Add PCRE2 support

This is still work in progress.

Work on oniguruma (https://github.com/kkos/oniguruma) has been discontinued
so we need to find some other regex implementation.

To build and test:
./configure --with-oniguruma=no --with-pcre2
make
make check
./jq --run-tests < ./tests/onig.test
./jq --run-tests < ./tests/pcre2.test

Known TODOs:
- Proper configure options (skip vendor for pcre2?)
- What about libjq usage? dont builtin and require libpcre2? (seems to be what debian does already for libonig)
- pcre2 in ci/release workflows
- Update Dockerfile
- Update docs

Build status:
- macos: Switch to native arm64/arm64 build hosts to simplify build a bit
- windows: Build fails to find libs and headers installed by pcre2-devel, no idea why, but pcre2-config seem to run
- linux native: Test fails due to old version of pcre2 (see below)
- linux cross: Maybe switch to docker build --platform

Good references:
- https://github.com/PCRE2Project/pcre2/blob/master/src/pcre2demo.c
Pre pcre2_next_match usage:
- https://github.com/PCRE2Project/pcre2/blob/eb3bd3cf1418cb1a0eabf984b0b1e80b6bdd9314~1/src/pcre2demo.c (pre pcre2_next_match usage)
- pcre2test cli tools is a good Playaround for pcre2

Notes:
".+?\\b" test in onig.test seems to be behaves differently depending on pcre2 version.
I suspect this fix in 10.43:
PCRE2Project/pcre2@0a55280

I noticed a clang -fsanitize=memory use-of-uninitialized-valu issue but it seems to go awa with pcre2 master

Related to #3313

Author: wader

.github/workflows/ci.yml

@@ -81,14 +81,16 @@ jobs:
       - name: Install packages
         run: |
           sudo apt-get update
-          sudo apt-get install -y automake autoconf libtool crossbuild-essential-${{ matrix.arch }}
+          # crossbuild libpcre2?
+          sudo apt-get install -y automake autoconf libtool libpcre2-dev libpcre2-8-0 crossbuild-essential-${{ matrix.arch }}
       - name: Build
         run: |
           autoreconf -i
           ./configure \
             --host=${{ matrix.CC }} \
             --disable-docs \
-            --with-oniguruma=builtin \
+            --with-oniguruma=no \
+            --with-pcre2 \
             --enable-static \
             --enable-all-static \
             CFLAGS="-O2 -pthread -fstack-protector-all"
@@ -127,10 +129,10 @@ jobs:
           - arm64
         include:
           - arch: amd64
-            target: x86_64-apple-darwin
+            runs_on: macos-15-intel
           - arch: arm64
-            target: arm64-apple-darwin
-    runs-on: macos-14
+            runs_on: macos-15
+    runs-on: ${{ matrix.runs_on }}
     env:
       LDFLAGS: -dead_strip
       SUFFIX: macos-${{ matrix.arch }}
@@ -143,17 +145,14 @@ jobs:
         run: |
           # brew update sometimes fails with "Fetching /usr/local/Homebrew/Library/Taps/homebrew/homebrew-cask failed!"
           brew update || brew update-reset
-          brew install autoconf automake libtool
-      - name: Set CC
-        run: |
-          echo "CC=clang -target ${{ matrix.target }}$(uname -r)" >> "$GITHUB_ENV"
+          brew install autoconf automake libtool pcre2
       - name: Build
         run: |
           autoreconf -i
           ./configure \
-            --host="${{ matrix.target }}$(uname -r)" \
             --disable-docs \
-            --with-oniguruma=builtin \
+            --with-oniguruma=no \
+            --with-pcre2 \
             --disable-shared \
             --enable-static \
             --enable-all-static \
@@ -226,14 +225,17 @@ jobs:
             autoconf
             automake
             libtool
+            pcre2
+            pcre2-devel
           pacboy: >-
             toolchain:p
       - name: Build
         run: |
           autoreconf -i
           ./configure \
             --disable-docs \
-            --with-oniguruma=builtin \
+            --with-oniguruma=no \
+            --with-pcre2 \
             --disable-shared \
             --enable-static \
             --enable-all-static \
@@ -272,13 +274,14 @@ jobs:
       - name: Install packages
         run: |
           sudo apt-get update -qq
-          sudo apt-get install -y automake autoconf
+          sudo apt-get install -y automake autoconf libpcre2-dev libpcre2-8-0
       - name: Create dist
         run: |
           autoreconf -i
           ./configure \
             --disable-docs \
-            --with-oniguruma=builtin
+            --with-oniguruma=no \
+            --with-pcre2 \
           make distcheck
           make dist dist-zip
           git diff --exit-code

.github/workflows/scanbuild.yml

@@ -16,12 +16,12 @@ jobs:
       - name: Install packages
         run: |
           sudo apt-get update -qq
-          sudo apt-get install -y automake autoconf gdb valgrind clang clang-tools
+          sudo apt-get install -y automake autoconf gdb valgrind clang clang-tools libpcre2-dev libpcre2-8-0
           echo "CC=clang" >> "$GITHUB_ENV"
       - name: Build
         run: |
           autoreconf -i
-          ./configure --enable-valgrind CFLAGS="-g -O0"
+          ./configure --enable-valgrind --with-oniguruma=no --with-pcre2 CFLAGS="-g -O0"
           scan-build --keep-going --exclude vendor/ make -j"$(nproc)"
       - name: Test
         run: |

.github/workflows/valgrind.yml

@@ -16,14 +16,15 @@ jobs:
       - name: Install packages
         run: |
           sudo apt-get update
-          sudo apt-get install -y automake autoconf libtool valgrind
+          sudo apt-get install -y automake autoconf libtool libpcre2-dev libpcre2-8-0 valgrind
       - name: Build
         run: |
           autoreconf -i
           ./configure \
             --disable-docs \
             --enable-valgrind \
-            --with-oniguruma=builtin
+            --with-oniguruma=no \
+            --with-pcre2
           make -j"$(nproc)"
           file ./jq
       - name: Test

Dockerfile

@@ -10,6 +10,8 @@ RUN apt-get update \
       build-essential \
       autoconf \
       libtool \
+      libpcre2-dev \
+      libpcre2-8-0 \
       git \
  && apt-get clean \
  && rm -rf /var/lib/apt/lists/*
@@ -19,7 +21,8 @@ COPY . /app
 RUN autoreconf -i \
  && ./configure \
       --disable-docs \
-      --with-oniguruma=builtin \
+      --with-oniguruma=no \
+      --with-pcre2 \
       --enable-static \
       --enable-all-static \
       --prefix=/usr/local \

Makefile.am

@@ -56,7 +56,7 @@ AM_YFLAGS = --warnings=all -Wno-yacc -d
 lib_LTLIBRARIES = libjq.la
 libjq_la_SOURCES = ${LIBJQ_SRC}
 libjq_la_LIBADD = -lm
-libjq_la_LDFLAGS = $(onig_LDFLAGS) -export-symbols-regex '^j[qv]_' -version-info 1:4:0
+libjq_la_LDFLAGS = $(onig_LDFLAGS) $(pcre2_LDFLAGS) -export-symbols-regex '^j[qv]_' -version-info 1:4:0
 
 if WIN32
 libjq_la_LIBADD += -lshlwapi
@@ -198,11 +198,17 @@ SUBDIRS = vendor/oniguruma
 endif
 
 AM_CFLAGS += $(onig_CFLAGS)
+AM_CFLAGS += $(pcre2_CFLAGS)
 
 if WITH_ONIGURUMA
 TESTS += tests/onigtest tests/manonigtest
 endif
 
+if WITH_PCRE2
+TESTS += tests/onigtest tests/manonigtest
+TESTS += tests/pcre2test
+endif
+
 ### Packaging
 
 install-binaries: $(BUILT_SOURCES)
@@ -230,7 +236,7 @@ EXTRA_DIST = $(DOC_FILES) $(man_MANS) $(TESTS) $(TEST_LOG_COMPILER)     \
         tests/onig.supp tests/local.supp                                \
         tests/setup tests/torture/input0.json                           \
         tests/optional.test tests/man.test tests/manonig.test           \
-        tests/jq.test tests/onig.test tests/base64.test tests/uri.test  \
+        tests/jq.test tests/onig.test tests/pcre2.test tests/base64.test tests/uri.test  \
         tests/jq-f-test.sh                                              \
         tests/no-main-program.jq tests/yes-main-program.jq
 

configure.ac

@@ -283,7 +283,44 @@ AC_SUBST(onig_LDFLAGS)
 AM_CONDITIONAL([BUILD_ONIGURUMA], [test "x$build_oniguruma" = xyes])
 AM_CONDITIONAL([WITH_ONIGURUMA], [test "x$with_oniguruma" != xno])
 
+dnl PCRE2
+dnl TODO: proper check
+AC_ARG_WITH([pcre2],
+   [AS_HELP_STRING([--with-pcre2], [Use PCRE2 library])],
+   [with_pcre2=yes],
+   [with_pcre2=no])
+
+if test "$with_pcre2" = yes; then
+   AC_CHECK_PROG(PCRE2_CONFIG, pcre2-config, pcre2-config)
+   AC_ARG_WITH(pcre2-config,
+               AS_HELP_STRING([--with-pcre2-config=PATH],
+                              [Location of PCRE2 pcre2-config (auto)]),
+               [pcre2_config="$withval"],
+               [pcre2_config=""])
+
+   if test "x$pcre2_config" != "x" ; then
+      AC_MSG_CHECKING(for $pcre2_config)
+
+      if test -f $pcre2_config ; then
+         PCRE2_CONFIG=$pcre2_config
+         AC_MSG_RESULT(yes)
+      else
+         AC_MSG_RESULT(no - searching PATH)
+      fi
+   fi
+   if test "x$PCRE2_CONFIG" = "x"; then
+      AC_CHECK_PROGS(PCRE2_CONFIG, pcre2-config)
+   fi
+   pcre2_CFLAGS=`$PCRE2_CONFIG --cflags`
+   pcre2_LDFLAGS=`$PCRE2_CONFIG --libs8`
+   AC_SUBST(pcre2_CFLAGS)
+   AC_SUBST(pcre2_LDFLAGS)
+   dnl condtional?
+   AC_DEFINE([HAVE_PCRE2],1,[Define to 1 if the system includes pcre2])
+fi
+
+AM_CONDITIONAL([WITH_PCRE2], [test "x$with_pcre2" != xno])
+
 AC_CONFIG_MACRO_DIRS([config/m4 m4])
 AC_CONFIG_FILES([Makefile libjq.pc])
 AC_OUTPUT
-