Add GCC compatible preprocessing function

cxxheaderparser/lexer.py

@@ -17,7 +17,7 @@ class LexError(CxxParseError):
 else:
     Protocol = object
 
-_line_re = re.compile(r'^\#[\t ]*line (\d+) "(.*)"')
+_line_re = re.compile(r'^\#[\t ]*(line)? (\d+) "(.*)"')
 _multicomment_re = re.compile("\n[\\s]+\\*")
 
 
@@ -448,8 +448,8 @@ def t_PP_DIRECTIVE(self, t: LexToken):
         # handle line macros
         m = _line_re.match(t.value)
         if m:
-            self.filename = m.group(2)
-            self.line_offset = 1 + self.lex.lineno - int(m.group(1))
+            self.filename = m.group(3)
+            self.line_offset = 1 + self.lex.lineno - int(m.group(2))
             return None
         # ignore C++23 warning directive
         if t.value.startswith("#warning"):

cxxheaderparser/preprocessor.py

@@ -1,15 +1,99 @@
 """
-Contains optional preprocessor support via pcpp
+Contains optional preprocessor support functions
 """
 
 import io
 import re
 import os
+import subprocess
+import sys
 import typing
+
 from .options import PreprocessorFunction
 
 from pcpp import Preprocessor, OutputDirective, Action
 
+#
+# GCC preprocessor support
+#
+
+
+def _gcc_filter(fname: str, fp: typing.TextIO) -> str:
+    new_output = io.StringIO()
+    keep = True
+
+    for line in fp:
+        if line.startswith("# "):
+            last_quote = line.rfind('"')
+            if last_quote != -1:
+                keep = line[:last_quote].endswith(fname)
+
+        if keep:
+            new_output.write(line)
+
+    new_output.seek(0)
+    return new_output.read()
+
+
+def make_gcc_preprocessor(
+    *,
+    defines: typing.List[str] = [],
+    include_paths: typing.List[str] = [],
+    retain_all_content: bool = False,
+    encoding: typing.Optional[str] = None,
+    gcc_args: typing.List[str] = ["g++"],
+    print_cmd: bool = True,
+) -> PreprocessorFunction:
+    """
+    Creates a preprocessor function that uses g++ to preprocess the input text.
+
+    gcc is a high performance and accurate precompiler, but if an #include
+    directive or other oddity exists in your input it will throw an error.
+
+    :param defines: list of #define macros specified as "key value"
+    :param include_paths: list of directories to search for included files
+    :param retain_all_content: If False, only the parsed file content will be retained
+    :param encoding: If specified any include files are opened with this encoding
+    :param gcc_args: This is the path to G++ and any extra args you might want
+    :param print_cmd: Prints the gcc command as its executed
+
+    """
+
+    if not encoding:
+        encoding = "utf-8"
+
+    def _preprocess_file(filename: str, content: str) -> str:
+        cmd = gcc_args + ["-w", "-E", "-C"]
+
+        for p in include_paths:
+            cmd.append(f"-I{p}")
+        for d in defines:
+            cmd.append(f"-D{d.replace(' ', '=')}")
+
+        kwargs = {"encoding": encoding}
+        if filename == "<str>":
+            cmd.append("-")
+            filename = "<stdin>"
+            kwargs["input"] = content
+        else:
+            cmd.append(filename)
+
+        if print_cmd:
+            print("+", " ".join(cmd), file=sys.stderr)
+
+        result: str = subprocess.check_output(cmd, **kwargs)  # type: ignore
+        if not retain_all_content:
+            result = _gcc_filter(filename, io.StringIO(result))
+
+        return result
+
+    return _preprocess_file
+
+
+#
+# PCPP preprocessor support
+#
+
 
 class PreprocessorError(Exception):
     pass
@@ -36,7 +120,7 @@ def on_comment(self, *ignored):
         return True
 
 
-def _filter_self(fname: str, fp: typing.TextIO) -> str:
+def _pcpp_filter(fname: str, fp: typing.TextIO) -> str:
     # the output of pcpp includes the contents of all the included files, which
     # isn't what a typical user of cxxheaderparser would want, so we strip out
     # the line directives and any content that isn't in our original file
@@ -69,6 +153,13 @@ def make_pcpp_preprocessor(
     Creates a preprocessor function that uses pcpp (which must be installed
     separately) to preprocess the input text.
 
+    If missing #include files are encountered, this preprocessor will ignore the
+    error. This preprocessor is pure python so it's very portable, and is a good
+    choice if performance isn't critical.
+
+    :param defines: list of #define macros specified as "key value"
+    :param include_paths: list of directories to search for included files
+    :param retain_all_content: If False, only the parsed file content will be retained
     :param encoding: If specified any include files are opened with this encoding
     :param passthru_includes: If specified any #include directives that match the
                               compiled regex pattern will be part of the output.
@@ -119,6 +210,6 @@ def _preprocess_file(filename: str, content: str) -> str:
                         filename = filename.replace(os.sep, "/")
                     break
 
-            return _filter_self(filename, fp)
+            return _pcpp_filter(filename, fp)
 
     return _preprocess_file

tests/test_preprocessor.py

@@ -1,9 +1,12 @@
 import os
 import pathlib
+import pytest
 import re
+import shutil
+import typing
 
-from cxxheaderparser.options import ParserOptions
-from cxxheaderparser.preprocessor import make_pcpp_preprocessor
+from cxxheaderparser.options import ParserOptions, PreprocessorFunction
+from cxxheaderparser import preprocessor
 from cxxheaderparser.simple import (
     NamespaceScope,
     ParsedData,
@@ -22,12 +25,29 @@
 )
 
 
-def test_basic_preprocessor() -> None:
+@pytest.fixture(params=["gcc", "pcpp"])
+def make_pp(request) -> typing.Callable[..., PreprocessorFunction]:
+    param = request.param
+    if param == "gcc":
+        gcc_path = shutil.which("g++")
+        if not gcc_path:
+            pytest.skip("g++ not found")
+        return preprocessor.make_gcc_preprocessor
+    elif param == "pcpp":
+        return preprocessor.make_pcpp_preprocessor
+    else:
+        assert False
+
+
+def test_basic_preprocessor(
+    make_pp: typing.Callable[..., PreprocessorFunction]
+) -> None:
     content = """
       #define X 1
       int x = X;
     """
-    options = ParserOptions(preprocessor=make_pcpp_preprocessor())
+
+    options = ParserOptions(preprocessor=make_pp())
     data = parse_string(content, cleandoc=True, options=options)
 
     assert data == ParsedData(
@@ -45,7 +65,10 @@ def test_basic_preprocessor() -> None:
     )
 
 
-def test_preprocessor_omit_content(tmp_path: pathlib.Path) -> None:
+def test_preprocessor_omit_content(
+    make_pp: typing.Callable[..., PreprocessorFunction],
+    tmp_path: pathlib.Path,
+) -> None:
     """Ensure that content in other headers is omitted"""
     h_content = '#include "t2.h"' "\n" "int x = X;\n"
     h2_content = "#define X 2\n" "int omitted = 1;\n"
@@ -56,7 +79,7 @@ def test_preprocessor_omit_content(tmp_path: pathlib.Path) -> None:
     with open(tmp_path / "t2.h", "w") as fp:
         fp.write(h2_content)
 
-    options = ParserOptions(preprocessor=make_pcpp_preprocessor())
+    options = ParserOptions(preprocessor=make_pp())
     data = parse_file(tmp_path / "t1.h", options=options)
 
     assert data == ParsedData(
@@ -74,7 +97,10 @@ def test_preprocessor_omit_content(tmp_path: pathlib.Path) -> None:
     )
 
 
-def test_preprocessor_omit_content2(tmp_path: pathlib.Path) -> None:
+def test_preprocessor_omit_content2(
+    make_pp: typing.Callable[..., PreprocessorFunction],
+    tmp_path: pathlib.Path,
+) -> None:
     """
     Ensure that content in other headers is omitted while handling pcpp
     relative path quirk
@@ -91,9 +117,7 @@ def test_preprocessor_omit_content2(tmp_path: pathlib.Path) -> None:
     with open(tmp_path2 / "t2.h", "w") as fp:
         fp.write(h2_content)
 
-    options = ParserOptions(
-        preprocessor=make_pcpp_preprocessor(include_paths=[str(tmp_path)])
-    )
+    options = ParserOptions(preprocessor=make_pp(include_paths=[str(tmp_path)]))
 
     # Weirdness happens here
     os.chdir(tmp_path)
@@ -114,7 +138,9 @@ def test_preprocessor_omit_content2(tmp_path: pathlib.Path) -> None:
     )
 
 
-def test_preprocessor_encoding(tmp_path: pathlib.Path) -> None:
+def test_preprocessor_encoding(
+    make_pp: typing.Callable[..., PreprocessorFunction], tmp_path: pathlib.Path
+) -> None:
     """Ensure we can handle alternate encodings"""
     h_content = b"// \xa9 2023 someone\n" b'#include "t2.h"' b"\n" b"int x = X;\n"
 
@@ -126,7 +152,7 @@ def test_preprocessor_encoding(tmp_path: pathlib.Path) -> None:
     with open(tmp_path / "t2.h", "wb") as fp:
         fp.write(h2_content)
 
-    options = ParserOptions(preprocessor=make_pcpp_preprocessor(encoding="cp1252"))
+    options = ParserOptions(preprocessor=make_pp(encoding="cp1252"))
     data = parse_file(tmp_path / "t1.h", options=options, encoding="cp1252")
 
     assert data == ParsedData(
@@ -155,7 +181,9 @@ def test_preprocessor_passthru_includes(tmp_path: pathlib.Path) -> None:
         fp.write("")
 
     options = ParserOptions(
-        preprocessor=make_pcpp_preprocessor(passthru_includes=re.compile(".+"))
+        preprocessor=preprocessor.make_pcpp_preprocessor(
+            passthru_includes=re.compile(".+")
+        )
     )
     data = parse_file(tmp_path / "t1.h", options=options)