Merge branch develop into master

Author: Matthew Barr

CHANGELOG.md

@@ -2,6 +2,13 @@
 
 This is a list of notable changes to Hyperscan, in reverse chronological order.
 
+## [4.5.2] 2017-07-26
+- Bugfix for issue #57: Treat characters between `\Q.\E` as codepoints in
+  UTF8 mode.
+- Bugfix for issue #60: Use a portable flag for mktemp for fat runtime builds.
+- Bugfix for fat runtime builds on AVX-512 capable machines with Hyperscan's
+  AVX-512 support disabled.
+
 ## [4.5.1] 2017-06-16
 - Bugfix for issue #56: workaround for gcc-4.8 C++11 defect.
 - Bugfix for literal matching table generation, reversing a regression in

CMakeLists.txt

@@ -3,7 +3,7 @@ project (hyperscan C CXX)
 
 set (HS_MAJOR_VERSION 4)
 set (HS_MINOR_VERSION 5)
-set (HS_PATCH_VERSION 1)
+set (HS_PATCH_VERSION 2)
 set (HS_VERSION ${HS_MAJOR_VERSION}.${HS_MINOR_VERSION}.${HS_PATCH_VERSION})
 
 set(CMAKE_MODULE_PATH ${PROJECT_SOURCE_DIR}/cmake)

cmake/build_wrapper.sh

@@ -11,8 +11,8 @@ shift 2
 # $@ contains the actual build command
 OUT=$(echo "$@" | sed 's/.* -o \(.*\.o\).*/\1/')
 trap cleanup INT QUIT EXIT
-SYMSFILE=$(mktemp --tmpdir ${PREFIX}_rename.syms.XXXXX)
-KEEPSYMS=$(mktemp --tmpdir keep.syms.XXXXX)
+SYMSFILE=$(mktemp -p /tmp ${PREFIX}_rename.syms.XXXXX)
+KEEPSYMS=$(mktemp -p /tmp keep.syms.XXXXX)
 # find the libc used by gcc
 LIBC_SO=$("$@" --print-file-name=libc.so.6)
 cp ${KEEPSYMS_IN} ${KEEPSYMS}

cmake/config.h.in

@@ -21,6 +21,9 @@
 /* Define if building "fat" runtime. */
 #cmakedefine FAT_RUNTIME
 
+/* Define if building AVX-512 in the fat runtime. */
+#cmakedefine BUILD_AVX512
+
 /* Define to 1 if `backtrace' works. */
 #cmakedefine HAVE_BACKTRACE
 

src/parser/Parser.rl

@@ -1155,6 +1155,40 @@ unichar readUtf8CodePoint4c(const char *s) {
               '\\E' => {
                   fgoto main;
               };
+
+              #unicode chars
+              utf8_2c when is_utf8 => {
+                  assert(mode.utf8);
+                  /* leverage ComponentClass to generate the vertices */
+                  auto cc = getComponentClass(mode);
+                  cc->add(readUtf8CodePoint2c(ts));
+                  cc->finalize();
+                  currentSeq->addComponent(move(cc));
+              };
+
+              utf8_3c when is_utf8 => {
+                  assert(mode.utf8);
+                  /* leverage ComponentClass to generate the vertices */
+                  auto cc = getComponentClass(mode);
+                  cc->add(readUtf8CodePoint3c(ts));
+                  cc->finalize();
+                  currentSeq->addComponent(move(cc));
+              };
+
+              utf8_4c when is_utf8 => {
+                  assert(mode.utf8);
+                  /* leverage ComponentClass to generate the vertices */
+                  auto cc = getComponentClass(mode);
+                  cc->add(readUtf8CodePoint4c(ts));
+                  cc->finalize();
+                  currentSeq->addComponent(move(cc));
+              };
+
+              hi_byte when is_utf8 => {
+                  assert(mode.utf8);
+                  throwInvalidUtf8();
+              };
+
               # Literal character
               any => {
                   addLiteral(currentSeq, *ts, mode);
@@ -1169,6 +1203,31 @@ unichar readUtf8CodePoint4c(const char *s) {
               '\\E' => {
                   fret;
               };
+
+              #unicode chars
+              utf8_2c when is_utf8 => {
+                  assert(mode.utf8);
+                  currentCls->add(readUtf8CodePoint2c(ts));
+                  inCharClassEarly = false;
+              };
+
+              utf8_3c when is_utf8 => {
+                  assert(mode.utf8);
+                  currentCls->add(readUtf8CodePoint3c(ts));
+                  inCharClassEarly = false;
+              };
+
+              utf8_4c when is_utf8 => {
+                  assert(mode.utf8);
+                  currentCls->add(readUtf8CodePoint4c(ts));
+                  inCharClassEarly = false;
+              };
+
+              hi_byte when is_utf8 => {
+                  assert(mode.utf8);
+                  throwInvalidUtf8();
+              };
+
               # Literal character
               any => {
                   currentCls->add(*ts);

src/util/cpuid_flags.c

@@ -192,7 +192,8 @@ u64a cpuid_flags(void) {
     cap &= ~HS_CPU_FEATURES_AVX2;
 #endif
 
-#if !defined(FAT_RUNTIME) && !defined(HAVE_AVX512)
+#if (!defined(FAT_RUNTIME) && !defined(HAVE_AVX512)) ||                        \
+    (defined(FAT_RUNTIME) && !defined(BUILD_AVX512))
     cap &= ~HS_CPU_FEATURES_AVX512;
 #endif
 

unit/hyperscan/bad_patterns.txt

@@ -142,3 +142,5 @@
 145:/abc/8{edit_distance=1} #UTF-8 is disallowed for approximate matching.
 146:/(*UTF8)abc/{edit_distance=1} #UTF-8 is disallowed for approximate matching.
 147:/\b\BMYBt/s{edit_distance=1} #Pattern can never match.
+148:/\Q�\Eaaaa/8 #Expression is not valid UTF-8.
+149:/[\Q�\Eaaaa]/8 #Expression is not valid UTF-8.