Merge pull request OpenSCAP#2015 from evgenyz/pcre2

PCRE2

.github/workflows/build.yml

@@ -34,14 +34,14 @@ jobs:
     - name: Install Deps
       run: |
         sudo apt-get update
-        sudo apt-get -y install lcov swig xsltproc rpm-common lua5.3 libyaml-dev libapt-pkg-dev libdbus-1-dev libdbus-glib-1-dev libcurl4-openssl-dev libgcrypt-dev libselinux1-dev libgconf2-dev libacl1-dev libblkid-dev libcap-dev libxml2-dev libxslt1-dev libxml-parser-perl libxml-xpath-perl libperl-dev librpm-dev librtmp-dev libxmlsec1-dev libxmlsec1-openssl python3-dbusmock
+        sudo apt-get -y install lcov swig xsltproc rpm-common lua5.3 libpcre2-dev libyaml-dev libapt-pkg-dev libdbus-1-dev libdbus-glib-1-dev libcurl4-openssl-dev libgcrypt-dev libselinux1-dev libgconf2-dev libacl1-dev libblkid-dev libcap-dev libxml2-dev libxslt1-dev libxml-parser-perl libxml-xpath-perl libperl-dev librpm-dev librtmp-dev libxmlsec1-dev libxmlsec1-openssl python3-dbusmock
         sudo apt-get -y remove rpm
 
     # Runs a set of commands using the runners shell
     - name: Build
       working-directory: ./build
       run: |
-        cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo ../
+        cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo -DWITH_PCRE2=True ../
         make all
 
     - name: Test
@@ -57,15 +57,15 @@ jobs:
       image: fedora:latest
     steps:
     - name: Install Deps
-      run: dnf install -y cmake git dbus-devel GConf2-devel libacl-devel libblkid-devel libcap-devel libcurl-devel libgcrypt-devel libselinux-devel libxml2-devel libxslt-devel libattr-devel make openldap-devel pcre-devel perl-XML-Parser perl-XML-XPath perl-devel python3-devel python3-dbusmock rpm-devel swig bzip2-devel gcc-c++ libyaml-devel xmlsec1-devel xmlsec1-openssl-devel hostname bzip2 lua rpm-build which strace apt-devel
+      run: dnf install -y cmake git dbus-devel GConf2-devel libacl-devel libblkid-devel libcap-devel libcurl-devel libgcrypt-devel libselinux-devel libxml2-devel libxslt-devel libattr-devel make openldap-devel pcre2-devel perl-XML-Parser perl-XML-XPath perl-devel python3-devel python3-dbusmock rpm-devel swig bzip2-devel gcc-c++ libyaml-devel xmlsec1-devel xmlsec1-openssl-devel hostname bzip2 lua rpm-build which strace apt-devel
     - name: Checkout
       uses: actions/checkout@v3
       with:
         submodules: recursive
     - name: Build
       working-directory: ./build
       run: |
-        cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo ../
+        cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo -DWITH_PCRE2=True ../
         make all
     - name: Test
       working-directory: ./build
@@ -97,13 +97,13 @@ jobs:
         brew install swig
         brew install libxmlsec1
         brew install openssl
-        brew install pcre
+        brew install pcre2
 
     # Runs a set of commands using the runners shell
     - name: Build
       run: |
         cd $GITHUB_WORKSPACE/build
-        cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo -DENABLE_PROBES=false ../
+        cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo -DWITH_PCRE2=True -DENABLE_PROBES=False ../
         make all
 
     - name: Test

CMakeLists.txt

@@ -106,7 +106,11 @@ find_package(Doxygen)
 find_package(GConf)
 find_package(Ldap)
 find_package(OpenDbx)
-find_package(PCRE REQUIRED)
+if(WITH_PCRE2)
+	find_package(PCRE2 REQUIRED)
+else()
+	find_package(PCRE REQUIRED)
+endif()
 find_package(PerlLibs)
 find_package(Popt)
 find_package(Systemd)
@@ -193,6 +197,13 @@ else()
 	set(YAML_FILTER_FOUND FALSE)
 endif()
 
+if(PCRE2_FOUND)
+	set(HAVE_PCRE2 1)
+	message("-- Using PCRE2")
+elseif(PCRE_FOUND)
+	message("-- Using PCRE")
+endif()
+
 check_library_exists(rt clock_gettime "" HAVE_CLOCK_GETTIME)
 check_function_exists(posix_memalign HAVE_POSIX_MEMALIGN)
 check_function_exists(memalign HAVE_MEMALIGN)
@@ -330,6 +341,8 @@ cmake_dependent_option(ENABLE_OSCAP_UTIL_CHROOT "enables the oscap-chroot utilit
 option(ENABLE_OSCAP_UTIL_AUTOTAILOR "enables the autotailor utility that is able to perform command-line tailoring" TRUE)
 option(ENABLE_OSCAP_REMEDIATE_SERVICE "enables the oscap-remediate service" FALSE)
 
+option(WITH_PCRE2 "use PCRE2 library" FALSE)
+
 # ---------- TEST-SUITE SWITCHES
 
 # Tests will be turned off on Windows, because the test suite uses bash
@@ -517,6 +530,7 @@ endif()
 
 if (${CMAKE_C_COMPILER_ID} STREQUAL "GNU" OR ${CMAKE_C_COMPILER_ID} STREQUAL "Clang")
 	set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -pipe -W -Wall -Wnonnull -Wshadow -Wformat -Wundef -Wno-unused-parameter -Wmissing-prototypes -Wno-unknown-pragmas -Wno-int-conversion -Werror=implicit-function-declaration -D_GNU_SOURCE -std=c99")
+	add_link_options(-Wl,-z,now)
 endif()
 if(${CMAKE_SYSTEM_NAME} STREQUAL "FreeBSD")
 	add_link_options(-lkvm -lm -lprocstat)
@@ -563,6 +577,7 @@ include_directories(
 	${XMLSEC_INCLUDE_DIRS}
 	${OPENSSL_INCLUDE_DIR}
 	${PCRE_INCLUDE_DIRS}
+	${PCRE2_INCLUDE_DIRS}
 )
 
 # Honor visibility properties for all target types

cmake/FindPCRE2.cmake

@@ -0,0 +1,27 @@
+# - Find pcre
+# Find the native PCRE2 headers and libraries.
+#
+# PCRE2_INCLUDE_DIRS	- where to find pcre2.h, etc.
+# PCRE2_LIBRARIES	- List of libraries when using pcre.
+# PCRE2_FOUND	- True if pcre found.
+
+# Look for the header file.
+FIND_PATH(PCRE2_INCLUDE_DIR pcre2.h)
+
+# Look for the library.
+FIND_LIBRARY(PCRE2_LIBRARY NAMES libpcre2.a pcre2-8)
+
+# Handle the QUIETLY and REQUIRED arguments and set PCRE2_FOUND to TRUE if all listed variables are TRUE.
+INCLUDE(FindPackageHandleStandardArgs)
+FIND_PACKAGE_HANDLE_STANDARD_ARGS(PCRE2 DEFAULT_MSG PCRE2_LIBRARY PCRE2_INCLUDE_DIR)
+
+# Copy the results to the output variables.
+IF(PCRE2_FOUND)
+    SET(PCRE2_LIBRARIES ${PCRE2_LIBRARY})
+    SET(PCRE2_INCLUDE_DIRS ${PCRE2_INCLUDE_DIR})
+ELSE(PCRE2_FOUND)
+    SET(PCRE_LIBRARIES)
+    SET(PCRE_INCLUDE_DIRS)
+ENDIF(PCRE2_FOUND)
+
+MARK_AS_ADVANCED(PCRE2_INCLUDE_DIRS PCRE2_LIBRARIES)

config.h.in

@@ -29,6 +29,8 @@
 #define OSCAP_DEFAULT_CPE_PATH "@OSCAP_DEFAULT_CPE_PATH@"
 #define OSCAP_TEMP_DIR "@OSCAP_TEMP_DIR@"
 
+#cmakedefine HAVE_PCRE2
+
 #cmakedefine HAVE_ATOMIC_BUILTINS
 
 #cmakedefine HAVE_ACL_EXTENDED_FILE

docs/developer/developer.adoc

@@ -89,6 +89,12 @@ libxml-xpath-perl libperl-dev libbz2-dev librpm-dev g++ libapt-pkg-dev libyaml-d
 libxmlsec1-dev libxmlsec1-openssl
 ----
 
+Since version 1.3.9 OpenSCAP can use either PCRE or PCRE2 library to handle regular
+expressions. The default behaviour for 1.3.x line of versions is to try and link
+with PCRE. In order to switxh the build system to PCRE2 one must configure
+CMake with '-DWITH_PCRE2=True' parameter. Dependencies: Debian/Ubuntu — pcre2-dev,
+Fedora/RHEL — pcre2-devel.
+
 When you have all the build dependencies installed you can build the library.
 --
 
@@ -319,7 +325,7 @@ behaviour.
 * *OSCAP_FULL_VALIDATION=1* - validate all exported documents (slower)
 * *SEXP_VALIDATE_DISABLE=1* - do not validate SEXP expressions (faster)
 * *OSCAP_PCRE_EXEC_RECURSION_LIMIT* - override default recursion limit
-  for match in pcre_exec call in textfilecontent(54) probes.
+  for match in pcre_exec/pcre2_match calls in textfilecontent(54) probes.
 
 
 

docs/manual/manual.adoc

@@ -1613,7 +1613,7 @@ not considered local by the scanner:
 * `OSCAP_EVALUATION_TARGET` - Change value of target facts `urn:xccdf:fact:identifier` and `urn:xccdf:fact:asset:identifier:ein` in XCCDF results. Used during offline scanning to pass the name of the target system.
 * `OSCAP_FULL_VALIDATION` - If set, XML schema validation will be performed in every step of SCAP content processing.
 * `OSCAP_OVAL_COMMAND_OPTIONS` - Additional command line options for `oscap oval` module. The value of this environment variable is appended to the actual command line options of `oscap` command.
-* `OSCAP_PCRE_EXEC_RECURSION_LIMIT` - Set recursion limit of regular expression matching using `pcre_exec` function.
+* `OSCAP_PCRE_EXEC_RECURSION_LIMIT` - Set recursion limit of regular expression matching using `pcre_exec`/`pcre2_match` functions.
 * `OSCAP_PROBE_ROOT` - Path to a directory which contains mounted filesystem to be evaluated. Used for offline scanning.
 * `SEXP_VALIDATE_DISABLE` - If set, `oscap` will not validate SEXP expressions during its execution.
 * `SOURCE_DATE_EPOCH` - Timestamp in seconds since epoch. This timestamp will be used instead of the current time to populate `timestamp` attributes in SCAP source data streams created by `oscap ds sds-compose` sub-module. This is used for reproducible builds of data streams.

openscap.spec

@@ -18,7 +18,11 @@ BuildRequires:  apt-devel
 %endif
 BuildRequires:  rpm-devel
 BuildRequires:  libgcrypt-devel
+%if 0%{?fedora}
+BuildRequires:  pcre2-devel
+%else
 BuildRequires:  pcre-devel
+%endif
 BuildRequires:  libacl-devel
 BuildRequires:  libselinux-devel
 BuildRequires:  libcap-devel
@@ -136,6 +140,9 @@ Tool for scanning Atomic containers.
 # gconf is a legacy system not used any more, and it blocks testing of oscap-anaconda-addon
 # as gconf is no longer part of the installation medium
 %cmake \
+%if 0%{?fedora}
+    -DWITH_PCRE2=ON \
+%endif
     -DENABLE_PERL=OFF \
     -DENABLE_DOCS=ON \
     -DOPENSCAP_PROBE_UNIX_GCONF=OFF \

src/CMakeLists.txt

@@ -81,7 +81,7 @@ set_target_properties(openscap PROPERTIES
 	C_VISIBILITY_PRESET hidden
 )
 
-target_link_libraries(openscap ${LIBXML2_LIBRARIES} ${LIBXSLT_LIBRARIES} ${XMLSEC_LIBRARIES} ${OPENSSL_LIBRARIES} ${LIBXSLT_EXSLT_LIBRARIES} ${PCRE_LIBRARIES} ${CURL_LIBRARIES} ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(openscap ${LIBXML2_LIBRARIES} ${LIBXSLT_LIBRARIES} ${XMLSEC_LIBRARIES} ${OPENSSL_LIBRARIES} ${LIBXSLT_EXSLT_LIBRARIES} ${PCRE_LIBRARIES} ${PCRE2_LIBRARIES} ${CURL_LIBRARIES} ${CMAKE_THREAD_LIBS_INIT})
 if (BZIP2_FOUND)
 	target_link_libraries(openscap ${BZIP2_LIBRARIES})
 endif()

src/CPE/cpename.c

@@ -37,11 +37,11 @@
 
 #include <string.h>
 #include <stdio.h>
-#include <pcre.h>
 #include <ctype.h>
 
 #include "cpe_name.h"
 #include "common/util.h"
+#include "common/oscap_pcre.h"
 #include "oscap_helpers.h"
 
 #define CPE_URI_SUPPORTED "2.3"
@@ -549,8 +549,8 @@ cpe_format_t cpe_name_get_format_of_str(const char *str)
 	if (str == NULL)
 		return CPE_FORMAT_UNKNOWN;
 
-	pcre *re;
-	const char *error;
+	oscap_pcre_t *re;
+	char *error;
 	int erroffset;
 	int rc;
 	int ovector[30];
@@ -559,26 +559,38 @@ cpe_format_t cpe_name_get_format_of_str(const char *str)
 	// http://scap.nist.gov/schema/cpe/2.3/cpe-naming_2.3.xsd
 	// [c] was replaced with [cC] here and in the schemas
 
-	re = pcre_compile("^[cC][pP][eE]:/[AHOaho]?(:[A-Za-z0-9\\._\\-~%]*){0,6}$", 0, &error, &erroffset, NULL);
-	rc = pcre_exec(re, NULL, str, strlen(str), 0, 0, ovector, 30);
-	pcre_free(re);
+	re = oscap_pcre_compile("^[cC][pP][eE]:/[AHOaho]?(:[A-Za-z0-9\\._\\-~%]*){0,6}$", 0, &error, &erroffset);
+	if (re == NULL) {
+		oscap_pcre_err_free(error);
+		return CPE_FORMAT_UNKNOWN;
+	}
+	rc = oscap_pcre_exec(re, str, strlen(str), 0, 0, ovector, 30);
+	oscap_pcre_free(re);
 
 	if (rc >= 0)
 		return CPE_FORMAT_URI;
 
 	// The regex was taken from the official XSD at
 	// http://scap.nist.gov/schema/cpe/2.3/cpe-naming_2.3.xsd
-	re = pcre_compile("^cpe:2\\.3:[aho\\*\\-](:(((\\?*|\\*?)([a-zA-Z0-9\\-\\._]|(\\\\[\\\\\\*\\?!\"#$$%&'\\(\\)\\+,/:;<=>@\\[\\]\\^`\\{\\|}~]))+(\\?*|\\*?))|[\\*\\-])){5}(:(([a-zA-Z]{2,3}(-([a-zA-Z]{2}|[0-9]{3}))?)|[\\*\\-]))(:(((\\?*|\\*?)([a-zA-Z0-9\\-\\._]|(\\\\[\\\\\\*\\?!\"#$$%&'\\(\\)\\+,/:;<=>@\\[\\]\\^`\\{\\|}~]))+(\\?*|\\*?))|[\\*\\-])){4}$", 0, &error, &erroffset, NULL);
-	rc = pcre_exec(re, NULL, str, strlen(str), 0, 0, ovector, 30);
-	pcre_free(re);
+	re = oscap_pcre_compile("^cpe:2\\.3:[aho\\*\\-](:(((\\?*|\\*?)([a-zA-Z0-9\\-\\._]|(\\\\[\\\\\\*\\?!\"#$$%&'\\(\\)\\+,/:;<=>@\\[\\]\\^`\\{\\|}~]))+(\\?*|\\*?))|[\\*\\-])){5}(:(([a-zA-Z]{2,3}(-([a-zA-Z]{2}|[0-9]{3}))?)|[\\*\\-]))(:(((\\?*|\\*?)([a-zA-Z0-9\\-\\._]|(\\\\[\\\\\\*\\?!\"#$$%&'\\(\\)\\+,/:;<=>@\\[\\]\\^`\\{\\|}~]))+(\\?*|\\*?))|[\\*\\-])){4}$", 0, &error, &erroffset);
+	if (re == NULL) {
+		oscap_pcre_err_free(error);
+		return CPE_FORMAT_UNKNOWN;
+	}
+	rc = oscap_pcre_exec(re, str, strlen(str), 0, 0, ovector, 30);
+	oscap_pcre_free(re);
 
 	if (rc >= 0)
 		return CPE_FORMAT_STRING;
 
 	// FIXME: This should be way more strict
-	re = pcre_compile("^wfn:\\[.+\\]$", PCRE_CASELESS, &error, &erroffset, NULL);
-	rc = pcre_exec(re, NULL, str, strlen(str), 0, 0, ovector, 30);
-	pcre_free(re);
+	re = oscap_pcre_compile("^wfn:\\[.+\\]$", OSCAP_PCRE_OPTS_CASELESS, &error, &erroffset);
+	if (re == NULL) {
+		oscap_pcre_err_free(error);
+		return CPE_FORMAT_UNKNOWN;
+	}
+	rc = oscap_pcre_exec(re, str, strlen(str), 0, 0, ovector, 30);
+	oscap_pcre_free(re);
 
 	if (rc >= 0)
 		return CPE_FORMAT_WFN;

src/OVAL/oval_component.c

@@ -55,8 +55,8 @@
 #include "common/debug_priv.h"
 #include "common/_error.h"
 #include "common/oscap_string.h"
+#include "common/oscap_pcre.h"
 #include "oval_glob_to_regex.h"
-#include <pcre.h>
 
 #if !defined(OVAL_PROBES_ENABLED)
 const char *oval_subtype_to_str(oval_subtype_t subtype);
@@ -1980,12 +1980,16 @@ static long unsigned int _parse_fmt_sse(char *dt)
 static bool _match(const char *pattern, const char *string)
 {
 	bool match = false;
-	pcre *re;
-	const char *error;
+	oscap_pcre_t *re;
+	char *error;
 	int erroffset = -1, ovector[60], ovector_len = sizeof (ovector) / sizeof (ovector[0]);
-	re = pcre_compile(pattern, PCRE_UTF8, &error, &erroffset, NULL);
-	match = (pcre_exec(re, NULL, string, strlen(string), 0, 0, ovector, ovector_len) >= 0);
-	pcre_free(re);
+	re = oscap_pcre_compile(pattern, OSCAP_PCRE_OPTS_UTF8, &error, &erroffset);
+	if (re == NULL) {
+		oscap_pcre_err_free(error);
+		return false;
+	}
+	match = (oscap_pcre_exec(re, string, strlen(string), 0, 0, ovector, ovector_len) >= 0);
+	oscap_pcre_free(re);
 	return match;
 }
 
@@ -2208,13 +2212,14 @@ static oval_syschar_collection_flag_t _oval_component_evaluate_REGEX_CAPTURE(ova
 	int rc;
 	char *pattern;
 	int erroffset = -1;
-	pcre *re = NULL;
-	const char *error;
+	oscap_pcre_t *re = NULL;
+	char *error;
 
 	pattern = oval_component_get_regex_pattern(component);
-	re = pcre_compile(pattern, PCRE_UTF8, &error, &erroffset, NULL);
+	re = oscap_pcre_compile(pattern, OSCAP_PCRE_OPTS_UTF8, &error, &erroffset);
 	if (re == NULL) {
-		dE("pcre_compile() failed: \"%s\".", error);
+		dE("oscap_pcre_compile() failed: \"%s\".", error);
+		oscap_pcre_err_free(error);
 		return SYSCHAR_FLAG_ERROR;
 	}
 
@@ -2233,9 +2238,9 @@ static oval_syschar_collection_flag_t _oval_component_evaluate_REGEX_CAPTURE(ova
 			for (i = 0; i < ovector_len; ++i)
 				ovector[i] = -1;
 
-			rc = pcre_exec(re, NULL, text, strlen(text), 0, 0, ovector, ovector_len);
+			rc = oscap_pcre_exec(re, text, strlen(text), 0, 0, ovector, ovector_len);
 			if (rc < -1) {
-				dE("pcre_exec() failed: %d.", rc);
+				dE("oscap_pcre_exec() failed: %d.", rc);
 				flag = SYSCHAR_FLAG_ERROR;
 				break;
 			}
@@ -2263,7 +2268,7 @@ static oval_syschar_collection_flag_t _oval_component_evaluate_REGEX_CAPTURE(ova
 		oval_collection_free_items(subcoll, (oscap_destruct_func) oval_value_free);
 	}
 	oval_component_iterator_free(subcomps);
-	pcre_free(re);
+	oscap_pcre_free(re);
 	return flag;
 }
 

src/OVAL/oval_schema_version.c

@@ -29,12 +29,11 @@
 #include <string.h>
 #include <sys/types.h>
 #include "common/util.h"
+#include "common/oscap_pcre.h"
 #include "common/debug_priv.h"
 #include "public/oval_schema_version.h"
 
-#include <pcre.h>
-
-#define OVECTOR_LEN 30 // must be a multiple of 30
+#define OVECTOR_LEN 60 // must be a multiple of 30
 
 static int _parse_int(const char *substring, size_t substring_length)
 {
@@ -55,16 +54,17 @@ oval_schema_version_t oval_schema_version_from_cstr(const char *ver_str)
 		return version;
 	}
 	const char *pattern = "([0-9]+)\\.([0-9]+)(?:\\.([0-9]+))?(?::([0-9]+)\\.([0-9]+)(?:\\.([0-9]+))?)?";
-	const char *error;
+	char *error;
 	int erroffset;
-	pcre *re = pcre_compile(pattern, 0, &error, &erroffset, NULL);
+	oscap_pcre_t *re = oscap_pcre_compile(pattern, 0, &error, &erroffset);
 	if (re == NULL) {
 		dE("Regular expression compilation failed with %s", pattern);
+		oscap_pcre_err_free(error);
 		return version;
 	}
 	int ovector[OVECTOR_LEN];
-	int rc = pcre_exec(re, NULL, ver_str, strlen(ver_str), 0, 0, ovector, OVECTOR_LEN);
-	pcre_free(re);
+	int rc = oscap_pcre_exec(re, ver_str, strlen(ver_str), 0, 0, ovector, OVECTOR_LEN);
+	oscap_pcre_free(re);
 	if (rc < 0) {
 		dE("Regular expression %s did not match string %s", pattern, ver_str);
 		return version;