Add support for running on Windows

.gitattributes

@@ -0,0 +1 @@
+*.sql binary

.github/workflows/ci.yml

@@ -14,7 +14,7 @@ jobs:
         valgrind: [valgrind,no-valgrind]
     steps:
     - name: Check out code
-      uses: actions/checkout@v2
+      uses: actions/checkout@v4
     - name: Cache protobuf library
       if: matrix.protobuf_lib == 'protobuf-cpp'
       id: cache-protobuf
@@ -71,3 +71,40 @@ jobs:
       run: make $FLAGS
       env:
         FLAGS: ${{ format('{0} {1} {2}', steps.make_flags.outputs.proto_flags, steps.make_flags.outputs.compiler_flags, steps.make_flags.outputs.debug_flags) }}
+  build_windows_msvc:
+    runs-on: windows-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        arch: [x64, x86]
+    steps:
+    - name: Configure MSVC developer console
+      uses: ilammy/msvc-dev-cmd@v1
+      with:
+        arch: ${{ matrix.arch }}
+    - name: Check out code
+      uses: actions/checkout@v4
+    - name: Build and run tests
+      run: nmake /F Makefile.msvc
+  build_windows_msys2:
+    runs-on: windows-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        sys: [mingw64, mingw32, ucrt64, clang64]
+    steps:
+    - name: Set up MSYS2 and compiler
+      uses: msys2/setup-msys2@v2
+      with:
+        msystem: ${{matrix.sys}}
+        pacboy: >-
+          toolchain:p
+        install: >-
+          make
+          diffutils
+    - name: Check out code
+      uses: actions/checkout@v4
+    - name: Build and run tests
+      shell: msys2 {0}
+      run: |
+        make

.gitignore

@@ -3,6 +3,8 @@ postgres.tar.bz2
 
 *.o
 *.a
+*.obj
+pg_query.lib
 
 examples/*
 !examples/*.c
@@ -17,3 +19,5 @@ test/*.dSYM
 
 tmp/*
 !tmp/.gitkeep
+
+.vs/*

Makefile

@@ -8,6 +8,7 @@ PGDIRBZ2 = $(root_dir)/tmp/postgres.tar.bz2
 
 PG_VERSION = 16.1
 PG_VERSION_MAJOR = $(call word-dot,$(PG_VERSION),1)
+PG_VERSION_NUM = 160001
 PROTOC_VERSION = 25.1
 
 VERSION = 5.0.0
@@ -30,10 +31,13 @@ else
 endif
 
 SRC_FILES := $(wildcard src/*.c src/postgres/*.c) vendor/protobuf-c/protobuf-c.c vendor/xxhash/xxhash.c protobuf/pg_query.pb-c.c
-NOT_OBJ_FILES := src/pg_query_enum_defs.o src/pg_query_fingerprint_defs.o src/pg_query_fingerprint_conds.o src/pg_query_outfuncs_defs.o src/pg_query_outfuncs_conds.o src/pg_query_readfuncs_defs.o src/pg_query_readfuncs_conds.o src/postgres/guc-file.o src/postgres/scan.o src/pg_query_json_helper.o $(patsubst %.c,%.o,$(wildcard src/postgres/src_backend_nodes_*.funcs.c))
-OBJ_FILES := $(filter-out $(NOT_OBJ_FILES), $(SRC_FILES:.c=.o))
+OBJ_FILES := $(SRC_FILES:.c=.o)
 
-override CFLAGS += -g -I. -I./vendor -I./src/postgres/include -Wall -Wno-unused-function -Wno-unused-value -Wno-unused-variable -fno-strict-aliasing -fwrapv -fPIC
+override CFLAGS += -g -I. -I./vendor -I./src/include -I./src/postgres/include -Wall -Wno-unused-function -Wno-unused-value -Wno-unused-variable -fno-strict-aliasing -fwrapv -fPIC
+
+ifeq ($(OS),Windows_NT)
+override CFLAGS += -I./src/postgres/include/port/win32
+endif
 
 override PG_CONFIGURE_FLAGS += -q --without-readline --without-zlib --without-icu
 
@@ -80,7 +84,7 @@ CC ?= cc
 # Experimental use of Protobuf C++ library, primarily used to validate JSON output matches Protobuf JSON mapping
 CXX_SRC_FILES := src/pg_query_outfuncs_protobuf_cpp.cc protobuf/pg_query.pb.cc
 ifeq ($(USE_PROTOBUF_CPP),1)
-	override CXXFLAGS += `pkg-config --cflags protobuf` -I. -I./src/postgres/include -DHAVE_PTHREAD -std=c++17 -Wall -Wno-unused-function -Wno-zero-length-array -Wno-c99-extensions -fwrapv -fPIC
+	override CXXFLAGS += `pkg-config --cflags protobuf` -I. -I./src/include -I./src/postgres/include -DHAVE_PTHREAD -std=c++17 -Wall -Wno-unused-function -Wno-zero-length-array -Wno-c99-extensions -fwrapv -fPIC
 	ifeq ($(DEBUG),1)
 		override CXXFLAGS += -O0 -g
 	else
@@ -90,7 +94,7 @@ ifeq ($(USE_PROTOBUF_CPP),1)
 
 	# Don't use regular Protobuf-C or JSON implementation (instead implement the same methods using the C++ library)
 	SRC_FILES := $(filter-out src/pg_query_outfuncs_json.c src/pg_query_outfuncs_protobuf.c, $(SRC_FILES))
-	OBJ_FILES := $(filter-out $(NOT_OBJ_FILES), $(SRC_FILES:.c=.o)) $(CXX_SRC_FILES:.cc=.o)
+	OBJ_FILES := $(SRC_FILES:.c=.o) $(CXX_SRC_FILES:.cc=.o)
 else
 	# Make sure we always clean C++ object files
 	CLEANOBJS += $(CXX_SRC_FILES:.cc=.o)
@@ -132,21 +136,38 @@ $(PGDIR):
 	echo "#define StaticAssertDecl(condition, errmessage)" >> $(PGDIR)/src/include/c.h
 	# Add pg_config.h overrides
 	cat scripts/pg_config_overrides.h >> $(PGDIR)/src/include/pg_config.h
+	# Only define strlcpy when needed
+	sed -i "" '$(shell echo 's/\#include "c.h"/#include "c.h"\n#if HAVE_DECL_STRLCPY == 0/')' $(PGDIR)/src/port/strlcpy.c
+	echo "#endif // HAVE_DECL_STRLCPY == 0" >> $(PGDIR)/src/port/strlcpy.c
+	# Define symbols needed by elog.c that are commonly defined by win32/signal.c
+	echo "#ifdef WIN32" >> $(PGDIR)/src/backend/utils/error/elog.c
+	echo "volatile int pg_signal_queue;" >> $(PGDIR)/src/backend/utils/error/elog.c
+	echo "int pg_signal_mask;" >> $(PGDIR)/src/backend/utils/error/elog.c
+	echo "void pgwin32_dispatch_queued_signals(void) {}" >> $(PGDIR)/src/backend/utils/error/elog.c
+	echo "#endif" >> $(PGDIR)/src/backend/utils/error/elog.c
 
 extract_source: $(PGDIR)
 	-@ $(RM) -rf ./src/postgres/
 	mkdir ./src/postgres
 	mkdir ./src/postgres/include
 	LIBCLANG=/Library/Developer/CommandLineTools/usr/lib/libclang.dylib ruby ./scripts/extract_source.rb $(PGDIR)/ ./src/postgres/
-	cp $(PGDIR)/src/include/storage/dsm_impl.h ./src/postgres/include/storage
-	cp $(PGDIR)/src/include/port/atomics/arch-x86.h ./src/postgres/include/port/atomics
-	cp $(PGDIR)/src/include/port/atomics/arch-arm.h ./src/postgres/include/port/atomics
-	cp $(PGDIR)/src/include/port/atomics/arch-ppc.h ./src/postgres/include/port/atomics
-	touch ./src/postgres/guc-file.c
+	# Override OS-specific pg_config_os.h to only load Win32 logic (the primary port logic that matters for libpg_query), if needed
+	echo "#if defined(_WIN32) || defined(_WIN64)" > ./src/postgres/include/pg_config_os.h
+	echo "#include \"port/win32.h\"" >> ./src/postgres/include/pg_config_os.h
+	# Don't mark anything as visible based on how Postgres defines it
+	echo "#undef PGDLLIMPORT" >> ./src/postgres/include/pg_config_os.h
+	echo "#undef PGDLLEXPORT" >> ./src/postgres/include/pg_config_os.h
+	# Avoid getting incorrect sigsetjmp overrides
+	echo "#ifdef __clang__" >> ./src/postgres/include/pg_config_os.h
+	echo "#undef __MINGW64__" >> ./src/postgres/include/pg_config_os.h
+	echo "#endif /* __clang__ */" >> ./src/postgres/include/pg_config_os.h
+	echo "#endif" >> ./src/postgres/include/pg_config_os.h
+	# Adjust version string to ignore differences in build environments
+	sed -i "" '$(shell echo 's/\#define PG_VERSION_STR .*/#define PG_VERSION_STR "PostgreSQL $(PG_VERSION) \(libpg_query\)"/')' ./src/postgres/include/pg_config.h
 	# Copy version information so its easily accessible
-	sed -i "" '$(shell echo 's/\#define PG_MAJORVERSION .*/'`grep "\#define PG_MAJORVERSION " ./src/postgres/include/pg_config.h`'/')' pg_query.h
-	sed -i "" '$(shell echo 's/\#define PG_VERSION .*/'`grep "\#define PG_VERSION " ./src/postgres/include/pg_config.h`'/')' pg_query.h
-	sed -i "" '$(shell echo 's/\#define PG_VERSION_NUM .*/'`grep "\#define PG_VERSION_NUM " ./src/postgres/include/pg_config.h`'/')' pg_query.h
+	sed -i "" '$(shell echo 's/\#define PG_MAJORVERSION .*/#define PG_MAJORVERSION "$(PG_VERSION_MAJOR)"/')' pg_query.h
+	sed -i "" '$(shell echo 's/\#define PG_VERSION .*/#define PG_VERSION "$(PG_VERSION)"/')' pg_query.h
+	sed -i "" '$(shell echo 's/\#define PG_VERSION_NUM .*/#define PG_VERSION_NUM $(PG_VERSION_NUM)/')' pg_query.h
 	# Copy regress SQL files so we can use them in tests
 	rm -f ./test/sql/postgres_regress/*.sql
 	cp $(PGDIR)/src/test/regress/sql/*.sql ./test/sql/postgres_regress/

Makefile.msvc

@@ -0,0 +1,108 @@
+# Makefile for "nmake", part of Microsoft Visual Studio Compiler (MSVC) on Windows
+
+TARGET = pg_query
+ARLIB = $(TARGET).lib
+
+SRC_FILES = src/*.c src/postgres/*.c vendor/protobuf-c/protobuf-c.c vendor/xxhash/xxhash.c protobuf/pg_query.pb-c.c
+
+CFLAGS = -I. -I./vendor -I./src/postgres/include -I./src/include -I./src/postgres/include/port/win32 -I./src/postgres/include/port/win32_msvc
+
+RM = del
+
+all: examples test build
+
+build: $(ARLIB)
+
+clean:
+	$(RM) *.obj
+	$(RM) pg_query.lib
+
+.PHONY: all clean build build_shared extract_source examples test install
+
+$(ARLIB): clean $(SRC_FILES)
+	$(CC) $(CFLAGS) /c $(SRC_FILES)
+    lib /OUT:pg_query.lib *.obj
+
+EXAMPLES = examples/simple examples/scan examples/normalize examples/simple_error examples/normalize_error examples/simple_plpgsql
+examples: $(EXAMPLES)
+	.\examples\simple
+	.\examples\scan
+	.\examples\normalize
+	.\examples\simple_error
+	.\examples\normalize_error
+	.\examples\simple_plpgsql
+
+examples/simple: examples/simple.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ examples/simple.c $(ARLIB)
+
+examples/scan: examples/scan.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ examples/scan.c $(ARLIB)
+
+examples/normalize: examples/normalize.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ examples/normalize.c $(ARLIB)
+
+examples/simple_error: examples/simple_error.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ examples/simple_error.c $(ARLIB)
+
+examples/normalize_error: examples/normalize_error.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ examples/normalize_error.c $(ARLIB)
+
+examples/simple_plpgsql: examples/simple_plpgsql.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ examples/simple_plpgsql.c $(ARLIB)
+
+TESTS = test/deparse test/fingerprint test/fingerprint_opts test/normalize test/parse test/parse_opts test/parse_protobuf test/parse_protobuf_opts test/parse_plpgsql test/scan test/split
+test: $(TESTS)
+	.\test\deparse
+	.\test\fingerprint
+	.\test\fingerprint_opts
+	.\test\normalize
+	.\test\parse
+	.\test\parse_opts
+	.\test\parse_protobuf
+	.\test\parse_protobuf_opts
+	.\test\scan
+	.\test\split
+
+# Doesn't work because of C2026: string too big, trailing characters truncated
+#test/complex: test/complex.c $(ARLIB)
+# We have "-Isrc/" because this test uses pg_query_fingerprint_with_opts
+#	$(CC) $(CFLAGS) -o $@ -Isrc/ test/complex.c $(ARLIB)
+
+# Doesn't work since this requires pthreads
+#test/concurrency: test/concurrency.c test/parse_tests.c $(ARLIB)
+#	$(CC) $(CFLAGS) -o $@ test/concurrency.c $(ARLIB)
+
+test/deparse: test/deparse.c test/deparse_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/deparse.c $(ARLIB)
+
+test/fingerprint: test/fingerprint.c test/fingerprint_tests.c $(ARLIB)
+# We have "-Isrc/" because this test uses pg_query_fingerprint_with_opts
+	$(CC) $(CFLAGS) -o $@ -Isrc/ test/fingerprint.c $(ARLIB)
+
+test/fingerprint_opts: test/fingerprint_opts.c test/fingerprint_opts_tests.c $(ARLIB)
+# We have "-Isrc/" because this test uses pg_query_fingerprint_with_opts
+	$(CC) $(CFLAGS) -o $@ -Isrc/ test/fingerprint_opts.c $(ARLIB)
+
+test/normalize: test/normalize.c test/normalize_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/normalize.c $(ARLIB)
+
+test/parse: test/parse.c test/parse_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/parse.c $(ARLIB)
+
+test/parse_opts: test/parse_opts.c test/parse_opts_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/parse_opts.c $(ARLIB)
+
+test/parse_plpgsql: test/parse_plpgsql.c test/parse_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/parse_plpgsql.c $(ARLIB)
+
+test/parse_protobuf: test/parse_protobuf.c test/parse_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/parse_protobuf.c $(ARLIB)
+
+test/parse_protobuf_opts: test/parse_protobuf_opts.c test/parse_opts_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/parse_protobuf_opts.c $(ARLIB)
+
+test/scan: test/scan.c test/scan_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/scan.c $(ARLIB)
+
+test/split: test/split.c test/split_tests.c $(ARLIB)
+	$(CC) $(CFLAGS) -o $@ test/split.c $(ARLIB)

scripts/extract_source.rb

@@ -103,7 +103,7 @@ def run
       @basepath + 'src/port/win32env.c', # Win32 only
       @basepath + 'src/port/win32security.c', # Win32 only
       @basepath + 'src/port/gettimeofday.c', # Win32 only
-      @basepath + 'src/port/strlcpy.c', # Not needed and conflicts with available function
+      @basepath + 'src/port/strnlen.c', # Not needed and conflicts with available function
       @basepath + 'src/port/strlcat.c', # Not needed and conflicts with available function
       @basepath + 'src/port/unsetenv.c', # Not needed and conflicts with available function
       @basepath + 'src/port/getaddrinfo.c', # Not needed and conflicts with available function
@@ -214,14 +214,33 @@ def self.analysis_filename(filename, basepath)
 
   def analyze_file(file)
     index = FFI::Clang::Index.new(true, true)
-    translation_unit = index.parse_translation_unit(file, [
+    flags = [
       '-I', @basepath + 'src/include',
       '-I', '/usr/local/opt/openssl/include',
       '-I', `xcrun --sdk macosx --show-sdk-path`.strip + '/usr/include',
       '-DDLSUFFIX=".bundle"',
       '-g',
-      '-DUSE_ASSERT_CHECKING'
-    ])
+      '-ferror-limit=0',
+      '-DUSE_ASSERT_CHECKING',
+      # EXEC_BACKEND is used on Windows, and can always be safely set during code analysis
+      '-DEXEC_BACKEND',
+    ]
+
+    # For certain files, use WIN32 define - we can't always do this since it pulls unnecessary code in other cases
+    if file == @basepath + 'src/backend/utils/error/elog.c' || file == @basepath + 'src/backend/utils/mb/mbutils.c'
+      flags << '-DWIN32'
+      flags << '-D__CYGWIN__' # Avoid pulling in win32_port.h (which includes a bunch of system headers we don't actually have)
+    end
+
+    # To override built-ins, we must avoid pulling in any system headers, so pretend we already defined c.h
+    if file == @basepath + 'src/port/strlcpy.c'
+      flags << '-DC_H'
+      flags << '-DHAVE_DECL_STRLCPY=0'
+      flags << '-Dsize_t=unsigned'
+    end
+
+    translation_unit = index.parse_translation_unit(file, flags)
+
     cursor = translation_unit.cursor
 
     func_cursor = nil
@@ -363,7 +382,7 @@ def deep_resolve(method_name, depth: 0, trail: [], global_resolved_by_parent: []
   end
 
   def special_include_file?(filename)
-    filename[/\/(reg(c|e)_[\w_]+|guc-file|qsort_tuple|repl_scanner|levenshtein|bootscanner|like_match)\.c$/] || filename[/\/[\w_]+_impl.h$/]
+    filename[/\/(reg(c|e)_[\w_]+|guc-file|qsort_tuple|repl_scanner|levenshtein|bootscanner|like_match)\.c$/] || filename[/\/[\w_]+\.funcs.c$/] || filename[/\/[\w_]+_impl.h$/]
   end
 
   def write_out
@@ -427,20 +446,22 @@ def write_out
       end
       all_thread_local_variables += file_thread_local_variables
 
-      if special_include_file?(filename)
-        out_name = File.basename(filename)
-      else
+      unless special_include_file?(filename)
         out_name = filename.gsub(%r{^#{@basepath}}, '').gsub('/', '_')
+        File.write(@out_path + out_name, str)
       end
-
-      File.write(@out_path + out_name, str)
     end
 
     #return
 
-    @include_files_to_output.each do |include_file|
-      next if special_include_file?(include_file)
+    additional_includes = Dir.glob(@basepath + 'src/include/storage/dsm_impl.h') +
+      Dir.glob(@basepath + 'src/include/port/atomics/**/*.h') +
+      Dir.glob(@basepath + 'src/include/port/win32/**/*.h') +
+      Dir.glob(@basepath + 'src/include/port/win32_msvc/**/*.h') +
+      Dir.glob(@basepath + 'src/include/port/win32.h') +
+      Dir.glob(@basepath + 'src/include/port/win32_port.h')
 
+    (@include_files_to_output + additional_includes).each do |include_file|
       if include_file.start_with?(@basepath + 'src/include')
         out_file = @out_path + include_file.gsub(%r{^#{@basepath}src/}, '')
       else
@@ -449,7 +470,7 @@ def write_out
 
       code = File.read(include_file)
       all_thread_local_variables.each do |variable|
-        code.gsub!(/(PGDLLIMPORT|extern)\s+(const|volatile)?\s*(\w+)\s+(\*{0,2})#{variable}(\[\])?;/, "\\1 __thread \\2 \\3 \\4#{variable}\\5;")
+        code.gsub!(/(extern\s+)(PGDLLIMPORT\s+)?(const\s+)?(volatile\s+)?(\w+)\s+(\*{0,2})#{variable}(\[\])?;/, "\\1\\2__thread \\3\\4\\5 \\6#{variable}\\7;")
       end
 
       FileUtils.mkdir_p File.dirname(out_file)
@@ -539,6 +560,20 @@ def write_out
 }
 )) # We're always working with fn_rettype = VOIDOID, due to our use of plpgsql_compile_inline
 
+# Mocks REQUIRED for Windows support
+runner.mock('write_stderr', %(
+void
+write_stderr(const char *fmt,...)
+{
+	va_list	ap;
+	va_start(ap, fmt);
+	vfprintf(stderr, fmt, ap);
+	fflush(stderr);
+	va_end(ap);
+}
+)) # Avoid pulling in write_console/write_eventlog, and instead always output to stderr (like on POSIX)
+runner.mock('should_output_to_client', 'static inline bool should_output_to_client(int elevel) { return false; }') # Avoid pulling in postmaster.c, which has a bunch of Windows-specific code hidden behind a define
+
 ## ---
 
 # SQL Parsing
@@ -592,6 +627,17 @@ def write_out
 runner.deep_resolve('pg_rightmost_one_pos32')
 runner.deep_resolve('pg_popcount32')
 
+# Required for Windows support
+runner.deep_resolve('newNodeMacroHolder')
+runner.deep_resolve('pg_leftmost_one_pos')
+runner.deep_resolve('pg_rightmost_one_pos')
+runner.deep_resolve('pg_number_of_ones')
+runner.deep_resolve('GetMessageEncoding')
+runner.deep_resolve('strlcpy')
+runner.deep_resolve('pg_signal_queue')
+runner.deep_resolve('pg_signal_mask')
+runner.deep_resolve('pgwin32_dispatch_queued_signals')
+
 runner.write_out
 
 #puts runner.unresolved.inspect