_list *fast_requested_char_simd(compiler_common *common, PCRE2_UCHAR char1, PCRE2_UCHAR char2)
 {
@@ -381,10 +395,10 @@ struct sljit_jump *quit;
 jump_list *not_found = NULL;
 vector_compare_type compare_type = vector_compare_match1;
 sljit_s32 tmp1_reg_ind = sljit_get_register_index(SLJIT_GP_REGISTER, TMP1);
-sljit_s32 data_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR0);
-sljit_s32 cmp1_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR1);
-sljit_s32 cmp2_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR2);
-sljit_s32 tmp_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR3);
+sljit_s32 data_ind = sljit_get_register_index(reg_type, SLJIT_VR0);
+sljit_s32 cmp1_ind = sljit_get_register_index(reg_type, SLJIT_VR1);
+sljit_s32 cmp2_ind = sljit_get_register_index(reg_type, SLJIT_VR2);
+sljit_s32 tmp_ind = sljit_get_register_index(reg_type, SLJIT_VR3);
 sljit_u32 bit = 0;
 int i;
 
@@ -472,7 +486,14 @@ return not_found;
 
 #ifndef _WIN64
 
+/* The AVX2 code path is currently disabled.
 #define JIT_HAS_FAST_FORWARD_CHAR_PAIR_SIMD (sljit_has_cpu_feature(SLJIT_HAS_SIMD))
+*/
+#if defined(SLJIT_CONFIG_X86_64) && SLJIT_CONFIG_X86_64
+#define JIT_HAS_FAST_FORWARD_CHAR_PAIR_SIMD 1
+#else
+#define JIT_HAS_FAST_FORWARD_CHAR_PAIR_SIMD (sljit_has_cpu_feature(SLJIT_HAS_FPU))
+#endif
 
 static void fast_forward_char_pair_simd(compiler_common *common, sljit_s32 offs1,
   PCRE2_UCHAR char1a, PCRE2_UCHAR char1b, sljit_s32 offs2, PCRE2_UCHAR char2a, PCRE2_UCHAR char2b)
@@ -489,14 +510,14 @@ sljit_u32 bit1 = 0;
 sljit_u32 bit2 = 0;
 sljit_u32 diff = IN_UCHARS(offs1 - offs2);
 sljit_s32 tmp1_reg_ind = sljit_get_register_index(SLJIT_GP_REGISTER, TMP1);
-sljit_s32 data1_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR0);
-sljit_s32 data2_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR1);
-sljit_s32 cmp1a_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR2);
-sljit_s32 cmp2a_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR3);
-sljit_s32 cmp1b_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR4);
-sljit_s32 cmp2b_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR5);
-sljit_s32 tmp1_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_VR6);
-sljit_s32 tmp2_ind = sljit_get_register_index(SLJIT_SIMD_REG_128, SLJIT_TMP_DEST_VREG);
+sljit_s32 data1_ind = sljit_get_register_index(reg_type, SLJIT_VR0);
+sljit_s32 data2_ind = sljit_get_register_index(reg_type, SLJIT_VR1);
+sljit_s32 cmp1a_ind = sljit_get_register_index(reg_type, SLJIT_VR2);
+sljit_s32 cmp2a_ind = sljit_get_register_index(reg_type, SLJIT_VR3);
+sljit_s32 cmp1b_ind = sljit_get_register_index(reg_type, SLJIT_VR4);
+sljit_s32 cmp2b_ind = sljit_get_register_index(reg_type, SLJIT_VR5);
+sljit_s32 tmp1_ind = sljit_get_register_index(reg_type, SLJIT_VR6);
+sljit_s32 tmp2_ind = sljit_get_register_index(reg_type, SLJIT_TMP_DEST_VREG);
 struct sljit_label *start;
 #if defined SUPPORT_UNICODE && PCRE2_CODE_UNIT_WIDTH != 32
 struct sljit_label *restart;

testdata/testinput12

@@ -368,7 +368,7 @@
 
 /\p{BC: Aሴ}/utf
 
-# A special extra option allows excaped surrogate code points in 32-bit mode,
+# A special extra option allows escaped surrogate code points in 32-bit mode,
 # but subjects containing them must not be UTF-checked. These patterns give
 # errors in 16-bit mode.
 

testdata/testoutput12-16

@@ -1428,7 +1428,7 @@ Failed: error 146 at offset 7: malformed \P or \p sequence
 Failed: error 146 at offset 9: malformed \P or \p sequence
         here: \p{BC: Aሴ |<--| }
 
-# A special extra option allows excaped surrogate code points in 32-bit mode,
+# A special extra option allows escaped surrogate code points in 32-bit mode,
 # but subjects containing them must not be UTF-checked. These patterns give
 # errors in 16-bit mode.
 

testdata/testoutput12-32

@@ -1421,7 +1421,7 @@ Failed: error 146 at offset 7: malformed \P or \p sequence
 Failed: error 146 at offset 9: malformed \P or \p sequence
         here: \p{BC: Aሴ |<--| }
 
-# A special extra option allows excaped surrogate code points in 32-bit mode,
+# A special extra option allows escaped surrogate code points in 32-bit mode,
 # but subjects containing them must not be UTF-checked. These patterns give
 # errors in 16-bit mode.