Release 3.1.

release.sh: make the script NixOS-friendly, fix quoting in echo command.
Makefile.lib.am: remove nonexistent header from the list.
2023-07-19 22:50:22 +01:00 · 2023-07-19 22:42:25 +01:00 · 2023-07-19 21:47:14 +01:00 · 2023-07-18 22:56:45 +01:00 · 2023-07-13 23:22:55 +01:00 · 2023-07-13 22:59:47 +01:00
80 changed files with 6161 additions and 4450 deletions
@@ -11,6 +11,7 @@
 /m4
 /test_*
 /.build*
+/bazel-*

 # Ignore user defined CMake presets
 CMakeUserPresets.json
@@ -0,0 +1,45 @@
+package(default_visibility = ["//visibility:private"])
+
+genrule(
+    name = "re2c_config",
+    srcs = [
+        "CMakeLists.txt",
+        "//bazel:re2c_config.h",
+    ],
+    outs = ["config.h"],
+    cmd = """
+        cat $(location //:CMakeLists.txt) |\\
+            sed -rne 's/project\\(re2c VERSION ([^ ]+).*/#define PACKAGE_VERSION "\\1-bootstrap"/p' >> "$@"
+        cat $(location //bazel:re2c_config.h) >> "$@"
+    """,
+)
+
+cc_library(
+    name = "re2c_config_cc",
+    hdrs = ["config.h"],
+    defines = ["""RE2C_STDLIB_DIR='"$(BINDIR)/re2c.runfiles/re2c/include"'"""],
+    data = ["include/unicode_categories.re"],
+)
+
+cc_library(
+    name = "re2c_bootstrap_cc",
+    includes = ["bootstrap"],
+    hdrs = glob([
+        "bootstrap/src/**/*.h",
+        "src/**/*.h",
+    ]),
+    srcs = glob([
+        "bootstrap/src/**/*.cc",
+        "src/**/*.cc",
+    ]),
+    deps = [
+        "re2c_config_cc",  # Needed for 're2c_config_cc' data files
+    ],
+)
+
+cc_binary(
+    name = "re2c",
+    srcs = ["src/main.cc"],
+    deps = [":re2c_bootstrap_cc"],
+    visibility = ["//visibility:public"],
+)
@@ -52,9 +52,13 @@ The configure script has many options (to see them all, run
    Build re2rust (identical to `re2c --lang rust`). This is on by default.

  * `--enable-lexers`
-    Enable regeneration of *.re* files (as opposed to using bootstrap files).
-    This requires setting `RE2C_FOR_BUILD` to an existing re2c executable
-    path.
+    Enable regeneration of lexers from *.re* files (as opposed to using
+    pre-generated bootstrap files). This requires setting `RE2C_FOR_BUILD` to
+    an existing re2c executable path.
+
+  * `--enable-parsers`
+    Enable regeneration of parsers from *.ypp* files with bison (as opposed to
+    using pre-generated bootstrap files).

  * `--enable-docs`
    Enable regeneration of documentation (requires Python 3 Docutils).
@@ -152,13 +156,13 @@ specific to re2c:
    default.

  * `-DRE2C_REBUILD_LEXERS=yes`
-    Enable regeneration of *.re* files (as opposed to using bootstrap files).
-    This requires setting `-DRE2C_FOR_BUILD` to an existing re2c executable
-    path.
+    Enable regeneration of lexers from *.re* files (as opposed to using
+    pre-generated bootstrap files). This requires setting `-DRE2C_FOR_BUILD` to
+    an existing re2c executable path.

  * `-DRE2C_REBUILD_PARSERS=yes`
-    Enable regeneration of `parser.cc` and `parser.h` files (as opposed to
-    using bootstrap files). This requires an available Bison parser generator.
+    Enable regeneration of parsers from *.ypp* files with bison (as opposed to
+    using pre-generated bootstrap files).

  * `-DRE2C_REBUILD_DOCS=yes`
    Enable regeneration of documentation (requires Python 3 Docutils).
@@ -6,6 +6,104 @@ Changelog
 3.0x
 ----

+3.1 (2023-07-19)
+~~~~~~~~~~~~~~~~
+
+- Added capturing groups with leftmost greedy semantics:
+
+  + Enabled with ``--leftmost-captures`` option or ``re2c:leftmost_captures``
+    configuration
+    (`55de79d8 <https://github.com/skvadrik/re2c/commit/55de79d8>`_,
+    `3a98b543 <https://github.com/skvadrik/re2c/commit/3a98b543>`_).
+
+- Added non-capturing groups:
+
+  + Added new syntax ``(! ...)`` for non-capturing groups
+    (`1edd25d3 <https://github.com/skvadrik/re2c/commit/1edd25d3>`_,
+    `b813c9b4 <https://github.com/skvadrik/re2c/commit/b813c9b4>`_,
+    `338806b9 <https://github.com/skvadrik/re2c/commit/338806b9>`_).
+
+  + Added the ability to flip defaults: make ``(...)`` capturing and ``(! ...)``
+    non-capturing with ``--invert-captures`` option or ``re2c:invert_captures``
+    configuration
+    (`20030ff1 <https://github.com/skvadrik/re2c/commit/20030ff1>`_,
+    `ce756195 <https://github.com/skvadrik/re2c/commit/ce756195>`_).
+
+- Regenerated Unicode include header to support a newer standard
+  (`e3ec2597 <https://github.com/skvadrik/re2c/commit/e3ec2597>`_).
+
+- Published TDFA paper: https://arxiv.org/abs/2206.01398, co-authored with
+  Angelo Borsotti
+  (`fa94d9c7 <https://github.com/skvadrik/re2c/commit/fa94d9c7>`_).
+
+- Removed experimental algorithms that are superseded by TDFA(1) and
+  generally less efficient:
+
+  + Removed staDFA algorithm and deprecated ``--stadfa`` option
+    (`ac5c06cc <https://github.com/skvadrik/re2c/commit/ac5c06cc>`_).
+
+  + Removed TDFA(0) algorithm and deprecated ``--no-lookahead`` option
+    (`dc8f264a <https://github.com/skvadrik/re2c/commit/dc8f264a>`_).
+
+  + (libre2c) Removed backward-matching algorithm
+    (`27256be1 <https://github.com/skvadrik/re2c/commit/27256be1>`_).
+
+  + (libre2c) Removed Kuklewicz POSIX disambiguation algorithm
+    (`aa97b014 <https://github.com/skvadrik/re2c/commit/aa97b014>`_).
+
+  + (libre2c) Removed GTOP shortest path finding algorithm
+    (`511a030c <https://github.com/skvadrik/re2c/commit/511a030c>`_).
+
+- Bug fixes:
+
+  + Fixed parsing of raw UTF-8 characters in Flex compatibility mode
+    (`d87f86ed <https://github.com/skvadrik/re2c/commit/d87f86ed>`_).
+
+  + Added header file to the dependencies generated with ``--depfile`` option
+    (`f807f763 <https://github.com/skvadrik/re2c/commit/f807f763>`_
+    and `2dda36aa <https://github.com/skvadrik/re2c/commit/2dda36aa>`_).
+
+  + Fixed stack overflow on large regular expressions by rewriting recursive
+    functions in iterative form
+    (`46a9b4c4 <https://github.com/skvadrik/re2c/commit/46a9b4c4>`_,
+    `aaf68292 <https://github.com/skvadrik/re2c/commit/aaf68292>`_,
+    `02e5d797 <https://github.com/skvadrik/re2c/commit/02e5d797>`_,
+    `5fffb187 <https://github.com/skvadrik/re2c/commit/5fffb187>`_)
+    and limited stack to 256K on GithubActions CI
+    (`111ee5da <https://github.com/skvadrik/re2c/commit/111ee5da>`_).
+
+- Build system:
+
+  + Added minimal http://bazel.build integration
+    (`3205c867 <https://github.com/skvadrik/re2c/commit/3205c867>`_).
+
+  + Added configure option ``--enable-parsers`` that regenerates bison parsers
+    (`9e0dbd3c <https://github.com/skvadrik/re2c/commit/9e0dbd3c>`_).
+
+  + Added CMake option ``RE2C_REBUILD_PARSERS``
+    (`6e91c22d <https://github.com/skvadrik/re2c/commit/6e91c22d>`_).
+
+  + With CMake, fixed documentation generation on Windows.
+
+- Codebase improvements:
+
+  + Moved the entire codebase to C++11.
+  + Added uniform error handling (return codes are now properly checked and
+    returned to the caller).
+  + Reorganized codegen subsystem in four well-defined phases (analyze,
+    generate, fixup, render) and separated codegen from parsing phase.
+  + Improved memory allocation by using slab allocators instead of global free
+    lists.
+  + Moved to pure API for bison parsers.
+  + Unified code style.
+
+- Testing:
+
+  + Added ``--verbose`` flag to run_tests.py and suppressed verbose output by
+    default.
+  + Multiple improvements of continuous testing with GithubActions.
+
+
 3.0 (2022-01-27)
 ~~~~~~~~~~~~~~~~

@@ -1,5 +1,5 @@
 cmake_minimum_required(VERSION 3.12)
-project(re2c VERSION 3.0 HOMEPAGE_URL "https://re2c.org/")
+project(re2c VERSION 3.1 HOMEPAGE_URL "https://re2c.org/")

 list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/cmake")

@@ -49,9 +49,10 @@ if(RE2C_REBUILD_DOCS)
    endif()
 endif()

-# checks for C++ compiler flags
+# Use C++11 standard. Note that without `set(CMAKE_CXX_EXTENSIONS OFF)` this enables
+# -std=gnu++11, not -std=c++11. Extensions are needed to enable POSIX functions on
+# Cygwin and NetBSD.
 set(CMAKE_CXX_STANDARD 11)
-set(CMAKE_CXX_EXTENSIONS OFF)
 set(CMAKE_POSITION_INDEPENDENT_CODE ON) # make sure object libraries work with shared libraries

 # needed for POSIX file API
@@ -38,7 +38,6 @@ libre2c_la_HDR = \
 	src/regexp/regexp.h \
 	src/regexp/rule.h \
 	src/regexp/tag.h \
-	src/compile.h \
 	src/skeleton/path.h \
 	src/skeleton/skeleton.h \
 	src/parse/ast.h \
@@ -0,0 +1,14 @@
+workspace(name = "re2c")
+
+load("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
+
+http_archive(
+    name = "bazel_skylib",
+    urls = [
+        "https://mirror.bazel.build/github.com/bazelbuild/bazel-skylib/releases/download/1.3.0/bazel-skylib-1.3.0.tar.gz",
+        "https://github.com/bazelbuild/bazel-skylib/releases/download/1.3.0/bazel-skylib-1.3.0.tar.gz",
+    ],
+    sha256 = "74d544d96f4a5bb630d465ca8bbcfe231e3594e5aae57e1edbf17a6eb3ca2506",
+)
+load("@bazel_skylib//:workspace.bzl", "bazel_skylib_workspace")
+bazel_skylib_workspace()
@@ -0,0 +1,13 @@
+package(default_visibility = ["//visibility:private"])
+
+load("@bazel_skylib//:bzl_library.bzl", "bzl_library")
+
+exports_files([
+    "re2c_config.h",
+])
+
+bzl_library(
+    name = "re2c_bzl",
+    srcs = ["re2c.bzl"],
+    visibility = ["//visibility:public"],
+)
@@ -0,0 +1,70 @@
+# re2c generator rules - assuming that re2c is available as '@re2c//:re2c'.
+
+load("@bazel_skylib//lib:paths.bzl", "paths")
+
+visibility("public")
+
+# Define supported extensions as defined by:
+# . https://bazel.build/reference/be/c-cpp#cc_library_args
+# . https://github.com/bazelbuild/bazel/blob/master/src/main/java/com/google/devtools/build/lib/rules/cpp/CppFileTypes.java
+_C_SRC_EXT = [".c", ".cc", ".cpp", ".cxx", ".c++", ".C", ".cu", ".cl"]
+_C_HDR_EXT = [".h", ".hh", ".hpp", ".ipp", ".hxx", ".h++", ".inc", ".inl", ".tlh", ".tli", ".H", ".tcc"]
+_RE2C_EXT = [".re", ".re2c"]
+
+def _re2c_impl(ctx):
+    srcs = ctx.files.srcs
+    if len(srcs) != 1:
+        fail("Attribute `srcs` must contain exactly 1 re2c source file ({exts}).".format(exts=", ".join(_RE2C_EXT)))
+    outs = ctx.outputs.outs
+    if not outs or [o for o in outs if paths.split_extension(o.path)[1] not in _C_SRC_EXT + _C_HDR_EXT]:
+        fail("Attribute `outs` must contain at least 1 C/C++ shource or header file ({exts}).".format(exts=", ".join(_C_SRC_EXT + _C_HDR_EXT)))
+    ctx.actions.run_shell(
+        outputs = outs + [err_file],
+        tools = [ctx.executable._re2c_tool],
+        inputs = ctx.files.srcs + ctx.files.data,
+        command = """
+            "{re2c}" "{src}" -o "{out}" {re2c_args} {includes}
+        """.format(
+            re2c = ctx.executable._re2c_tool.path,
+            src = srcs[0].path,
+            out = outs[0].path,
+            includes = " ".join(list(dict([("-I" + f.dirname, 1) for f in ctx.files.data]))),
+            re2c_args = " ".join(ctx.attr.re2c_args),
+        ),
+        mnemonic = "RE2C",
+        progress_message = "Generating re2c %{label}",
+    )
+
+_re2c_attr = {
+    "srcs": attr.label_list(
+        doc = "The re2c source.",
+        allow_files = _RE2C_EXT,
+        mandatory = True,
+    ),
+    "data": attr.label_list(
+        doc = "Other input data files (e.g. include files).",
+        allow_files = _C_SRC_EXT + _C_HDR_EXT,
+        mandatory = False,
+    ),
+    "outs": attr.output_list(
+        doc = "The generated output source.",
+        mandatory = True,
+    ),
+    "re2c_args": attr.string_list(
+        doc = "Arguments to re2c invocation.",
+    ),
+    "_re2c_tool": attr.label(
+        doc = "The target of the 're2c' executable.",
+        default = Label("//:re2c"),
+        allow_single_file = True,
+        executable = True,
+        cfg = "exec",
+    ),
+}
+
+re2c_gen = rule(
+    attrs = _re2c_attr,
+    doc = """Use re2c tool to create an output file (e.g. '.cc') from a '.re' input.""",
+    output_to_genfiles = True,
+    implementation = _re2c_impl,
+)
@@ -0,0 +1,4 @@
+#define HAVE_SYS_TYPES_H 1
+#define HAVE_SYS_STAT_H 1
+#define HAVE_FCNTL_H 1
+#define HAVE_UNISTD_H 1
@@ -438,10 +438,18 @@ sequence of code units is a sequence of standalone 1\-byte characters.
 With \fButf8\fP re2c handles input as UTF8\-encoded and recognizes multibyte
 characters.
 .TP
+.B \fB\-\-invert\-captures\fP
+Invert the meaning of capturing and non\-capturing groups. By default
+\fB(...)\fP is capturing and \fB(! ...)\fP is non\-capturing. With this option
+\fB(! ...)\fP is capturing and \fB(...)\fP is non\-capturing.
+.TP
 .B \fB\-\-lang <c | go | rust>\fP
 Specify the output language. Supported languages are C, Go and Rust.
 The default is C for re2c, Go for re2go and Rust for re2rust.
 .TP
+.B \fB\-\-leftmost\-captures\fP
+Enable submatch extraction with leftmost greedy capturing groups.
+.TP
 .B \fB\-\-location\-format <gnu | msvc>\fP
 Specify location format in messages.
 With \fBgnu\fP locations are printed as \(aqfilename:line:column: ...\(aq.
@@ -1343,6 +1351,10 @@ To disable indentation entirely, set this configuration to an empty string.
 Specifies the minimum amount of indentation to use. The default value is
 zero. The value should be a non\-negative integer number.
 .TP
+.B \fBre2c:invert\-captures\fP
+Same as the \fB\-\-invert\-captures\fP option, but can be configured on per\-block
+basis.
+.TP
 .B \fBre2c:label:prefix\fP, \fBre2c:labelprefix\fP
 Specifies the prefix used for DFA state labels. The default is \fByy\fP\&.
 .TP
@@ -1367,6 +1379,10 @@ Specifies the name of the optional label that follows \fBYYGETSTATE\fP switch
 in storable state mode (enabled with \fBre2c:state:nextlabel\fP). The default
 is \fByyNext\fP\&.
 .TP
+.B \fBre2c:leftmost\-captures\fP
+Same as the \fB\-\-leftmost\-captures\fP option, but can be configured on
+per\-block basis.
+.TP
 .B \fBre2c:lookahead\fP, \fBre2c:flags:lookahead\fP
 Deprecated (see the deprecated \fB\-\-no\-lookahead\fP option).
 .TP
@@ -1507,7 +1523,13 @@ re2c uses the following syntax for regular expressions:
 .IP \(bu 2
 \fBR{n,m}\fP repetition of \fBR\fP from \fBn\fP to \fBm\fP times
 .IP \(bu 2
-\fB(R)\fP just \fBR\fP; parentheses are used to override precedence or for POSIX\-style submatch
+\fB(R)\fP just \fBR\fP; parentheses are used to override precedence.
+If submatch extraction is enabled, \fB(R)\fP is a capturing or a
+non\-capturing group depending on \fB\-\-invert\-captures\fP option.
+.IP \(bu 2
+\fB(!R)\fP
+If submatch extraction is enabled, \fB(!R)\fP is a non\-capturing or a
+capturing group depending on \fB\-\-invert\-captures\fP option.
 .IP \(bu 2
 \fBR S\fP concatenation: \fBR\fP followed by \fBS\fP
 .IP \(bu 2
@@ -1515,17 +1537,23 @@ re2c uses the following syntax for regular expressions:
 .IP \(bu 2
 \fBR / S\fP lookahead: \fBR\fP followed by \fBS\fP, but \fBS\fP is not consumed
 .IP \(bu 2
-\fBname\fP the regular expression defined as \fBname\fP (or literal string \fB\(dqname\(dq\fP in Flex compatibility mode)
+\fBname\fP the regular expression defined as \fBname\fP (or literal string
+\fB\(dqname\(dq\fP in Flex compatibility mode)
 .IP \(bu 2
-\fB{name}\fP the regular expression defined as \fBname\fP in Flex compatibility mode
+\fB{name}\fP the regular expression defined as \fBname\fP in Flex
+compatibility mode
 .IP \(bu 2
-\fB@stag\fP an \fIs\-tag\fP: saves the last input position at which \fB@stag\fP matches in a variable named \fBstag\fP
+\fB@stag\fP an \fIs\-tag\fP: saves the last input position at which \fB@stag\fP
+matches in a variable named \fBstag\fP
 .IP \(bu 2
-\fB#mtag\fP an \fIm\-tag\fP: saves all input positions at which \fB#mtag\fP matches in a variable named \fBmtag\fP
+\fB#mtag\fP an \fIm\-tag\fP: saves all input positions at which \fB#mtag\fP matches
+in a variable named \fBmtag\fP
 .UNINDENT
 .sp
-Character classes and string literals may contain the following escape sequences:
-\fB\ea\fP, \fB\eb\fP, \fB\ef\fP, \fB\en\fP, \fB\er\fP, \fB\et\fP, \fB\ev\fP, \fB\e\e\fP, octal escapes \fB\eooo\fP and hexadecimal escapes \fB\exhh\fP, \fB\euhhhh\fP and \fB\eUhhhhhhhh\fP\&.
+Character classes and string literals may contain the following escape
+sequences: \fB\ea\fP, \fB\eb\fP, \fB\ef\fP, \fB\en\fP, \fB\er\fP, \fB\et\fP, \fB\ev\fP, \fB\e\e\fP,
+octal escapes \fB\eooo\fP and hexadecimal escapes \fB\exhh\fP, \fB\euhhhh\fP and
+\fB\eUhhhhhhhh\fP\&.
 .SH HANDLING THE END OF INPUT
 .sp
 One of the main problems for the lexer is to know when to stop.
@@ -442,10 +442,18 @@ sequence of code units is a sequence of standalone 1\-byte characters.
 With \fButf8\fP re2c handles input as UTF8\-encoded and recognizes multibyte
 characters.
 .TP
+.B \fB\-\-invert\-captures\fP
+Invert the meaning of capturing and non\-capturing groups. By default
+\fB(...)\fP is capturing and \fB(! ...)\fP is non\-capturing. With this option
+\fB(! ...)\fP is capturing and \fB(...)\fP is non\-capturing.
+.TP
 .B \fB\-\-lang <c | go | rust>\fP
 Specify the output language. Supported languages are C, Go and Rust.
 The default is C for re2c, Go for re2go and Rust for re2rust.
 .TP
+.B \fB\-\-leftmost\-captures\fP
+Enable submatch extraction with leftmost greedy capturing groups.
+.TP
 .B \fB\-\-location\-format <gnu | msvc>\fP
 Specify location format in messages.
 With \fBgnu\fP locations are printed as \(aqfilename:line:column: ...\(aq.
@@ -1347,6 +1355,10 @@ To disable indentation entirely, set this configuration to an empty string.
 Specifies the minimum amount of indentation to use. The default value is
 zero. The value should be a non\-negative integer number.
 .TP
+.B \fBre2c:invert\-captures\fP
+Same as the \fB\-\-invert\-captures\fP option, but can be configured on per\-block
+basis.
+.TP
 .B \fBre2c:label:prefix\fP, \fBre2c:labelprefix\fP
 Specifies the prefix used for DFA state labels. The default is \fByy\fP\&.
 .TP
@@ -1371,6 +1383,10 @@ Specifies the name of the optional label that follows \fBYYGETSTATE\fP switch
 in storable state mode (enabled with \fBre2c:state:nextlabel\fP). The default
 is \fByyNext\fP\&.
 .TP
+.B \fBre2c:leftmost\-captures\fP
+Same as the \fB\-\-leftmost\-captures\fP option, but can be configured on
+per\-block basis.
+.TP
 .B \fBre2c:lookahead\fP, \fBre2c:flags:lookahead\fP
 Deprecated (see the deprecated \fB\-\-no\-lookahead\fP option).
 .TP
@@ -1511,7 +1527,13 @@ re2c uses the following syntax for regular expressions:
 .IP \(bu 2
 \fBR{n,m}\fP repetition of \fBR\fP from \fBn\fP to \fBm\fP times
 .IP \(bu 2
-\fB(R)\fP just \fBR\fP; parentheses are used to override precedence or for POSIX\-style submatch
+\fB(R)\fP just \fBR\fP; parentheses are used to override precedence.
+If submatch extraction is enabled, \fB(R)\fP is a capturing or a
+non\-capturing group depending on \fB\-\-invert\-captures\fP option.
+.IP \(bu 2
+\fB(!R)\fP
+If submatch extraction is enabled, \fB(!R)\fP is a non\-capturing or a
+capturing group depending on \fB\-\-invert\-captures\fP option.
 .IP \(bu 2
 \fBR S\fP concatenation: \fBR\fP followed by \fBS\fP
 .IP \(bu 2
@@ -1519,17 +1541,23 @@ re2c uses the following syntax for regular expressions:
 .IP \(bu 2
 \fBR / S\fP lookahead: \fBR\fP followed by \fBS\fP, but \fBS\fP is not consumed
 .IP \(bu 2
-\fBname\fP the regular expression defined as \fBname\fP (or literal string \fB\(dqname\(dq\fP in Flex compatibility mode)
+\fBname\fP the regular expression defined as \fBname\fP (or literal string
+\fB\(dqname\(dq\fP in Flex compatibility mode)
 .IP \(bu 2
-\fB{name}\fP the regular expression defined as \fBname\fP in Flex compatibility mode
+\fB{name}\fP the regular expression defined as \fBname\fP in Flex
+compatibility mode
 .IP \(bu 2
-\fB@stag\fP an \fIs\-tag\fP: saves the last input position at which \fB@stag\fP matches in a variable named \fBstag\fP
+\fB@stag\fP an \fIs\-tag\fP: saves the last input position at which \fB@stag\fP
+matches in a variable named \fBstag\fP
 .IP \(bu 2
-\fB#mtag\fP an \fIm\-tag\fP: saves all input positions at which \fB#mtag\fP matches in a variable named \fBmtag\fP
+\fB#mtag\fP an \fIm\-tag\fP: saves all input positions at which \fB#mtag\fP matches
+in a variable named \fBmtag\fP
 .UNINDENT
 .sp
-Character classes and string literals may contain the following escape sequences:
-\fB\ea\fP, \fB\eb\fP, \fB\ef\fP, \fB\en\fP, \fB\er\fP, \fB\et\fP, \fB\ev\fP, \fB\e\e\fP, octal escapes \fB\eooo\fP and hexadecimal escapes \fB\exhh\fP, \fB\euhhhh\fP and \fB\eUhhhhhhhh\fP\&.
+Character classes and string literals may contain the following escape
+sequences: \fB\ea\fP, \fB\eb\fP, \fB\ef\fP, \fB\en\fP, \fB\er\fP, \fB\et\fP, \fB\ev\fP, \fB\e\e\fP,
+octal escapes \fB\eooo\fP and hexadecimal escapes \fB\exhh\fP, \fB\euhhhh\fP and
+\fB\eUhhhhhhhh\fP\&.
 .SH HANDLING THE END OF INPUT
 .sp
 One of the main problems for the lexer is to know when to stop.
@@ -458,10 +458,18 @@ sequence of code units is a sequence of standalone 1\-byte characters.
 With \fButf8\fP re2c handles input as UTF8\-encoded and recognizes multibyte
 characters.
 .TP
+.B \fB\-\-invert\-captures\fP
+Invert the meaning of capturing and non\-capturing groups. By default
+\fB(...)\fP is capturing and \fB(! ...)\fP is non\-capturing. With this option
+\fB(! ...)\fP is capturing and \fB(...)\fP is non\-capturing.
+.TP
 .B \fB\-\-lang <c | go | rust>\fP
 Specify the output language. Supported languages are C, Go and Rust.
 The default is C for re2c, Go for re2go and Rust for re2rust.
 .TP
+.B \fB\-\-leftmost\-captures\fP
+Enable submatch extraction with leftmost greedy capturing groups.
+.TP
 .B \fB\-\-location\-format <gnu | msvc>\fP
 Specify location format in messages.
 With \fBgnu\fP locations are printed as \(aqfilename:line:column: ...\(aq.
@@ -1363,6 +1371,10 @@ To disable indentation entirely, set this configuration to an empty string.
 Specifies the minimum amount of indentation to use. The default value is
 zero. The value should be a non\-negative integer number.
 .TP
+.B \fBre2c:invert\-captures\fP
+Same as the \fB\-\-invert\-captures\fP option, but can be configured on per\-block
+basis.
+.TP
 .B \fBre2c:label:prefix\fP, \fBre2c:labelprefix\fP
 Specifies the prefix used for DFA state labels. The default is \fByy\fP\&.
 .TP
@@ -1387,6 +1399,10 @@ Specifies the name of the optional label that follows \fBYYGETSTATE\fP switch
 in storable state mode (enabled with \fBre2c:state:nextlabel\fP). The default
 is \fByyNext\fP\&.
 .TP
+.B \fBre2c:leftmost\-captures\fP
+Same as the \fB\-\-leftmost\-captures\fP option, but can be configured on
+per\-block basis.
+.TP
 .B \fBre2c:lookahead\fP, \fBre2c:flags:lookahead\fP
 Deprecated (see the deprecated \fB\-\-no\-lookahead\fP option).
 .TP
@@ -1527,7 +1543,13 @@ re2c uses the following syntax for regular expressions:
 .IP \(bu 2
 \fBR{n,m}\fP repetition of \fBR\fP from \fBn\fP to \fBm\fP times
 .IP \(bu 2
-\fB(R)\fP just \fBR\fP; parentheses are used to override precedence or for POSIX\-style submatch
+\fB(R)\fP just \fBR\fP; parentheses are used to override precedence.
+If submatch extraction is enabled, \fB(R)\fP is a capturing or a
+non\-capturing group depending on \fB\-\-invert\-captures\fP option.
+.IP \(bu 2
+\fB(!R)\fP
+If submatch extraction is enabled, \fB(!R)\fP is a non\-capturing or a
+capturing group depending on \fB\-\-invert\-captures\fP option.
 .IP \(bu 2
 \fBR S\fP concatenation: \fBR\fP followed by \fBS\fP
 .IP \(bu 2
@@ -1535,17 +1557,23 @@ re2c uses the following syntax for regular expressions:
 .IP \(bu 2
 \fBR / S\fP lookahead: \fBR\fP followed by \fBS\fP, but \fBS\fP is not consumed
 .IP \(bu 2
-\fBname\fP the regular expression defined as \fBname\fP (or literal string \fB\(dqname\(dq\fP in Flex compatibility mode)
+\fBname\fP the regular expression defined as \fBname\fP (or literal string
+\fB\(dqname\(dq\fP in Flex compatibility mode)
 .IP \(bu 2
-\fB{name}\fP the regular expression defined as \fBname\fP in Flex compatibility mode
+\fB{name}\fP the regular expression defined as \fBname\fP in Flex
+compatibility mode
 .IP \(bu 2
-\fB@stag\fP an \fIs\-tag\fP: saves the last input position at which \fB@stag\fP matches in a variable named \fBstag\fP
+\fB@stag\fP an \fIs\-tag\fP: saves the last input position at which \fB@stag\fP
+matches in a variable named \fBstag\fP
 .IP \(bu 2
-\fB#mtag\fP an \fIm\-tag\fP: saves all input positions at which \fB#mtag\fP matches in a variable named \fBmtag\fP
+\fB#mtag\fP an \fIm\-tag\fP: saves all input positions at which \fB#mtag\fP matches
+in a variable named \fBmtag\fP
 .UNINDENT
 .sp
-Character classes and string literals may contain the following escape sequences:
-\fB\ea\fP, \fB\eb\fP, \fB\ef\fP, \fB\en\fP, \fB\er\fP, \fB\et\fP, \fB\ev\fP, \fB\e\e\fP, octal escapes \fB\eooo\fP and hexadecimal escapes \fB\exhh\fP, \fB\euhhhh\fP and \fB\eUhhhhhhhh\fP\&.
+Character classes and string literals may contain the following escape
+sequences: \fB\ea\fP, \fB\eb\fP, \fB\ef\fP, \fB\en\fP, \fB\er\fP, \fB\et\fP, \fB\ev\fP, \fB\e\e\fP,
+octal escapes \fB\eooo\fP and hexadecimal escapes \fB\exhh\fP, \fB\euhhhh\fP and
+\fB\eUhhhhhhhh\fP\&.
 .SH HANDLING THE END OF INPUT
 .sp
 One of the main problems for the lexer is to know when to stop.
@@ -1,4 +1,4 @@
-/* Generated by re2c 3.0 */
+/* Generated by re2c 3.1 */
 #line 1 "../lib/lex.re"
 #include <stdint.h>
 #include <stdio.h>
@@ -1144,13 +1144,13 @@ yyreduce:

  case 13: /* primary: '(' ')'  */
 #line 77 "../lib/parse.ypp"
-               { (yyval.regexp) = ast.cap(ast.nil(NOWHERE)); }
+               { (yyval.regexp) = ast.cap(ast.nil(NOWHERE), CAPTURE); }
 #line 1149 "lib/parse.cc"
    break;

  case 14: /* primary: '(' expr ')'  */
 #line 78 "../lib/parse.ypp"
-               { (yyval.regexp) = ast.cap((yyvsp[-1].regexp)); }
+               { (yyval.regexp) = ast.cap((yyvsp[-1].regexp), CAPTURE); }
 #line 1155 "lib/parse.cc"
    break;

@@ -51,7 +51,7 @@ extern int yydebug;
 #include "src/parse/ast.h"
 namespace re2c {
    struct AstNode;
-};
+}

 #line 57 "lib/parse.h"

@@ -1,4 +1,4 @@
-/* Generated by re2c 3.0 */
+/* Generated by re2c 3.1 */
 #line 1 "../lib/test_helper.re"
 #include "lib/test_helper.h"

@@ -116,11 +116,21 @@ const char* help =
 "        units is a sequence of standalone 1-byte characters. With utf8 re2c\n"
 "        handles input as UTF8-encoded and recognizes multibyte characters.\n"
 "\n"
+"    --invert-captures\n"
+"\n"
+"        Invert the meaning of capturing and non-capturing groups. By default\n"
+"        (...) is capturing and (! ...) is non-capturing. With this option (!\n"
+"        ...) is capturing and (...) is non-capturing.\n"
+"\n"
 "    --lang <c | go | rust>\n"
 "\n"
 "        Specify the output language. Supported languages are C, Go and Rust.\n"
 "        The default is C for re2c, Go for re2go and Rust for re2rust.\n"
 "\n"
+"    --leftmost-captures\n"
+"\n"
+"        Enable submatch extraction with leftmost greedy capturing groups.\n"
+"\n"
 "    --location-format <gnu | msvc>\n"
 "\n"
 "        Specify location format in messages. With gnu locations are printed as\n"
@@ -1,4 +1,4 @@
-/* Generated by re2c 3.0 */
+/* Generated by re2c 3.1 */
 #line 1 "../src/msg/ver_to_vernum.re"
 #include <stddef.h>
 #include <stdint.h>
@@ -1,4 +1,4 @@
-/* Generated by re2c 3.0 */
+/* Generated by re2c 3.1 */

 #ifndef _RE2C_PARSE_LEX_
 #define _RE2C_PARSE_LEX_
@@ -144,21 +144,22 @@ enum yysymbol_kind_t
  YYSYMBOL_26_ = 26,                       /* '?'  */
  YYSYMBOL_27_ = 27,                       /* '('  */
  YYSYMBOL_28_ = 28,                       /* ')'  */
-  YYSYMBOL_YYACCEPT = 29,                  /* $accept  */
-  YYSYMBOL_spec = 30,                      /* spec  */
-  YYSYMBOL_def = 31,                       /* def  */
-  YYSYMBOL_name = 32,                      /* name  */
-  YYSYMBOL_enddef = 33,                    /* enddef  */
-  YYSYMBOL_rule = 34,                      /* rule  */
-  YYSYMBOL_ccode = 35,                     /* ccode  */
-  YYSYMBOL_trailexpr = 36,                 /* trailexpr  */
-  YYSYMBOL_expr = 37,                      /* expr  */
-  YYSYMBOL_diff = 38,                      /* diff  */
-  YYSYMBOL_term = 39,                      /* term  */
-  YYSYMBOL_factor = 40,                    /* factor  */
-  YYSYMBOL_closes = 41,                    /* closes  */
-  YYSYMBOL_close = 42,                     /* close  */
-  YYSYMBOL_primary = 43                    /* primary  */
+  YYSYMBOL_29_ = 29,                       /* '!'  */
+  YYSYMBOL_YYACCEPT = 30,                  /* $accept  */
+  YYSYMBOL_spec = 31,                      /* spec  */
+  YYSYMBOL_def = 32,                       /* def  */
+  YYSYMBOL_name = 33,                      /* name  */
+  YYSYMBOL_enddef = 34,                    /* enddef  */
+  YYSYMBOL_rule = 35,                      /* rule  */
+  YYSYMBOL_ccode = 36,                     /* ccode  */
+  YYSYMBOL_trailexpr = 37,                 /* trailexpr  */
+  YYSYMBOL_expr = 38,                      /* expr  */
+  YYSYMBOL_diff = 39,                      /* diff  */
+  YYSYMBOL_term = 40,                      /* term  */
+  YYSYMBOL_factor = 41,                    /* factor  */
+  YYSYMBOL_closes = 42,                    /* closes  */
+  YYSYMBOL_close = 43,                     /* close  */
+  YYSYMBOL_primary = 44                    /* primary  */
 };
 typedef enum yysymbol_kind_t yysymbol_kind_t;

@@ -486,10 +487,10 @@ union yyalloc
 /* YYFINAL -- State number of the termination state.  */
 #define YYFINAL  2
 /* YYLAST -- Last index in YYTABLE.  */
-#define YYLAST   64
+#define YYLAST   67

 /* YYNTOKENS -- Number of terminals.  */
-#define YYNTOKENS  29
+#define YYNTOKENS  30
 /* YYNNTS -- Number of nonterminals.  */
 #define YYNNTS  15
 /* YYNRULES -- Number of rules.  */
@@ -515,7 +516,7 @@ static const yytype_int8 yytranslate[] =
       0,     2,     2,     2,     2,     2,     2,     2,     2,     2,
       2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
       2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
-       2,     2,     2,     2,     2,     2,    22,     2,     2,     2,
+       2,     2,     2,    29,     2,     2,    22,     2,     2,     2,
      27,    28,    21,    25,     2,     2,     2,    18,     2,     2,
       2,     2,     2,     2,     2,     2,     2,     2,     2,    20,
       2,    19,     2,    26,     2,     2,     2,     2,     2,     2,
@@ -548,9 +549,9 @@ static const yytype_uint8 yyrline[] =
 {
       0,    76,    76,    77,    81,    85,    86,    87,    91,    94,
     100,   101,   104,   104,   107,   110,   113,   116,   122,   128,
-     134,   140,   147,   148,   153,   158,   159,   164,   165,   169,
-     170,   174,   176,   180,   181,   194,   200,   201,   205,   206,
-     207,   211,   212,   221,   222
+     134,   140,   147,   148,   153,   158,   159,   165,   166,   170,
+     171,   175,   177,   181,   182,   195,   201,   202,   206,   207,
+     208,   212,   213,   222,   223
 };
 #endif

@@ -571,8 +572,8 @@ static const char *const yytname[] =
  "TOKEN_CLOSESIZE", "TOKEN_CODE", "TOKEN_CONF", "TOKEN_ID", "TOKEN_FID",
  "TOKEN_FID_END", "TOKEN_LINE_INFO", "TOKEN_REGEXP", "TOKEN_BLOCK",
  "TOKEN_ERROR", "'/'", "'='", "';'", "'*'", "'$'", "'|'", "'\\\\'", "'+'",
-  "'?'", "'('", "')'", "$accept", "spec", "def", "name", "enddef", "rule",
-  "ccode", "trailexpr", "expr", "diff", "term", "factor", "closes",
+  "'?'", "'('", "')'", "'!'", "$accept", "spec", "def", "name", "enddef",
+  "rule", "ccode", "trailexpr", "expr", "diff", "term", "factor", "closes",
  "close", "primary", YY_NULLPTR
 };

@@ -597,13 +598,13 @@ yysymbol_name (yysymbol_kind_t yysymbol)
   STATE-NUM.  */
 static const yytype_int8 yypact[] =
 {
-     -20,     0,   -20,    14,    -1,    36,   -20,   -18,   -20,   -20,
-     -20,   -20,    11,    45,    22,   -20,    23,   -20,    52,    24,
-     -15,   -20,    23,     5,   -20,    36,    36,    36,   -20,   -20,
-      53,   -20,   -20,   -20,   -20,   -20,    23,    29,    33,   -20,
-      23,    23,    23,   -20,   -20,   -20,   -20,   -20,    34,   -20,
-     -20,   -20,   -20,   -20,    35,   -20,   -20,   -20,   -20,   -20,
-       9,   -15,   -20,   -20,   -20
+     -20,     0,   -20,    19,    -8,    46,   -20,   -11,   -20,   -20,
+     -20,   -20,     9,    17,     2,   -20,    24,   -20,    33,    35,
+     -15,   -20,    24,    11,   -20,    46,    46,    46,   -20,   -20,
+      38,   -20,   -20,   -20,   -20,   -20,    24,    -3,    25,   -20,
+      24,    24,    24,   -20,   -20,   -20,   -20,   -20,    31,   -20,
+     -20,   -20,   -20,   -20,    36,   -20,   -20,   -20,   -20,   -20,
+      10,   -15,   -20,   -20,   -20
 };

 /* YYDEFACT[STATE-NUM] -- Default reduction number in state STATE-NUM.
@@ -623,8 +624,8 @@ static const yytype_int8 yydefact[] =
 /* YYPGOTO[NTERM-NUM].  */
 static const yytype_int8 yypgoto[] =
 {
-     -20,   -20,   -20,   -20,   -20,   -20,    -8,    40,   -12,     3,
-     -19,   -20,   -20,    16,   -20
+     -20,   -20,   -20,   -20,   -20,   -20,    40,    41,   -12,    13,
+     -19,   -20,   -20,    12,   -20
 };

 /* YYDEFGOTO[NTERM-NUM].  */
@@ -639,47 +640,47 @@ static const yytype_int8 yydefgoto[] =
   number is the opposite.  If YYTABLE_NINF, syntax error.  */
 static const yytype_int8 yytable[] =
 {
-       2,    33,    37,    43,    38,     3,     4,     5,    28,    42,
-       6,     7,     8,    44,     9,    10,    11,    50,    51,    52,
-      34,    12,    13,    62,    54,    24,    45,    14,    60,    10,
-      46,    47,    41,    24,    24,    25,    26,    10,    10,    29,
-      30,    14,    40,    27,    61,    31,    56,    41,    36,    14,
-      14,    57,    41,    58,    35,    45,    41,    55,    41,    46,
-      47,    39,    53,    64,    63
+       2,    28,    37,    43,    38,     3,     4,     5,    33,    42,
+       6,     7,     8,    24,     9,    10,    11,    10,    34,    44,
+      41,    12,    13,    62,    54,    55,    35,    14,    60,    14,
+      24,    36,    45,    41,    10,    24,    46,    47,    56,    10,
+      25,    26,    39,    57,    27,    58,    14,    53,    41,    29,
+      30,    14,    45,    40,    61,    31,    46,    47,    41,    41,
+      63,     0,     0,     0,    64,    50,    51,    52
 };

 static const yytype_int8 yycheck[] =
 {
-       0,    19,    14,    22,    16,     5,     6,     7,     9,    24,
-      10,    11,    12,     8,    14,    15,    16,    25,    26,    27,
-       9,    21,    22,    42,    36,    11,    21,    27,    40,    15,
-      25,    26,    23,    11,    11,    21,    22,    15,    15,     3,
-       4,    27,    18,     3,    41,     9,    13,    23,    26,    27,
-      27,    18,    23,    20,     9,    21,    23,    28,    23,    25,
-      26,     9,     9,    28,    48
+       0,     9,    14,    22,    16,     5,     6,     7,    19,    24,
+      10,    11,    12,    11,    14,    15,    16,    15,     9,     8,
+      23,    21,    22,    42,    36,    28,     9,    27,    40,    27,
+      11,    29,    21,    23,    15,    11,    25,    26,    13,    15,
+      21,    22,     9,    18,     3,    20,    27,     9,    23,     3,
+       4,    27,    21,    18,    41,     9,    25,    26,    23,    23,
+      48,    -1,    -1,    -1,    28,    25,    26,    27
 };

 /* YYSTOS[STATE-NUM] -- The symbol kind of the accessing symbol of
   state STATE-NUM.  */
 static const yytype_int8 yystos[] =
 {
-       0,    30,     0,     5,     6,     7,    10,    11,    12,    14,
-      15,    16,    21,    22,    27,    31,    32,    34,    36,    37,
-      38,    39,    40,    43,    11,    21,    22,    36,     9,     3,
-       4,     9,    35,    19,     9,     9,    26,    37,    37,     9,
-      18,    23,    24,    39,     8,    21,    25,    26,    41,    42,
-      35,    35,    35,     9,    37,    28,    13,    18,    20,    33,
-      37,    38,    39,    42,    28
+       0,    31,     0,     5,     6,     7,    10,    11,    12,    14,
+      15,    16,    21,    22,    27,    32,    33,    35,    37,    38,
+      39,    40,    41,    44,    11,    21,    22,    37,     9,     3,
+       4,     9,    36,    19,     9,     9,    29,    38,    38,     9,
+      18,    23,    24,    40,     8,    21,    25,    26,    42,    43,
+      36,    36,    36,     9,    38,    28,    13,    18,    20,    34,
+      38,    39,    40,    43,    28
 };

 /* YYR1[RULE-NUM] -- Symbol kind of the left-hand side of rule RULE-NUM.  */
 static const yytype_int8 yyr1[] =
 {
-       0,    29,    30,    30,    30,    30,    30,    30,    31,    31,
-      32,    32,    33,    33,    34,    34,    34,    34,    34,    34,
-      34,    34,    35,    35,    35,    36,    36,    37,    37,    38,
-      38,    39,    39,    40,    40,    40,    41,    41,    42,    42,
-      42,    43,    43,    43,    43
+       0,    30,    31,    31,    31,    31,    31,    31,    32,    32,
+      33,    33,    34,    34,    35,    35,    35,    35,    35,    35,
+      35,    35,    36,    36,    36,    37,    37,    38,    38,    39,
+      39,    40,    40,    41,    41,    41,    42,    42,    43,    43,
+      43,    44,    44,    44,    44
 };

 /* YYR2[RULE-NUM] -- Number of symbols on the right-hand side of rule RULE-NUM.  */
@@ -1172,7 +1173,7 @@ yyreduce:
    if (use_block(input, ast, opts, grams, ast.temp_blockname) != Ret::OK) YYABORT;
    ast.temp_blockname.clear();
 }
-#line 1176 "src/parse/parser.cc"
+#line 1177 "src/parse/parser.cc"
    break;

  case 4: /* spec: spec TOKEN_CONF  */
@@ -1181,7 +1182,7 @@ yyreduce:
    // Handled here, as the main lexer should not get access to opts under construction.
    if (input.lex_conf(opts) != Ret::OK) YYABORT;
 }
-#line 1185 "src/parse/parser.cc"
+#line 1186 "src/parse/parser.cc"
    break;

  case 8: /* def: name expr enddef  */
@@ -1189,7 +1190,7 @@ yyreduce:
                   {
    if (add_named_def(opts.symtab, (yyvsp[-2].cstr), (yyvsp[-1].regexp), input) != Ret::OK) YYABORT;
 }
-#line 1193 "src/parse/parser.cc"
+#line 1194 "src/parse/parser.cc"
    break;

  case 9: /* def: name expr '/'  */
@@ -1198,13 +1199,13 @@ yyreduce:
    input.error_at_tok("trailing contexts are not allowed in named definitions");
    YYABORT;
 }
-#line 1202 "src/parse/parser.cc"
+#line 1203 "src/parse/parser.cc"
    break;

  case 10: /* name: TOKEN_ID '='  */
 #line 100 "../src/parse/parser.ypp"
               { (yyval.cstr) = (yyvsp[-1].cstr); }
-#line 1208 "src/parse/parser.cc"
+#line 1209 "src/parse/parser.cc"
    break;

  case 14: /* rule: trailexpr TOKEN_CODE  */
@@ -1212,7 +1213,7 @@ yyreduce:
                       {
    find_or_add_gram(grams, "").rules.push_back(AstRule((yyvsp[-1].regexp), (yyvsp[0].semact)));
 }
-#line 1216 "src/parse/parser.cc"
+#line 1217 "src/parse/parser.cc"
    break;

  case 15: /* rule: '*' TOKEN_CODE  */
@@ -1220,7 +1221,7 @@ yyreduce:
                 {
    find_or_add_gram(grams, "").defs.push_back((yyvsp[0].semact));
 }
-#line 1224 "src/parse/parser.cc"
+#line 1225 "src/parse/parser.cc"
    break;

  case 16: /* rule: '$' TOKEN_CODE  */
@@ -1228,7 +1229,7 @@ yyreduce:
                 {
    find_or_add_gram(grams, "").eofs.push_back((yyvsp[0].semact));
 }
-#line 1232 "src/parse/parser.cc"
+#line 1233 "src/parse/parser.cc"
    break;

  case 17: /* rule: TOKEN_CLIST trailexpr ccode  */
@@ -1239,7 +1240,7 @@ yyreduce:
    }
    ast.temp_condlist.clear();
 }
-#line 1243 "src/parse/parser.cc"
+#line 1244 "src/parse/parser.cc"
    break;

  case 18: /* rule: TOKEN_CLIST '*' ccode  */
@@ -1250,7 +1251,7 @@ yyreduce:
    }
    ast.temp_condlist.clear();
 }
-#line 1254 "src/parse/parser.cc"
+#line 1255 "src/parse/parser.cc"
    break;

  case 19: /* rule: TOKEN_CLIST '$' ccode  */
@@ -1261,7 +1262,7 @@ yyreduce:
    }
    ast.temp_condlist.clear();
 }
-#line 1265 "src/parse/parser.cc"
+#line 1266 "src/parse/parser.cc"
    break;

  case 20: /* rule: TOKEN_CSETUP TOKEN_CODE  */
@@ -1272,7 +1273,7 @@ yyreduce:
    }
    ast.temp_condlist.clear();
 }
-#line 1276 "src/parse/parser.cc"
+#line 1277 "src/parse/parser.cc"
    break;

  case 21: /* rule: TOKEN_CZERO ccode  */
@@ -1282,7 +1283,7 @@ yyreduce:
    find_or_add_gram(grams, "0").rules.push_back(AstRule(r, (yyvsp[0].semact)));
    ast.temp_condlist.clear();
 }
-#line 1286 "src/parse/parser.cc"
+#line 1287 "src/parse/parser.cc"
    break;

  case 23: /* ccode: TOKEN_CNEXT TOKEN_CODE  */
@@ -1292,7 +1293,7 @@ yyreduce:
    // parser needs to to update condition, but for the rest of the code actions are immutable
    const_cast<SemAct*>((yyval.semact))->cond = (yyvsp[-1].cstr); 
 }
-#line 1296 "src/parse/parser.cc"
+#line 1297 "src/parse/parser.cc"
    break;

  case 24: /* ccode: TOKEN_CJUMP  */
@@ -1300,43 +1301,44 @@ yyreduce:
              {
    (yyval.semact) = ast.sem_act(input.tok_loc(), nullptr, (yyvsp[0].cstr), true);
 }
-#line 1304 "src/parse/parser.cc"
+#line 1305 "src/parse/parser.cc"
    break;

  case 25: /* trailexpr: expr  */
 #line 158 "../src/parse/parser.ypp"
-       { (yyval.regexp) = ast.cap((yyvsp[0].regexp)); }
-#line 1310 "src/parse/parser.cc"
+       { (yyval.regexp) = ast.cap((yyvsp[0].regexp), CAPTURE_IF_NOT_INVERTED); }
+#line 1311 "src/parse/parser.cc"
    break;

  case 26: /* trailexpr: expr '/' expr  */
 #line 159 "../src/parse/parser.ypp"
                {
-    (yyval.regexp) = ast.cat(ast.cap((yyvsp[-2].regexp)), ast.cat(ast.tag(input.tok_loc(), nullptr, false), (yyvsp[0].regexp)));
+    (yyval.regexp) = ast.cat(ast.cap((yyvsp[-2].regexp), CAPTURE_IF_NOT_INVERTED),
+                 ast.cat(ast.tag(input.tok_loc(), nullptr, false), (yyvsp[0].regexp)));
 }
-#line 1318 "src/parse/parser.cc"
+#line 1320 "src/parse/parser.cc"
    break;

  case 28: /* expr: expr '|' diff  */
-#line 165 "../src/parse/parser.ypp"
+#line 166 "../src/parse/parser.ypp"
                { (yyval.regexp) = ast.alt((yyvsp[-2].regexp), (yyvsp[0].regexp)); }
-#line 1324 "src/parse/parser.cc"
+#line 1326 "src/parse/parser.cc"
    break;

  case 30: /* diff: diff '\\' term  */
-#line 170 "../src/parse/parser.ypp"
+#line 171 "../src/parse/parser.ypp"
                 { (yyval.regexp) = ast.diff((yyvsp[-2].regexp), (yyvsp[0].regexp)); }
-#line 1330 "src/parse/parser.cc"
+#line 1332 "src/parse/parser.cc"
    break;

  case 32: /* term: factor term  */
-#line 176 "../src/parse/parser.ypp"
+#line 177 "../src/parse/parser.ypp"
              { (yyval.regexp) = ast.cat((yyvsp[-1].regexp), (yyvsp[0].regexp)); }
-#line 1336 "src/parse/parser.cc"
+#line 1338 "src/parse/parser.cc"
    break;

  case 34: /* factor: primary closes  */
-#line 181 "../src/parse/parser.ypp"
+#line 182 "../src/parse/parser.ypp"
                 {
    switch((yyvsp[0].op)) {
    case '*':
@@ -1350,69 +1352,69 @@ yyreduce:
        break;
    }
 }
-#line 1354 "src/parse/parser.cc"
+#line 1356 "src/parse/parser.cc"
    break;

  case 35: /* factor: primary TOKEN_CLOSESIZE  */
-#line 194 "../src/parse/parser.ypp"
+#line 195 "../src/parse/parser.ypp"
                          {
    (yyval.regexp) = ast.iter((yyvsp[-1].regexp), (yyvsp[0].bounds).min, (yyvsp[0].bounds).max);
 }
-#line 1362 "src/parse/parser.cc"
+#line 1364 "src/parse/parser.cc"
    break;

  case 37: /* closes: closes close  */
-#line 201 "../src/parse/parser.ypp"
+#line 202 "../src/parse/parser.ypp"
               { (yyval.op) = ((yyvsp[-1].op) == (yyvsp[0].op)) ? (yyvsp[-1].op) : '*'; }
-#line 1368 "src/parse/parser.cc"
+#line 1370 "src/parse/parser.cc"
    break;

  case 38: /* close: '*'  */
-#line 205 "../src/parse/parser.ypp"
+#line 206 "../src/parse/parser.ypp"
      { (yyval.op) = '*'; }
-#line 1374 "src/parse/parser.cc"
+#line 1376 "src/parse/parser.cc"
    break;

  case 39: /* close: '+'  */
-#line 206 "../src/parse/parser.ypp"
+#line 207 "../src/parse/parser.ypp"
      { (yyval.op) = '+'; }
-#line 1380 "src/parse/parser.cc"
+#line 1382 "src/parse/parser.cc"
    break;

  case 40: /* close: '?'  */
-#line 207 "../src/parse/parser.ypp"
+#line 208 "../src/parse/parser.ypp"
      { (yyval.op) = '?'; }
-#line 1386 "src/parse/parser.cc"
+#line 1388 "src/parse/parser.cc"
    break;

  case 42: /* primary: TOKEN_ID  */
-#line 212 "../src/parse/parser.ypp"
+#line 213 "../src/parse/parser.ypp"
           {
    (yyval.regexp) = find_def(opts.symtab, (yyvsp[0].cstr));
    if ((yyval.regexp) == nullptr) {
        input.error_at_tok("undefined symbol '%s'", (yyvsp[0].cstr));
        YYABORT;
    } else if (Ast::needs_wrap((yyval.regexp))) {
-        (yyval.regexp) = ast.ref((yyval.regexp));
+        (yyval.regexp) = ast.cap((yyval.regexp), NO_CAPTURE);
    }
 }
-#line 1400 "src/parse/parser.cc"
+#line 1402 "src/parse/parser.cc"
    break;

  case 43: /* primary: '(' expr ')'  */
-#line 221 "../src/parse/parser.ypp"
-                   { (yyval.regexp) = ast.cap((yyvsp[-1].regexp)); }
-#line 1406 "src/parse/parser.cc"
-    break;
-
-  case 44: /* primary: '(' '?' expr ')'  */
 #line 222 "../src/parse/parser.ypp"
-                   { (yyval.regexp) = ast.ref((yyvsp[-1].regexp)); }
-#line 1412 "src/parse/parser.cc"
+                   { (yyval.regexp) = ast.cap((yyvsp[-1].regexp), CAPTURE_IF_NOT_INVERTED); }
+#line 1408 "src/parse/parser.cc"
+    break;
+
+  case 44: /* primary: '(' '!' expr ')'  */
+#line 223 "../src/parse/parser.ypp"
+                   { (yyval.regexp) = ast.cap((yyvsp[-1].regexp), CAPTURE_IF_INVERTED); }
+#line 1414 "src/parse/parser.cc"
    break;


-#line 1416 "src/parse/parser.cc"
+#line 1418 "src/parse/parser.cc"

      default: break;
    }
@@ -1605,7 +1607,7 @@ yyreturnlab:
  return yyresult;
 }

-#line 225 "../src/parse/parser.ypp"
+#line 226 "../src/parse/parser.ypp"


 #pragma GCC diagnostic pop
@@ -52,7 +52,7 @@ extern int yydebug;
 namespace re2c {
    struct AstNode;
    struct SemAct;
-};
+}

 #line 58 "src/parse/parser.h"

@@ -12,5 +12,6 @@ cd $builddir
    --enable-debug \
    --enable-libs \
    --host i686-w64-mingw32 \
-    && make -j$(nproc)
+    && make -j$(nproc) \
+    && ../build/copy_wine_libs_on_nixos.sh
 cd ..
@@ -16,5 +16,6 @@ cd $builddir
    --enable-debug \
    --enable-libs \
    --host i686-w64-mingw32 \
-    && make -j$(nproc) LIBTOOL="slibtool"
+    && make -j$(nproc) LIBTOOL="slibtool" \
+    && ../build/copy_wine_libs_on_nixos.sh
 cd ..
@@ -12,5 +12,6 @@ cmake .. \
    -DBUILD_SHARED_LIBS=no \
    -DCMAKE_EXE_LINKER_FLAGS="-static -static-libstdc++ -static-libgcc" \
    -DCMAKE_TOOLCHAIN_FILE=cmake/Toolchain-cross-mingw32-linux.cmake \
-    && cmake --build . -j$(nproc)
+    && cmake --build . -j$(nproc) \
+    && ../build/copy_wine_libs_on_nixos.sh
 cd ..
@@ -0,0 +1,19 @@
+#!/bin/sh
+
+# On NixOS, Wine cannot find DLLs by their absolute paths in /nix/store,
+# so find and copy them in the same directory with re2c.exe.
+
+uname -a | grep -q NixOS || exit 0
+
+test -e re2c.exe || \
+    { echo "$0 must be ran from the build directory containing re2c.exe"; exit 1; }
+
+nix_wine="$(nix-build --no-link '<nixpkgs>' -A wine)"
+nix_mcfgthreads="$(nix-build --no-link '<nixpkgs>' -A pkgsCross.mingw32.windows.mcfgthreads)"
+
+uname -a | grep -q NixOS && (
+    cp -f "$nix_mcfgthreads/bin/mcfgthread"*".dll" .
+    cp -f "$nix_wine/lib/wine/i386-windows/win32u.dll" .
+    cp -f "$nix_wine/lib/wine/i386-windows/user32.dll" .
+    cp -f "$nix_wine/lib/wine/i386-windows/gdi32.dll" .
+)
@@ -1,4 +1,4 @@
-AC_INIT([re2c],[3.0],[re2c-general@lists.sourceforge.net])
+AC_INIT([re2c],[3.1],[re2c-general@lists.sourceforge.net])
 AC_CONFIG_AUX_DIR([build-aux])
 AM_INIT_AUTOMAKE([foreign subdir-objects dist-xz dist-lzip no-dist-gzip])
 AM_SILENT_RULES([yes])
@@ -15,7 +15,8 @@ AC_SUBST(PACKAGE_RELEASE, ${PACKAGE_RELEASE:-1})

 AM_PATH_PYTHON([3.7])

-# without this, --std=c++98 disables POSIX functions on Cygwin
+# without this, --std=c++11 disables POSIX functions on Cygwin and NetBSD
+# (alternatively we could use -std=gnu++11, as CMake does)
 AC_USE_SYSTEM_EXTENSIONS


@@ -250,6 +250,10 @@
    Specifies the minimum amount of indentation to use. The default value is
    zero. The value should be a non-negative integer number.

+``re2c:invert-captures``
+    Same as the ``--invert-captures`` option, but can be configured on per-block
+    basis.
+
 ``re2c:label:prefix``, ``re2c:labelprefix``
    Specifies the prefix used for DFA state labels. The default is ``yy``.

@@ -274,6 +278,10 @@
    in storable state mode (enabled with ``re2c:state:nextlabel``). The default
    is ``yyNext``.

+``re2c:leftmost-captures``
+    Same as the ``--leftmost-captures`` option, but can be configured on
+    per-block basis.
+
 ``re2c:lookahead``, ``re2c:flags:lookahead``
    Deprecated (see the deprecated ``--no-lookahead`` option).

@@ -89,10 +89,18 @@
    With ``utf8`` re2c handles input as UTF8-encoded and recognizes multibyte
    characters.

+``--invert-captures``
+    Invert the meaning of capturing and non-capturing groups. By default
+    ``(...)`` is capturing and ``(! ...)`` is non-capturing. With this option
+    ``(! ...)`` is capturing and ``(...)`` is non-capturing.
+
 ``--lang <c | go | rust>``
    Specify the output language. Supported languages are C, Go and Rust.
    The default is C for re2c, Go for re2go and Rust for re2rust.

+``--leftmost-captures``
+    Enable submatch extraction with leftmost greedy capturing groups.
+
 ``--location-format <gnu | msvc>``
    Specify location format in messages.
    With ``gnu`` locations are printed as 'filename:line:column: ...'.
@@ -23,7 +23,13 @@ re2c uses the following syntax for regular expressions:

 *    ``R{n,m}`` repetition of ``R`` from ``n`` to ``m`` times

-*    ``(R)`` just ``R``; parentheses are used to override precedence or for POSIX-style submatch
+*    ``(R)`` just ``R``; parentheses are used to override precedence.
+     If submatch extraction is enabled, ``(R)`` is a capturing or a
+     non-capturing group depending on ``--invert-captures`` option.
+
+*    ``(!R)``
+     If submatch extraction is enabled, ``(!R)`` is a non-capturing or a
+     capturing group depending on ``--invert-captures`` option.

 *    ``R S`` concatenation: ``R`` followed by ``S``

@@ -31,14 +37,20 @@ re2c uses the following syntax for regular expressions:

 *    ``R / S`` lookahead: ``R`` followed by ``S``, but ``S`` is not consumed

-*    ``name`` the regular expression defined as ``name`` (or literal string ``"name"`` in Flex compatibility mode)
+*    ``name`` the regular expression defined as ``name`` (or literal string
+     ``"name"`` in Flex compatibility mode)

-*    ``{name}`` the regular expression defined as ``name`` in Flex compatibility mode
+*    ``{name}`` the regular expression defined as ``name`` in Flex
+     compatibility mode

-*    ``@stag`` an *s-tag*: saves the last input position at which ``@stag`` matches in a variable named ``stag``
+*    ``@stag`` an *s-tag*: saves the last input position at which ``@stag``
+     matches in a variable named ``stag``

-*    ``#mtag`` an *m-tag*: saves all input positions at which ``#mtag`` matches in a variable named ``mtag``
+*    ``#mtag`` an *m-tag*: saves all input positions at which ``#mtag`` matches
+     in a variable named ``mtag``

-Character classes and string literals may contain the following escape sequences:
-``\a``, ``\b``, ``\f``, ``\n``, ``\r``, ``\t``, ``\v``, ``\\``, octal escapes ``\ooo`` and hexadecimal escapes ``\xhh``, ``\uhhhh`` and ``\Uhhhhhhhh``.
+Character classes and string literals may contain the following escape
+sequences: ``\a``, ``\b``, ``\f``, ``\n``, ``\r``, ``\t``, ``\v``, ``\\``,
+octal escapes ``\ooo`` and hexadecimal escapes ``\xhh``, ``\uhhhh`` and
+``\Uhhhhhhhh``.

@@ -3,7 +3,7 @@
 #include "src/parse/ast.h"
 namespace re2c {
    struct AstNode;
-};
+}
 }

 %{
@@ -74,8 +74,8 @@ factor

 primary
 : TOKEN_REGEXP
-| '(' ')'      { $$ = ast.cap(ast.nil(NOWHERE)); }
-| '(' expr ')' { $$ = ast.cap($2); }
+| '(' ')'      { $$ = ast.cap(ast.nil(NOWHERE), CAPTURE); }
+| '(' expr ')' { $$ = ast.cap($2, CAPTURE); }
 ;

 %%
@@ -66,7 +66,7 @@ void make_one_step(psimctx_t& ctx, uint32_t sym) {

        s->clos = NOCLOS;
        s->arcidx = 0;
-        DCHECK(s->status == GorPass::NOPASS && s->active == 0);
+        DCHECK(s->status == GOR_NOPASS && s->active == 0);

        if (s->kind == TnfaState::Kind::RAN) {
            for (const Range* r = s->ran; r; r = r->next()) {
@@ -105,7 +105,7 @@ void make_final_step(psimctx_t& ctx) {

        s->clos = NOCLOS;
        s->arcidx = 0;
-        DCHECK(s->status == GorPass::NOPASS && s->active == 0);
+        DCHECK(s->status == GOR_NOPASS && s->active == 0);

        if (s->kind == TnfaState::Kind::FIN) {
            update_offsets(ctx, *i, NONCORE);
@@ -87,7 +87,7 @@ void make_step(pzsimctx_t& ctx, uint32_t sym) {

        s->clos = NOCLOS;
        s->arcidx = 0;
-        DCHECK(s->status == GorPass::NOPASS && s->active == 0);
+        DCHECK(s->status == GOR_NOPASS && s->active == 0);

        if (s->kind == TnfaState::Kind::RAN) {
            for (const Range* r = s->ran; r; r = r->next()) {
@@ -113,7 +113,7 @@ void make_final_step(pzsimctx_t& ctx) {

        s->clos = NOCLOS;
        s->arcidx = 0;
-        DCHECK(s->status == GorPass::NOPASS && s->active == 0);
+        DCHECK(s->status == GOR_NOPASS && s->active == 0);

        if (s->kind == TnfaState::Kind::FIN) {
            ctx.marker = ctx.cursor;
@@ -1,10 +1,10 @@
-#!/bin/bash
+#!/usr/bin/env bash

 set -e

 # checks for which there are no scripts yet
 echo 'RUNME: cppcheck --enable=all --inconclusive --std=posix --quiet --force -I. src/'
-echo "RUNME: uselex `find <objdir> -type f -name '*.o'`"
+echo 'RUNME: uselex `find <objdir> -type f -name '"'*.o'"'`'
 echo 'RUNME: configure CFLAGS="-ffunction-sections -fdata-sections" LDFLAGS="-Wl,--gc-sections -Wl,--print-gc-sections"'

 if [[ $# -ne 2 ]]; then
@@ -31,13 +31,9 @@ TOP_BUILDIR = '@top_builddir@'
 RE2C_PATTERN = re.compile(r'.*re2c[ \t]*(.*)$')
 TYPE_HEADER_PATTERN1 = re.compile(r'--header[ \t]*')
 TYPE_HEADER_PATTERN2 = re.compile(r'.*--header[ \t]*([^ \t]*).*')
-START_OF_MESSAGE = b'^[a-z0-9_]+([\\\\/]+[a-z0-9_]+)+([.][a-z]+)+[(:]'
+FILEPATH = b'^[a-z0-9_]+([\\\\/]+[a-z0-9_]+)+([.][a-z]+)+'
 START_OF_LINEDIR = b'^(//|#)line '
-FILEPATH_CONTEXT_PATTERN = re.compile(b'('
-    + START_OF_MESSAGE
-    + b'|'
-    + START_OF_LINEDIR
-    + b')')
+FILEPATH_CONTEXT_PATTERN = re.compile(b'(' + START_OF_LINEDIR + b'|' + FILEPATH + b')')


@dataclass
@@ -361,12 +357,6 @@ def run_test(test):
    stem = os.path.splitext(outx)[0]
    outy = f'{stem}.{ext}'

-    # Known failure: on windows overflow-1.re fails due to small stack size.
-    if os.name == 'nt' and outx.endswith('overflow-1.re'):
-        if not _ctx.keep_temp_files:
-            remove(outx, outy)
-        return 0, 0, 0
-
    # Prepare re2c flags
    switches = head
    switches = switches.replace('re2go', 're2c --lang go')
@@ -4,6 +4,7 @@
 #include <stdint.h>
 #include <string.h>

+#include "src/constants.h"
 #include "src/dfa/tcmd.h"
 #include "src/msg/location.h"
 #include "src/util/allocator.h"
@@ -746,7 +746,7 @@ static void render(RenderContext& rctx, const Code* code) {
    }
 }

-uint32_t write_converting_newlines(const std::string& str, FILE* f) {
+static uint32_t write_converting_newlines(const std::string& str, FILE* f) {
    const char* s = str.c_str(), *e = s + str.length();
    uint32_t lines = 0;

@@ -53,6 +53,15 @@ enum class Minimization: uint32_t {
    MOORE
 };

+// Whether a group is capturing or non-capturng depending on --invert-capture option and
+// `re2c:invert_captures` configuration.
+enum CaptureMode : uint8_t {
+    NO_CAPTURE              = 0, // 00
+    CAPTURE_IF_INVERTED     = 1, // 01
+    CAPTURE_IF_NOT_INVERTED = 2, // 10
+    CAPTURE                 = CAPTURE_IF_INVERTED | CAPTURE_IF_NOT_INVERTED
+};
+
 // Algorithm for construction of POSIX precedence table for disambiguation.
 // Complexity estimates are in terms of TNFA size (m) and the number of tags (t).
 enum class PosixPrectable: uint32_t {
@@ -65,13 +65,13 @@ void closure_posix_gor1(ctx_t& ctx) {
        // 1st pass: depth-first postorder traversal of admissible subgraph
        for (; !topsort.empty(); ) {
            TnfaState* q = topsort.back();
-            if (q->status == GorPass::LINEAR) {
+            if (q->status == GOR_LINEAR) {
                topsort.pop_back();
            } else {
-                q->status = GorPass::TOPSORT;
+                q->status = GOR_TOPSORT;
                DINCCOUNT_CLSCANS(ctx);
                if (!scan(ctx, q, false)) {
-                    q->status = GorPass::LINEAR;
+                    q->status = GOR_LINEAR;
                    topsort.pop_back();
                    linear.push_back(q);
                }
@@ -88,7 +88,7 @@ void closure_posix_gor1(ctx_t& ctx) {
                DINCCOUNT_CLSCANS(ctx);
                scan(ctx, q, true);
            }
-            q->status = GorPass::NOPASS;
+            q->status = GOR_NOPASS;
        }
    }
 }
@@ -154,7 +154,7 @@ bool relax_gor1(ctx_t& ctx, const typename ctx_t::conf_t& x) {
    const uint32_t idx = q->clos;
    int32_t p1, p2;

-    if (q->status == GorPass::TOPSORT) {
+    if (q->status == GOR_TOPSORT) {
        return false;
    }

@@ -167,7 +167,7 @@ bool relax_gor1(ctx_t& ctx, const typename ctx_t::conf_t& x) {
        return false;
    }

-    if (q->status == GorPass::NOPASS) {
+    if (q->status == GOR_NOPASS) {
        ctx.gor1_topsort.push_back(q);
        q->arcidx = 0;
        return true;
@@ -181,7 +181,7 @@ template<>
 inline void closure_cleanup<pdetctx_t>(TnfaState* q) {
    q->clos = NOCLOS;
    q->arcidx = 0;
-    DCHECK(q->status == GorPass::NOPASS && q->active == 0);
+    DCHECK(q->status == GOR_NOPASS && q->active == 0);
 }

 } // namespace re2c
@@ -223,7 +223,7 @@ LOCAL_NODISCARD(Ret compile(int, char* argv[])) {

    out_alc.clear(); // Release memory used for codegen.

-    CHECK_RET(input.gen_dep_file());
+    CHECK_RET(input.gen_dep_file(output.total_opts->header_file));

    if (globopts.verbose) fprintf(stderr, "re2c: success\n");
    return Ret::OK;
@@ -15,11 +15,13 @@ namespace re2c {
 struct clos_t;

 // Passes of Goldberg-Radzik shortest path algorithm.
-enum class GorPass: uint32_t {
-    NOPASS = 0u,
-    TOPSORT = 1u,
-    LINEAR = 2u
-};
+//
+// This cannot be a `enum class` because it is used in a bitfield, which triggers compiler
+// warnings and errors on some toolchains, see https://github.com/skvadrik/re2c/issues/446.
+// It cannot be a plain `enum` defaulting to `int` as well because of -Wsigned-enum-bitfield.
+constexpr uint32_t GOR_NOPASS = 0u;
+constexpr uint32_t GOR_TOPSORT = 1u;
+constexpr uint32_t GOR_LINEAR = 2u;

 static constexpr uint32_t NOCLOS = ~0u;

@@ -33,7 +35,7 @@ struct TnfaState {
        tag_info_t tag;   // tag information (number and sign, only for TAG states)
    };
    uint32_t clos;        // GOR1: closure item for this stack
-    GorPass status  : 2;  // GOR1: status (values 0, 1, 2)
+    uint32_t status : 2;  // GOR1: status (values 0, 1, 2)
    uint32_t arcidx : 2;  // GOR1: index of the next transtion (maximum out-dergee is 2)
    uint32_t active : 1;  // GOR1: if the state is on stack (boolean)
    uint32_t indeg  : 27; // GOR1: in-degree (we are unlikely to have more than 2^27 states)
@@ -85,7 +87,7 @@ struct Tnfa {
        s->kind = kind;
        s->rule = rule;
        s->clos = NOCLOS;
-        s->status = GorPass::NOPASS;
+        s->status = GOR_NOPASS;
        s->arcidx = 0;
        s->active = 0;
        s->indeg = 0;
@@ -167,6 +167,7 @@ class Input;
    MUTOPT(bool,        tags_posix_syntax, false) \
    MUTOPT(bool,        tags_posix_semantics, false) \
    MUTOPT(std::string, tags_prefix, "yyt") \
+    MUTOPT(bool,        invert_captures, false) \
    /* labels */ \
    MUTOPT(std::string, label_fill, "yyFillLabel") \
    MUTOPT(std::string, label_loop, "") \
@@ -160,6 +160,7 @@ opt_long: /*!local:re2c
    "case-inverted"         end { opts.set_case_inverted(true);      goto opt; }
    "tags"                  end { opts.set_tags(true);               goto opt; }
    "no-unsafe"             end { opts.set_unsafe(false);            goto opt; }
+    "invert-captures"       end { opts.set_invert_captures(true);    goto opt; }

    "ebcdic" | "ecb"        end { opts.set_encoding(Enc::Type::EBCDIC, true); goto opt; }
    "utf32"  | "unicode"    end { opts.set_encoding(Enc::Type::UTF32, true);  goto opt; }
@@ -58,7 +58,7 @@ Ast::Ast(AstAllocator& ast_alc, OutAllocator& out_alc)
      temp_blockname(),
      blocks() {}

-AstNode* Ast::make(const loc_t& loc, AstKind kind, bool has_caps) {
+AstNode* Ast::make(const loc_t& loc, AstKind kind, uint8_t has_caps) {
    AstNode* p = ast_alc.alloct<AstNode>(1);
    p->kind = kind;
    p->loc = loc;
@@ -67,11 +67,11 @@ AstNode* Ast::make(const loc_t& loc, AstKind kind, bool has_caps) {
 }

 const AstNode* Ast::nil(const loc_t& loc) {
-    return make(loc, AstKind::NIL, false);
+    return make(loc, AstKind::NIL, NO_CAPTURE);
 }

 const AstNode* Ast::str(const loc_t& loc, bool icase) {
-    AstNode* ast = make(loc, AstKind::STR, false);
+    AstNode* ast = make(loc, AstKind::STR, NO_CAPTURE);
    ast->str.chars.init(temp_chars.data(), temp_chars.size(), ast_alc);
    ast->str.icase = icase;
    temp_chars.clear();
@@ -79,7 +79,7 @@ const AstNode* Ast::str(const loc_t& loc, bool icase) {
 }

 const AstNode* Ast::cls(const loc_t& loc, bool negated) {
-    AstNode* ast = make(loc, AstKind::CLS, false);
+    AstNode* ast = make(loc, AstKind::CLS, NO_CAPTURE);
    ast->cls.ranges.init(temp_ranges.data(), temp_ranges.size(), ast_alc);
    ast->cls.negated = negated;
    temp_ranges.clear();
@@ -87,17 +87,17 @@ const AstNode* Ast::cls(const loc_t& loc, bool negated) {
 }

 const AstNode* Ast::dot(const loc_t& loc) {
-    return make(loc, AstKind::DOT, false);
+    return make(loc, AstKind::DOT, NO_CAPTURE);
 }

 const AstNode* Ast::def(const loc_t& loc) {
-    return make(loc, AstKind::DEF, false);
+    return make(loc, AstKind::DEF, NO_CAPTURE);
 }

 const AstNode* Ast::alt(const AstNode* a1, const AstNode* a2) {
    if (!a1) return a2;
    if (!a2) return a1;
-    AstNode* ast = make(a1->loc, AstKind::ALT, a1->has_caps || a2->has_caps);
+    AstNode* ast = make(a1->loc, AstKind::ALT, a1->has_caps | a2->has_caps);
    ast->alt.ast1 = a1;
    ast->alt.ast2 = a2;
    return ast;
@@ -106,7 +106,7 @@ const AstNode* Ast::alt(const AstNode* a1, const AstNode* a2) {
 const AstNode* Ast::cat(const AstNode* a1, const AstNode* a2) {
    if (!a1) return a2;
    if (!a2) return a1;
-    AstNode* ast = make(a1->loc, AstKind::CAT, a1->has_caps || a2->has_caps);
+    AstNode* ast = make(a1->loc, AstKind::CAT, a1->has_caps | a2->has_caps);
    ast->cat.ast1 = a1;
    ast->cat.ast2 = a2;
    return ast;
@@ -122,28 +122,23 @@ const AstNode* Ast::iter(const AstNode* a, uint32_t n, uint32_t m) {
 }

 const AstNode* Ast::diff(const AstNode* a1, const AstNode* a2) {
-    AstNode* ast = make(a1->loc, AstKind::DIFF, a1->has_caps || a2->has_caps);
+    AstNode* ast = make(a1->loc, AstKind::DIFF, a1->has_caps | a2->has_caps);
    ast->cat.ast1 = a1;
    ast->cat.ast2 = a2;
    return ast;
 }

 const AstNode* Ast::tag(const loc_t& loc, const char* n, bool h) {
-    AstNode* ast = make(loc, AstKind::TAG, false);
+    AstNode* ast = make(loc, AstKind::TAG, NO_CAPTURE);
    ast->tag.name = n;
    ast->tag.history = h;
    return ast;
 }

-const AstNode* Ast::cap(const AstNode* a) {
-    AstNode* ast = make(a->loc, AstKind::CAP, true);
-    ast->cap = a;
-    return ast;
-}
-
-const AstNode* Ast::ref(const AstNode* a) {
-    AstNode* ast = make(a->loc, AstKind::REF, a->has_caps);
-    ast->ref = a;
+const AstNode* Ast::cap(const AstNode* a, CaptureMode mode) {
+    AstNode* ast = make(a->loc, AstKind::CAP, mode | a->has_caps);
+    ast->cap.ast = a;
+    ast->cap.mode = mode;
    return ast;
 }

@@ -180,7 +175,6 @@ bool Ast::needs_wrap(const AstNode* a) {
    case AstKind::ALT:
    case AstKind::CAT:
    case AstKind::DIFF:
-    case AstKind::REF:
        return true;
    }
    return false; // unreachable
@@ -32,8 +32,7 @@ enum class AstKind: uint32_t {
    ITER, // generalized repetition of two nodes: x{n,m} or x{n,} or x{n} or x* or x+
    DIFF, // difference of two node (only applies to character classes)
    TAG,  // a tag, like @t (s-tag, single-valued tag) or #t (m-tag. multi-valued tag)
-    CAP,  // capturing group (submatch group)
-    REF   // non-capturing group
+    CAP   // capturing or non-capturing group
 };

 // A character (symbol) in the abstract syntax tree.
@@ -92,11 +91,16 @@ struct AstNode {
            const char* name;
            bool history;
        } tag;
-        const AstNode* cap;
-        const AstNode* ref;
+        struct {
+            const AstNode* ast;
+            CaptureMode mode;
+        } cap;
    };
    loc_t loc;
-    bool has_caps; // whether this AST has nested capturing groups
+
+    // Whether this AST has nested capturing groups (a combination of `CaptureMode` bit flags
+    // accumulated from the nested regexp).
+    uint8_t has_caps;
 };

 // Semanic action associated with an AST for a regular expression.
@@ -166,7 +170,7 @@ class Ast {
    // survive until the main codegen pass (e.g. semantic actions, condition names, tag names).
    OutAllocator& out_alc;

-    AstNode* make(const loc_t& loc, AstKind kind, bool has_caps);
+    AstNode* make(const loc_t& loc, AstKind kind, uint8_t has_caps);

  public:
    // Temporary buffers for constructing character strings and classes in lexer/parser. Generally
@@ -197,8 +201,7 @@ class Ast {
    const AstNode* iter(const AstNode* a, uint32_t n, uint32_t m);
    const AstNode* diff(const AstNode* a1, const AstNode* a2);
    const AstNode* tag(const loc_t& loc, const char* n, bool h);
-    const AstNode* cap(const AstNode* a);
-    const AstNode* ref(const AstNode* a);
+    const AstNode* cap(const AstNode* a, CaptureMode mode);
    const SemAct* sem_act(const loc_t& loc, const char* text, const char* cond, bool autogen);

    // Local and global strings differ in their lifetimes: local ones live during AST construction,
@@ -247,14 +247,18 @@ bool Input::fill(size_t need) {
    return true;
 }

-Ret Input::gen_dep_file() const {
+Ret Input::gen_dep_file(const std::string& header) const {
    const std::string& fname = globopts->dep_file;
    if (fname.empty()) return Ret::OK;

    FILE* file = fopen(fname.c_str(), "w");
    if (file == nullptr) RET_FAIL(error("cannot open dep file %s", fname.c_str()));

-    fprintf(file, "%s:", escape_backslashes(globopts->output_file).c_str());
+    fprintf(file, "%s", escape_backslashes(globopts->output_file).c_str());
+    if (!header.empty()) {
+        fprintf(file, " %s", header.c_str());
+    }
+    fprintf(file, ":");
    for (const std::string& fdep : filedeps) {
        fprintf(file, " %s", fdep.c_str());
    }
@@ -68,7 +68,7 @@ class Input: private LexerState {
    loc_t cur_loc() const;
    Ret open(const std::string& filename, const std::string* parent) NODISCARD;
    Ret include(const std::string& filename, uint8_t* at) NODISCARD;
-    Ret gen_dep_file() const NODISCARD;
+    Ret gen_dep_file(const std::string& header) const NODISCARD;
    Ret lex_program(Output& out, std::string& block_name, InputBlock& kind) NODISCARD;
    Ret lex_block(YYSTYPE* yylval, Ast& ast, int& token) NODISCARD;
    Ret lex_conf(Opt& opts) NODISCARD;
@@ -428,7 +428,7 @@ scan:
        RET_TOK(TOKEN_REGEXP);
    }

-    [*+?()|;/\\=$] { RET_TOK(*tok); }
+    [*+?!()|;/\\=$] { RET_TOK(*tok); }

    "{" [0-9]+ "}" {
        if (!s_to_u32_unsafe (tok + 1, cur - 1, yylval->bounds.min)) {
@@ -532,7 +532,19 @@ scan:
        goto scan;
    }

-    * { RET_FAIL(error_at_tok("unexpected character: '%c'", *tok)); }
+    * {
+        if (globopts->flex_syntax && globopts->input_encoding == Enc::Type::UTF8) {
+            // Try to lex this as a raw UTF-8 code point (not captured by the `name` rule above
+            // because it is restricted to alphanumeric ASCII characters and underscore).
+            --cur;
+            uint32_t c;
+            CHECK_RET(lex_cls_chr(c));
+            ast.temp_chars.push_back({c, tok_loc()});
+            yylval->regexp = ast.str(tok_loc(), false);
+            RET_TOK(TOKEN_REGEXP);
+        }
+        RET_FAIL(error_at_tok("unexpected character: '%c'", *tok));
+    }
 */
 }

@@ -80,7 +80,12 @@ Ret Input::lex_conf(Opt& opts) {
    "yyfill:parameter" { RET_CONF_BOOL(fill_param_enable); }
    "yyfill:check"     { RET_CONF_BOOL(fill_check); }

-    "flags:"? "tags"           | "flags:T" { RET_CONF_BOOL(tags); }
+    "flags:"? "tags" | "flags:T" { RET_CONF_BOOL(tags); }
+    "flags:"? "leftmost-captures" {
+        CHECK_RET(lex_conf_bool(b));
+        opts.set_tags_posix_syntax(b);
+        return Ret::OK;
+    }
    "flags:"? "posix-captures" | "flags:P" {
        CHECK_RET(lex_conf_bool(b));
        opts.set_tags_posix_syntax(b);
@@ -89,6 +94,7 @@ Ret Input::lex_conf(Opt& opts) {
    }
    "tags:prefix"     { RET_CONF_STR(tags_prefix); }
    "tags:expression" { RET_CONF_STR(tags_expression); }
+    "invert-captures" { RET_CONF_BOOL(invert_captures); }

    "define:YYBACKUP"             { RET_CONF_STR(api_backup); }
    "define:YYBACKUPCTX"          { RET_CONF_STR(api_backup_ctx); }
@@ -4,7 +4,7 @@
 namespace re2c {
    struct AstNode;
    struct SemAct;
-};
+}
 }

 %{
@@ -155,9 +155,10 @@ ccode
 };

 trailexpr
-: expr { $$ = ast.cap($1); }
+: expr { $$ = ast.cap($1, CAPTURE_IF_NOT_INVERTED); }
 | expr '/' expr {
-    $$ = ast.cat(ast.cap($1), ast.cat(ast.tag(input.tok_loc(), nullptr, false), $3));
+    $$ = ast.cat(ast.cap($1, CAPTURE_IF_NOT_INVERTED),
+                 ast.cat(ast.tag(input.tok_loc(), nullptr, false), $3));
 };

 expr
@@ -215,11 +216,11 @@ primary
        input.error_at_tok("undefined symbol '%s'", $1);
        YYABORT;
    } else if (Ast::needs_wrap($$)) {
-        $$ = ast.ref($$);
+        $$ = ast.cap($$, NO_CAPTURE);
    }
 }
-| '('     expr ')' { $$ = ast.cap($2); } // capturing group (in POSIX mode)
-| '(' '?' expr ')' { $$ = ast.ref($3); } // non-capturing group
+| '('     expr ')' { $$ = ast.cap($2, CAPTURE_IF_NOT_INVERTED); }
+| '(' '!' expr ')' { $$ = ast.cap($3, CAPTURE_IF_INVERTED); }
 ;

 %%
@@ -83,6 +83,15 @@ LOCAL_NODISCARD(Regexp* fictive_tags(RESpec& spec, int32_t height)) {
    return re_cat(spec, t1, t2);
 }

+LOCAL_NODISCARD(bool is_capture_mode(uint8_t mode, const opt_t* opts)) {
+    return mode == CAPTURE ||
+        mode == (opts->invert_captures ? CAPTURE_IF_INVERTED : CAPTURE_IF_NOT_INVERTED);
+}
+
+LOCAL_NODISCARD(bool is_capture(const AstNode* a, const opt_t* opts)) {
+    return a->kind == AstKind::CAP && is_capture_mode(a->cap.mode, opts);
+}
+
 LOCAL_NODISCARD(Regexp* capture_tags(
        RESpec& spec, DfsAstToRe& x, bool orbit, const AstNode** psub, size_t* pncap)) {
    std::vector<Tag>& tags = spec.tags;
@@ -90,14 +99,14 @@ LOCAL_NODISCARD(Regexp* capture_tags(
    size_t ncap = *pncap, lcap = ncap;

    const AstNode* ast = *psub;
-    if (ast->kind == AstKind::CAP) {
+    if (is_capture(ast, spec.opts)) {
        // save the range of repeated captures and collapse them: (...(R)...) -> (R)
-        for (ast = ast->cap; ast && ast->kind == AstKind::CAP; ast = ast->cap) {
+        for (ast = ast->cap.ast; ast && is_capture(ast, spec.opts); ast = ast->cap.ast) {
            ++ncap;
        }
        // dereference to avoid future check for non-parenthesized rerefences
-        if (ast->kind == AstKind::REF) {
-            ast = ast->ref;
+        if (ast->kind == AstKind::CAP) { // non-capturing group
+            ast = ast->cap.ast;
        }
        *psub = ast;
    }
@@ -116,12 +125,12 @@ LOCAL_NODISCARD(Regexp* capture_tags(

 LOCAL_NODISCARD(Regexp* structural_tags(
        RESpec& spec, DfsAstToRe& x, const AstNode* sub, size_t* pncap)) {
-    if (sub->kind == AstKind::CAP) {
+    if (is_capture(sub, spec.opts)) {
        // If this sub-AST is already a capture, no need for structural tags.
    } else if (spec.opts->tags_automatic) {
        // Full parsing: automatically add tags as if this sub-regexp was a capture.
        return capture_tags(spec, x, false, &x.ast, pncap);
-    } else if (spec.opts->tags_posix_semantics && x.ast->has_caps) {
+    } else if (spec.opts->tags_posix_semantics && is_capture_mode(x.ast->has_caps, spec.opts)) {
        // POSIX submatch extraction: add fictive structural tags.
        // See note [POSIX subexpression hierarchy].
        return fictive_tags(spec, x.height);
@@ -314,12 +323,8 @@ LOCAL_NODISCARD(Ret diff_to_range(RESpec& spec,
            break;

        case AstKind::CAP:
-            if (spec.opts->tags_posix_syntax) goto error;
-            x.ast = ast->cap; // replace on stack
-            break;
-
-        case AstKind::REF:
-            x.ast = ast->ref; // replace on stack
+            if (is_capture(ast, spec.opts) && spec.opts->tags_posix_syntax) goto error;
+            x.ast = ast->cap.ast; // replace on stack
            break;

        case AstKind::DIFF:
@@ -380,7 +385,7 @@ LOCAL_NODISCARD(Ret ast_to_re(RESpec& spec,

        const AstNode* ast = x.ast;

-        if (ast->kind != AstKind::CAP && ast->kind != AstKind::REF) ++x.height;
+        if (ast->kind != AstKind::CAP) ++x.height;

        switch (ast->kind) {
        case AstKind::NIL:
@@ -432,27 +437,17 @@ LOCAL_NODISCARD(Ret ast_to_re(RESpec& spec,

        case AstKind::CAP:
            if (!opts->tags_posix_syntax) { // ordinary group, replace with subexpr on stack
-                x.ast = ast->cap;
-            } else { // capturing group
+                x.ast = ast->cap.ast;
+            } else { // capturing or non-capturing group
                if (x.succ == 0) { // 1st visit: push successor
                    ++x.succ;
-                    x.re1 = capture_tags(spec, x, false, &ast, pncap);
-                    stack.emplace_back(ast, x.height, x.in_iter);
-                } else { // 2nd visit: return
-                    re = insert_between_tags(spec, x.re1, re);
-                    stack.pop_back();
-                }
-            }
-            break;
-
-        case AstKind::REF:
-            if (!opts->tags_posix_semantics) { // ordinary group, replace with subexpr on stack
-                x.ast = ast->ref;
-            } else { // non-capturing group
-                if (x.succ == 0) { // 1st visit: push successor
-                    ++x.succ;
-                    x.re1 = structural_tags(spec, x, ast->ref, pncap);
-                    stack.emplace_back(ast->ref, x.height, x.in_iter);
+                    if (is_capture(ast, opts)) {
+                        x.re1 = capture_tags(spec, x, false, &ast, pncap);
+                        stack.emplace_back(ast, x.height, x.in_iter);
+                    } else {
+                        x.re1 = structural_tags(spec, x, ast->cap.ast, pncap);
+                        stack.emplace_back(ast->cap.ast, x.height, x.in_iter);
+                    }
                } else { // 2nd visit: return
                    re = insert_between_tags(spec, x.re1, re);
                    stack.pop_back();
@@ -500,8 +495,7 @@ LOCAL_NODISCARD(Ret ast_to_re(RESpec& spec,
                ++x.succ;
                const uint32_t m = ast->iter.max;
                ast = ast->iter.ast;
-                if ((opts->tags_posix_semantics && ast->kind == AstKind::CAP)
-                        || opts->tags_automatic) {
+                if ((opts->tags_posix_semantics && is_capture(ast, opts)) || opts->tags_automatic) {
                    x.re1 = capture_tags(spec, x, m > 1, &ast, pncap);
                }
                if (m > 0) {
@@ -1,19 +1,3 @@
-#pragma GCC diagnostic push
-#pragma GCC diagnostic ignored "-Wunused-macros"
-
-// See https://github.com/skvadrik/re2c/pull/376
-// _POSIX_SOURCE is needed for fdopen on Cygwin.
-#ifndef _POSIX_SOURCE
-#define _POSIX_SOURCE 1
-#endif
-// However the above definition of _POSIX_SOURCE breaks the build on FreeBSD,
-// which also needs _POSIX_C_SOURCE to be defined.
-#ifndef _POSIX_C_SOURCE
-#define _POSIX_C_SOURCE 200112L
-#endif
-
-#pragma GCC diagnostic pop
-
 #include "config.h"
 #include <time.h>
 #include "src/util/file_utils.h"
@@ -1,5 +1,5 @@
 /* Generated by re2c */
-// re2c $INPUT -o $OUTPUT -i --leftmost-captures --input custom
+// re2c $INPUT -o $OUTPUT -i --input custom
 // POSIX and leftmost policies produce different results.
 #include <stddef.h>
 #include <stdio.h>
@@ -66,4 +66,4 @@ int main(int argc, char **argv)
    }
    return 0;
 }
-captures/diff1_leftmost.re:24:22: warning: rule matches empty string [-Wmatch-empty-string]
+captures/diff1_leftmost.re:25:22: warning: rule matches empty string [-Wmatch-empty-string]
@@ -1,4 +1,4 @@
-// re2c $INPUT -o $OUTPUT -i --leftmost-captures --input custom
+// re2c $INPUT -o $OUTPUT -i --input custom
 // POSIX and leftmost policies produce different results.
 #include <stddef.h>
 #include <stdio.h>
@@ -19,6 +19,7 @@ static void lex(const char *s)
    /*!re2c
        re2c:define:YYCTYPE = char;
        re2c:yyfill:enable = 0;
+        re2c:leftmost-captures = 1;

        * { printf("error\n"); return; }
        ("a" | "aa")* {
@@ -1,5 +1,5 @@
 /* Generated by re2c */
-// re2c $INPUT -o $OUTPUT -i --posix-captures --input custom
+// re2c $INPUT -o $OUTPUT -i --input custom
 // POSIX and leftmost policies produce different results.
 #include <stddef.h>
 #include <stdio.h>
@@ -75,4 +75,4 @@ int main(int argc, char **argv)
    }
    return 0;
 }
-captures/diff1_posix.re:24:22: warning: rule matches empty string [-Wmatch-empty-string]
+captures/diff1_posix.re:25:22: warning: rule matches empty string [-Wmatch-empty-string]
@@ -1,4 +1,4 @@
-// re2c $INPUT -o $OUTPUT -i --posix-captures --input custom
+// re2c $INPUT -o $OUTPUT -i --input custom
 // POSIX and leftmost policies produce different results.
 #include <stddef.h>
 #include <stdio.h>
@@ -19,6 +19,7 @@ static void lex(const char *s)
    /*!re2c
        re2c:define:YYCTYPE = char;
        re2c:yyfill:enable = 0;
+        re2c:posix-captures = 1;

        * { printf("error\n"); return; }
        ("a" | "aa")* {
@@ -1,41 +1,41 @@
 // re2c $INPUT -o $OUTPUT -i --leftmost-captures

 /*!local:re2c
-    x = (?"a");
+    x = (!"a");
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" "b");
+    x = (!"a" "b");
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" ("b"));
+    x = (!"a" ("b"));
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" | "b");
+    x = (!"a" | "b");
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" | ("b"));
+    x = (!"a" | ("b"));
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" | "aa");
+    x = (!"a" | "aa");
    x*x* {}
 */

 /*!local:re2c
-    x = (?"a" | ("aa"));
+    x = (!"a" | ("aa"));
    x*x* {}
 */
@@ -0,0 +1,241 @@
+/* Generated by re2c */
+// re2c $INPUT -o $OUTPUT -i --leftmost-captures --invert-captures
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 2) YYFILL(2);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy3;
+		default: goto yy1;
+	}
+yy1:
+	++YYCURSOR;
+yy2:
+	{}
+yy3:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy4;
+		default: goto yy2;
+	}
+yy4:
+	++YYCURSOR;
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 4) YYFILL(4);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy8;
+		default: goto yy6;
+	}
+yy6:
+	++YYCURSOR;
+yy7:
+	{}
+yy8:
+	yych = *(YYMARKER = ++YYCURSOR);
+	switch (yych) {
+		case 'b': goto yy9;
+		default: goto yy7;
+	}
+yy9:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy11;
+		default: goto yy10;
+	}
+yy10:
+	YYCURSOR = YYMARKER;
+	goto yy7;
+yy11:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'b': goto yy12;
+		default: goto yy10;
+	}
+yy12:
+	++YYCURSOR;
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 4) YYFILL(4);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy16;
+		default: goto yy14;
+	}
+yy14:
+	++YYCURSOR;
+yy15:
+	{}
+yy16:
+	yych = *(YYMARKER = ++YYCURSOR);
+	switch (yych) {
+		case 'b': goto yy17;
+		default: goto yy15;
+	}
+yy17:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy19;
+		default: goto yy18;
+	}
+yy18:
+	YYCURSOR = YYMARKER;
+	goto yy15;
+yy19:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'b': goto yy20;
+		default: goto yy18;
+	}
+yy20:
+	++YYCURSOR;
+	yynmatch = 2;
+	yypmatch[0] = YYCURSOR - 3;
+	yypmatch[1] = YYCURSOR - 2;
+	yypmatch[2] = YYCURSOR - 1;
+	yypmatch[3] = YYCURSOR;
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 2) YYFILL(2);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a':
+		case 'b': goto yy24;
+		default: goto yy22;
+	}
+yy22:
+	++YYCURSOR;
+yy23:
+	{}
+yy24:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a':
+		case 'b': goto yy25;
+		default: goto yy23;
+	}
+yy25:
+	++YYCURSOR;
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 2) YYFILL(2);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy29;
+		case 'b': goto yy30;
+		default: goto yy27;
+	}
+yy27:
+	++YYCURSOR;
+yy28:
+	{}
+yy29:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a':
+			yyt1 = NULL;
+			goto yy31;
+		case 'b':
+			yyt1 = NULL;
+			goto yy33;
+		default: goto yy28;
+	}
+yy30:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a':
+			yyt1 = YYCURSOR;
+			goto yy31;
+		case 'b':
+			yyt1 = YYCURSOR;
+			goto yy33;
+		default: goto yy28;
+	}
+yy31:
+	++YYCURSOR;
+	yyt2 = NULL;
+yy32:
+	yynmatch = 2;
+	yypmatch[1] = yyt1;
+	yypmatch[3] = yyt2;
+	yypmatch[0] = yyt1;
+	if (yyt1 != NULL) yypmatch[0] -= 1;
+	yypmatch[2] = yyt2;
+	if (yyt2 != NULL) yypmatch[2] -= 1;
+	{}
+yy33:
+	++YYCURSOR;
+	yyt2 = YYCURSOR;
+	goto yy32;
+}
+
+
+
+{
+	YYCTYPE yych;
+	goto yy34;
+yy35:
+	++YYCURSOR;
+yy34:
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy35;
+		default: goto yy36;
+	}
+yy36:
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	goto yy37;
+yy38:
+	++YYCURSOR;
+yy37:
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy38;
+		default:
+			yyt1 = NULL;
+			goto yy39;
+	}
+yy39:
+	yynmatch = 2;
+	yypmatch[1] = yyt1;
+	yypmatch[3] = yyt1;
+	yypmatch[0] = yyt1;
+	if (yyt1 != NULL) yypmatch[0] -= 2;
+	yypmatch[2] = yyt1;
+	if (yyt1 != NULL) yypmatch[2] -= 2;
+	{}
+}
+
+captures/noncapturing/explicit2_leftmost_inverted.re:35:9: warning: rule matches empty string [-Wmatch-empty-string]
+captures/noncapturing/explicit2_leftmost_inverted.re:40:9: warning: rule matches empty string [-Wmatch-empty-string]
@@ -0,0 +1,41 @@
+// re2c $INPUT -o $OUTPUT -i --leftmost-captures --invert-captures
+
+/*!local:re2c
+    x = ("a");
+    x x {}
+    *   {}
+*/
+
+/*!local:re2c
+    x = ("a" "b");
+    x x {}
+    *   {}
+*/
+
+/*!local:re2c
+    x = ("a" (!"b"));
+    x x {}
+    *   {}
+*/
+
+/*!local:re2c
+    x = ("a" | "b");
+    x x {}
+    *   {}
+*/
+
+/*!local:re2c
+    x = ("a" | (!"b"));
+    x x {}
+    *   {}
+*/
+
+/*!local:re2c
+    x = ("a" | "aa");
+    x*x* {}
+*/
+
+/*!local:re2c
+    x = ("a" | (!"aa"));
+    x*x* {}
+*/
@@ -1,41 +1,41 @@
 // re2c $INPUT -o $OUTPUT -i --posix-captures

 /*!local:re2c
-    x = (?"a");
+    x = (!"a");
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" "b");
+    x = (!"a" "b");
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" ("b"));
+    x = (!"a" ("b"));
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" | "b");
+    x = (!"a" | "b");
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" | ("b"));
+    x = (!"a" | ("b"));
    x x {}
    *   {}
 */

 /*!local:re2c
-    x = (?"a" | "aa");
+    x = (!"a" | "aa");
    x*x* {}
 */

 /*!local:re2c
-    x = (?"a" | ("aa"));
+    x = (!"a" | ("aa"));
    x*x* {}
 */
@@ -2,40 +2,40 @@

 /*!local:re2c
    x = "a";
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" "b";
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" ("b");
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" | "b";
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" | ("b");
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" | "aa";
-    (x)*(?x)* {}
+    (x)*(!x)* {}
 */

 /*!local:re2c
    x = "a" | ("aa");
-    (x)*(?x)* {}
+    (x)*(!x)* {}
 */
@@ -2,40 +2,40 @@

 /*!local:re2c
    x = "a";
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" "b";
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" ("b");
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" | "b";
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" | ("b");
-    (x)(?x) {}
+    (x)(!x) {}
    *       {}
 */

 /*!local:re2c
    x = "a" | "aa";
-    (x)*(?x)* {}
+    (x)*(!x)* {}
 */

 /*!local:re2c
    x = "a" | ("aa");
-    (x)*(?x)* {}
+    (x)*(!x)* {}
 */
@@ -0,0 +1,303 @@
+/* Generated by re2c */
+// re2c $INPUT -o $OUTPUT -i --posix-captures --invert-captures
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 2) YYFILL(2);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy3;
+		default: goto yy1;
+	}
+yy1:
+	++YYCURSOR;
+yy2:
+	{}
+yy3:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy4;
+		default: goto yy2;
+	}
+yy4:
+	++YYCURSOR;
+	yynmatch = 1;
+	yypmatch[0] = YYCURSOR - 2;
+	yypmatch[1] = YYCURSOR - 1;
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 4) YYFILL(4);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy8;
+		default: goto yy6;
+	}
+yy6:
+	++YYCURSOR;
+yy7:
+	{}
+yy8:
+	yych = *(YYMARKER = ++YYCURSOR);
+	switch (yych) {
+		case 'b': goto yy9;
+		default: goto yy7;
+	}
+yy9:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy11;
+		default: goto yy10;
+	}
+yy10:
+	YYCURSOR = YYMARKER;
+	goto yy7;
+yy11:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'b': goto yy12;
+		default: goto yy10;
+	}
+yy12:
+	++YYCURSOR;
+	yynmatch = 1;
+	yypmatch[0] = YYCURSOR - 4;
+	yypmatch[1] = YYCURSOR - 2;
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 4) YYFILL(4);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy16;
+		default: goto yy14;
+	}
+yy14:
+	++YYCURSOR;
+yy15:
+	{}
+yy16:
+	yych = *(YYMARKER = ++YYCURSOR);
+	switch (yych) {
+		case 'b': goto yy17;
+		default: goto yy15;
+	}
+yy17:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy19;
+		default: goto yy18;
+	}
+yy18:
+	YYCURSOR = YYMARKER;
+	goto yy15;
+yy19:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'b': goto yy20;
+		default: goto yy18;
+	}
+yy20:
+	++YYCURSOR;
+	yynmatch = 3;
+	yypmatch[0] = YYCURSOR - 4;
+	yypmatch[1] = YYCURSOR - 2;
+	yypmatch[2] = YYCURSOR - 3;
+	yypmatch[3] = YYCURSOR - 2;
+	yypmatch[4] = YYCURSOR - 1;
+	yypmatch[5] = YYCURSOR;
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 2) YYFILL(2);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a':
+		case 'b': goto yy24;
+		default: goto yy22;
+	}
+yy22:
+	++YYCURSOR;
+yy23:
+	{}
+yy24:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a':
+		case 'b': goto yy25;
+		default: goto yy23;
+	}
+yy25:
+	++YYCURSOR;
+	yynmatch = 1;
+	yypmatch[0] = YYCURSOR - 2;
+	yypmatch[1] = YYCURSOR - 1;
+	{}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if ((YYLIMIT - YYCURSOR) < 2) YYFILL(2);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy29;
+		case 'b': goto yy30;
+		default: goto yy27;
+	}
+yy27:
+	++YYCURSOR;
+yy28:
+	{}
+yy29:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a':
+			yyt1 = NULL;
+			goto yy31;
+		case 'b':
+			yyt1 = NULL;
+			goto yy33;
+		default: goto yy28;
+	}
+yy30:
+	yych = *++YYCURSOR;
+	switch (yych) {
+		case 'a':
+			yyt1 = YYCURSOR;
+			goto yy31;
+		case 'b':
+			yyt1 = YYCURSOR;
+			goto yy33;
+		default: goto yy28;
+	}
+yy31:
+	++YYCURSOR;
+	yyt2 = NULL;
+yy32:
+	yynmatch = 3;
+	yypmatch[3] = yyt1;
+	yypmatch[5] = yyt2;
+	yypmatch[0] = YYCURSOR - 2;
+	yypmatch[1] = YYCURSOR - 1;
+	yypmatch[2] = yyt1;
+	if (yyt1 != NULL) yypmatch[2] -= 1;
+	yypmatch[4] = yyt2;
+	if (yyt2 != NULL) yypmatch[4] -= 1;
+	{}
+yy33:
+	++YYCURSOR;
+	yyt2 = YYCURSOR;
+	goto yy32;
+}
+
+
+
+{
+	YYCTYPE yych;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a':
+			yyt2 = YYCURSOR;
+			goto yy36;
+		default:
+			yyt1 = yyt2 = NULL;
+			goto yy35;
+	}
+yy35:
+	yynmatch = 1;
+	yypmatch[0] = yyt2;
+	yypmatch[1] = yyt1;
+	{}
+yy36:
+	++YYCURSOR;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy37;
+		default:
+			yyt1 = YYCURSOR;
+			goto yy35;
+	}
+yy37:
+	++YYCURSOR;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a':
+			yyt2 = YYCURSOR;
+			goto yy36;
+		default:
+			yyt1 = YYCURSOR;
+			goto yy35;
+	}
+}
+
+
+
+{
+	YYCTYPE yych;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a':
+			yyt4 = YYCURSOR;
+			goto yy40;
+		default:
+			yyt1 = yyt2 = yyt3 = yyt4 = NULL;
+			goto yy39;
+	}
+yy39:
+	yynmatch = 3;
+	yypmatch[0] = yyt4;
+	yypmatch[1] = yyt1;
+	yypmatch[3] = yyt2;
+	yypmatch[5] = yyt3;
+	yypmatch[2] = yyt2;
+	if (yyt2 != NULL) yypmatch[2] -= 2;
+	yypmatch[4] = yyt3;
+	if (yyt3 != NULL) yypmatch[4] -= 2;
+	{}
+yy40:
+	++YYCURSOR;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a': goto yy41;
+		default:
+			yyt2 = yyt3 = NULL;
+			yyt1 = YYCURSOR;
+			goto yy39;
+	}
+yy41:
+	++YYCURSOR;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'a':
+			yyt4 = YYCURSOR;
+			goto yy40;
+		default:
+			yyt3 = NULL;
+			yyt1 = yyt2 = YYCURSOR;
+			goto yy39;
+	}
+}
+
+captures/noncapturing/explicit3_posix_inverted.re:35:14: warning: rule matches empty string [-Wmatch-empty-string]
+captures/noncapturing/explicit3_posix_inverted.re:40:14: warning: rule matches empty string [-Wmatch-empty-string]
@@ -0,0 +1,41 @@
+// re2c $INPUT -o $OUTPUT -i --posix-captures --invert-captures
+
+/*!local:re2c
+    x = "a";
+    (!x)(x) {}
+    *       {}
+*/
+
+/*!local:re2c
+    x = "a" "b";
+    (!x)(x) {}
+    *       {}
+*/
+
+/*!local:re2c
+    x = "a" (!"b");
+    (!x)(x) {}
+    *       {}
+*/
+
+/*!local:re2c
+    x = "a" | "b";
+    (!x)(x) {}
+    *       {}
+*/
+
+/*!local:re2c
+    x = "a" | (!"b");
+    (!x)(x) {}
+    *       {}
+*/
+
+/*!local:re2c
+    x = "a" | "aa";
+    (!x)*(x)* {}
+*/
+
+/*!local:re2c
+    x = "a" | (!"aa");
+    (!x)*(x)* {}
+*/
@@ -2,40 +2,40 @@

 /*!local:re2c
    x = "a";
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" "b";
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" ("b");
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" | "b";
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" | ("b");
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" | "aa";
-    (?x)*x* {}
+    (!x)*x* {}
 */

 /*!local:re2c
    x = "a" | ("aa");
-    (?x)*x* {}
+    (!x)*x* {}
 */
@@ -2,40 +2,40 @@

 /*!local:re2c
    x = "a";
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" "b";
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" ("b");
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" | "b";
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" | ("b");
-    (?x)x {}
+    (!x)x {}
    *     {}
 */

 /*!local:re2c
    x = "a" | "aa";
-    (?x)*x* {}
+    (!x)*x* {}
 */

 /*!local:re2c
    x = "a" | ("aa");
-    (?x)*x* {}
+    (!x)*x* {}
 */
@@ -0,0 +1,83 @@
+/* Generated by re2c */
+// re2c $INPUT -o $OUTPUT -i --leftmost-captures --invert-captures
+
+// B is capturing
+
+{
+	YYCTYPE yych;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'A': goto yy2;
+		case 'B': goto yy3;
+		default: goto yy1;
+	}
+yy1:
+	++YYCURSOR;
+	{ * }
+yy2:
+	++YYCURSOR;
+	{ A }
+yy3:
+	++YYCURSOR;
+	yynmatch = 1;
+	yypmatch[0] = YYCURSOR - 1;
+	yypmatch[1] = YYCURSOR;
+	{ B }
+}
+
+
+// B is non-capturing, A is capturing, plus the implicit whole rule capture for A and B
+
+{
+	YYCTYPE yych;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'A': goto yy6;
+		case 'B': goto yy7;
+		default: goto yy5;
+	}
+yy5:
+	++YYCURSOR;
+	{ * }
+yy6:
+	++YYCURSOR;
+	yynmatch = 2;
+	yypmatch[0] = yypmatch[2] = YYCURSOR - 1;
+	yypmatch[1] = yypmatch[3] = YYCURSOR;
+	{ A }
+yy7:
+	++YYCURSOR;
+	yynmatch = 1;
+	yypmatch[0] = YYCURSOR - 1;
+	yypmatch[1] = YYCURSOR;
+	{ B }
+}
+
+
+// B is capturing
+
+{
+	YYCTYPE yych;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'A': goto yy10;
+		case 'B': goto yy11;
+		default: goto yy9;
+	}
+yy9:
+	++YYCURSOR;
+	{ * }
+yy10:
+	++YYCURSOR;
+	{ A }
+yy11:
+	++YYCURSOR;
+	yynmatch = 1;
+	yypmatch[0] = YYCURSOR - 1;
+	yypmatch[1] = YYCURSOR;
+	{ B }
+}
+
@@ -0,0 +1,24 @@
+// re2c $INPUT -o $OUTPUT -i --leftmost-captures --invert-captures
+
+// B is capturing
+/*!re2c
+    ([A])  { A }
+    (![B]) { B }
+    *      { * }
+*/
+
+// B is non-capturing, A is capturing, plus the implicit whole rule capture for A and B
+/*!re2c
+    re2c:invert-captures = 0;
+    ([A])  { A }
+    (![B]) { B }
+    *      { * }
+*/
+
+// B is capturing
+/*!re2c
+    re2c:invert-captures = 1;
+    ([A])  { A }
+    (![B]) { B }
+    *      { * }
+*/
@@ -1,5 +1,5 @@
 /* Generated by re2c */
 #line 1 "config/all_configs.re"
 // re2c $INPUT -o $OUTPUT
-#line 156 "config/all_configs.re"
+#line 158 "config/all_configs.re"

@@ -19,9 +19,11 @@
    re2c:tags = 0;
    re2c:flags:tags = 0;
    re2c:flags:T = 0;
+    re2c:leftmost-captures = 0;
    re2c:posix-captures = 0;
    re2c:flags:posix-captures = 0;
    re2c:flags:P = 0;
+    re2c:invert-captures = 0;
    re2c:tags:prefix = "";
    re2c:tags:expression = "";

@@ -1,4 +1,4 @@
-// re2c $INPUT -o $OUTPUT --start-conditions --conditions --emit-dot --storable-state --flex-syntax --verbose --no-debug-info --no-generation-date --no-version --skeleton --eager-skip --loop-switch --bit-vectors --debug-output --case-ranges --computed-gotos --nested-ifs --case-insensitive --case-inverted --tags --posix-captures --no-unsafe --ebcdic --ecb --utf32 --unicode --ucs2 --wide-chars --utf16 --utf-16 --utf8 --utf-8 --lang rust --lang go --lang c --header all_options.h --type-header all_options.h --depfile all_options.d --encoding-policy ignore --encoding-policy substitute --encoding-policy fail --api custom --api default --input custom --input default --empty-class match-empty --empty-class match-none --empty-class error --location-format gnu --location-format msvc --input-encoding ascii --input-encoding utf8 --single-pass --reusable --dfa-minimization table --dfa-minimization moore --posix-prectable naive --posix-prectable complex --fixed-tags none --fixed-tags toplevel --fixed-tags all  --no-optimize-tags --dump-nfa --dump-dfa-tree --dump-dfa-raw --dump-dfa-det --dump-dfa-tagopt --dump-dfa-min --dump-adfa --dump-cfg --dump-interf --dump-closure-stats -cDfFiSbdgsTeuwx8P1r -I . -t all_options.h
+// re2c $INPUT -o $OUTPUT --start-conditions --conditions --emit-dot --storable-state --flex-syntax --verbose --no-debug-info --no-generation-date --no-version --skeleton --eager-skip --loop-switch --bit-vectors --debug-output --case-ranges --computed-gotos --nested-ifs --case-insensitive --case-inverted --tags --leftmost-captures --posix-captures --invert-captures --no-unsafe --ebcdic --ecb --utf32 --unicode --ucs2 --wide-chars --utf16 --utf-16 --utf8 --utf-8 --lang rust --lang go --lang c --header all_options.h --type-header all_options.h --depfile all_options.d --encoding-policy ignore --encoding-policy substitute --encoding-policy fail --api custom --api default --input custom --input default --empty-class match-empty --empty-class match-none --empty-class error --location-format gnu --location-format msvc --input-encoding ascii --input-encoding utf8 --single-pass --reusable --dfa-minimization table --dfa-minimization moore --posix-prectable naive --posix-prectable complex --fixed-tags none --fixed-tags toplevel --fixed-tags all  --no-optimize-tags --dump-nfa --dump-dfa-tree --dump-dfa-raw --dump-dfa-det --dump-dfa-tagopt --dump-dfa-min --dump-adfa --dump-cfg --dump-interf --dump-closure-stats -cDfFiSbdgsTeuwx8P1r -I . -t all_options.h

 // Not really *all* options: -v, -V, -h and variations are missing, as their
 // output depends on the version, build configuration and OS (help foramtting).
@@ -0,0 +1,268 @@
+// Code generated by re2c, DO NOT EDIT.
+//go:generate re2go $INPUT -o $OUTPUT --utf8 --input-encoding utf8 --flex-syntax -i
+package main
+
+func parse(str string) int {
+  var cur, mar int
+  for {
+	
+{
+	var yych byte
+	yych = str[cur]
+	switch (yych) {
+	case 0x00:
+		goto yy1
+	case 'a':
+		goto yy4
+	case 'c':
+		goto yy5
+	case 'e':
+		goto yy6
+	case 'g':
+		goto yy7
+	default:
+		goto yy2
+	}
+yy1:
+	cur += 1
+	{ return -1 }
+yy2:
+	cur += 1
+yy3:
+	{ continue }
+yy4:
+	cur += 1
+	mar = cur
+	yych = str[cur]
+	switch (yych) {
+	case 'b':
+		goto yy8
+	case 0xC3:
+		goto yy9
+	default:
+		goto yy3
+	}
+yy5:
+	cur += 1
+	mar = cur
+	yych = str[cur]
+	switch (yych) {
+	case 0xC3:
+		goto yy11
+	default:
+		goto yy3
+	}
+yy6:
+	cur += 1
+	mar = cur
+	yych = str[cur]
+	switch (yych) {
+	case 0xC3:
+		goto yy12
+	default:
+		goto yy3
+	}
+yy7:
+	cur += 1
+	mar = cur
+	yych = str[cur]
+	switch (yych) {
+	case 0xC3:
+		goto yy13
+	default:
+		goto yy3
+	}
+yy8:
+	cur += 1
+	{ return 1 }
+yy9:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 0xBC:
+		goto yy14
+	default:
+		goto yy10
+	}
+yy10:
+	cur = mar
+	goto yy3
+yy11:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 0xBC:
+		goto yy15
+	default:
+		goto yy10
+	}
+yy12:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 0xBC:
+		goto yy16
+	default:
+		goto yy10
+	}
+yy13:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 0xBC:
+		goto yy17
+	default:
+		goto yy10
+	}
+yy14:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'b':
+		goto yy8
+	default:
+		goto yy10
+	}
+yy15:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 0xC3:
+		goto yy18
+	default:
+		goto yy10
+	}
+yy16:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'f':
+		goto yy19
+	default:
+		goto yy10
+	}
+yy17:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'n':
+		goto yy20
+	default:
+		goto yy10
+	}
+yy18:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 0xBC:
+		goto yy21
+	default:
+		goto yy10
+	}
+yy19:
+	cur += 1
+	{ return 3 }
+yy20:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'c':
+		goto yy22
+	default:
+		goto yy10
+	}
+yy21:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'd':
+		goto yy23
+	case 0xC3:
+		goto yy18
+	default:
+		goto yy10
+	}
+yy22:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'e':
+		goto yy24
+	default:
+		goto yy10
+	}
+yy23:
+	cur += 1
+	{ return 2 }
+yy24:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'l':
+		goto yy25
+	default:
+		goto yy10
+	}
+yy25:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'l':
+		goto yy26
+	default:
+		goto yy10
+	}
+yy26:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'e':
+		goto yy27
+	default:
+		goto yy10
+	}
+yy27:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'm':
+		goto yy28
+	case 'n':
+		goto yy29
+	default:
+		goto yy10
+	}
+yy28:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'e':
+		goto yy30
+	default:
+		goto yy10
+	}
+yy29:
+	cur += 1
+	yych = str[cur]
+	switch (yych) {
+	case 'm':
+		goto yy28
+	default:
+		goto yy10
+	}
+yy30:
+	cur += 1
+	{ return 0 }
+}
+
+  }
+}
+
+func main() {
+	assert_eq := func(x, y int) { if x != y { panic("error") } }
+	assert_eq(parse("<html><body><p><em>Son güncelleme: 5/5/2020</em></p></body></html>\000"), 0)
+	assert_eq(parse("ab"), 1)
+	assert_eq(parse("aüb"), 1)
+	assert_eq(parse("cüüüüd"), 2)
+	assert_eq(parse("eüf"), 3)
+}
@@ -0,0 +1,32 @@
+//go:generate re2go $INPUT -o $OUTPUT --utf8 --input-encoding utf8 --flex-syntax -i
+package main
+
+func parse(str string) int {
+  var cur, mar int
+  for {
+	/*!re2c
+	re2c:yyfill:enable    = 0;
+	re2c:define:YYCTYPE   = byte;
+	re2c:define:YYPEEK    = "str[cur]";
+	re2c:define:YYSKIP    = "cur += 1";
+	re2c:define:YYBACKUP  = "mar = cur";
+	re2c:define:YYRESTORE = "cur = mar";
+
+	güncellen?me { return 0 }
+	aü?b         { return 1 }
+	cüüü*d       { return 2 }
+	e[ü]f        { return 3 }
+	[\000]       { return -1 }
+	*            { continue }
+	*/
+  }
+}
+
+func main() {
+	assert_eq := func(x, y int) { if x != y { panic("error") } }
+	assert_eq(parse("<html><body><p><em>Son güncelleme: 5/5/2020</em></p></body></html>\000"), 0)
+	assert_eq(parse("ab"), 1)
+	assert_eq(parse("aüb"), 1)
+	assert_eq(parse("cüüüüd"), 2)
+	assert_eq(parse("eüf"), 3)
+}
@@ -0,0 +1,4 @@
+/*!re2c
+    [b] { b }
+    *   { * }
+*/
@@ -0,0 +1,4 @@
+/*!re2c
+    [c] { c }
+    *   { * }
+*/
@@ -0,0 +1,61 @@
+/* Generated by re2c */
+#line 1 "include/include008_header.re"
+// re2c $INPUT -o $OUTPUT --depfile $DEPFILE --header include008.h
+#line 1 "include/include008.re.b.inc"
+
+#line 7 "include/include008_header.c"
+{
+	YYCTYPE yych;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'b': goto yy2;
+		default: goto yy1;
+	}
+yy1:
+	++YYCURSOR;
+#line 3 "include/include008.re.b.inc"
+	{ * }
+#line 20 "include/include008_header.c"
+yy2:
+	++YYCURSOR;
+#line 2 "include/include008.re.b.inc"
+	{ b }
+#line 25 "include/include008_header.c"
+}
+#line 4 "include/include008.re.b.inc"
+
+#line 2 "include/include008_header.re"
+
+
+#line 6 "include/include008_header.re"
+
+/* Generated by re2c */
+
+#line 1 "include/include008.re.c.inc"
+
+#line 6 "include/include008.h"
+{
+	YYCTYPE yych;
+	if (YYLIMIT <= YYCURSOR) YYFILL(1);
+	yych = *YYCURSOR;
+	switch (yych) {
+		case 'c': goto yy5;
+		default: goto yy4;
+	}
+yy4:
+	++YYCURSOR;
+#line 3 "include/include008.re.c.inc"
+	{ * }
+#line 19 "include/include008.h"
+yy5:
+	++YYCURSOR;
+#line 2 "include/include008.re.c.inc"
+	{ c }
+#line 24 "include/include008.h"
+}
+#line 4 "include/include008.re.c.inc"
+
+#line 5 "include/include008_header.re"
+
+include/include008_header.c include/include008.h: include/include008.re.b.inc include/include008.re.c.inc include/include008_header.re
@@ -0,0 +1,6 @@
+// re2c $INPUT -o $OUTPUT --depfile $DEPFILE --header include008.h
+/*!include:re2c "include008.re.b.inc" */
+
+/*!header:re2c:on */
+/*!include:re2c "include008.re.c.inc" */
+/*!header:re2c:off */
Author	SHA1	Message	Date
Ulya Trofimovich	ece4bb76e3	Release 3.1.	2023-07-19 22:50:22 +01:00
Ulya Trofimovich	c3f267b248	release.sh: make the script NixOS-friendly, fix quoting in echo command.	2023-07-19 22:42:25 +01:00
Ulya Trofimovich	60b40d94bd	Makefile.lib.am: remove nonexistent header from the list.	2023-07-19 21:47:14 +01:00
Ulya Trofimovich	5abc35226b	Update CHANGELOG.	2023-07-18 22:56:45 +01:00
Ulya Trofimovich	e868cc58b9	Document configure option --enable-parsers in BUILD.md.	2023-07-13 23:22:55 +01:00
Ulya Trofimovich	c633d5de3e	Update documentation for capturing and non-capturing groups. Document the following options: --leftomost-captures --invert-captures And the following configurations: re2c:leftmost-captures re2c:invert-captures And the new `(! ...)` syntax.	2023-07-13 22:59:47 +01:00
Ulya Trofimovich	c86a8bcf5d	Add tests for `re2c:leftmost-captures` and `re2c:posix-captures`.	2023-07-13 22:27:13 +01:00
Ulya Trofimovich	3a98b54363	Add configuration `re2c:leftmost-captures`. It has the same meaning as the --leftmost-captures option.	2023-07-13 22:16:37 +01:00
Ulya Trofimovich	0384cffd9f	Fix build scripts for Wine to run on NixOS. Wine cannot find DLLs by their absolute paths on NixOS, so find and copy them before running re2c.exe.	2023-07-13 08:06:00 +01:00
Ulya Trofimovich	a303a8500d	Add missing header include to make src/codegen/code.h self-contained.	2023-07-13 06:22:48 +01:00
Ulya Trofimovich	8c2127d420	Replace `enum GorPass` with uint32_t as it is used in a bitfield. It cannot be a `enum class` because it's used in a bitfield and this triggers compiler warnings and errors on some toolchains (see https://github.com/skvadrik/re2c/issues/446 for details). It cannot be a plain `enum` defaulting to `int` as well because of -Wsigned-enum-bitfield warning (noticed when testing on OpenBSD with clang-13).	2023-07-12 22:48:57 +01:00
Ulya Trofimovich	54ffc58440	Make function `static` as it is used in one file.	2023-07-12 22:19:09 +01:00
Ulya Trofimovich	00cbfe90af	Drop _POSIX_SOURCE, enable -std=gnu++11 for CMake. On Unix systems, re2c uses POSIX functions (e.g. fdopen). Certain systems, such as Cygwin and NetBSD, enable these funtions conditionally based on certain preprocessor definitions. Previously https://github.com/skvadrik/re2c/pull/376 fixed the build on Cygwin by setting _POSIX_SOURCE. However, this broke the build on FreeBSD which was fixed by commit This worked for Autotools build system which uses -std=c++11 due to the use of AC_USE_SYSTEM_EXTENSIONS, but it didn't work for CMake that used -std=c++11 and disabled extensions with `set(CMAKE_CXX_EXTENSIONS OFF)`. This commit drops `set(CMAKE_CXX_EXTENSIONS OFF)`, which effectively enables -std=gnu++11 instead of -std=c++11 for CMake builds. This allows us to get rid of the problematic _POSIX_SOURCE that breaks the build on NetBSD. This was tested on NetBSD-9.3, FreeBSD-13.2 and Cygwin with both Autotools and CMake.	2023-07-11 23:34:53 +01:00
Ulya Trofimovich	bb143b980c	libre2c: omit redundant semicolon after namespace. This gets rid of compiler warnings on some older toolchains (e.g. GCC-10 emits pedantic diagnostic about it, which was fixed in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96068).	2023-07-08 15:02:40 +01:00
Ulya Trofimovich	fce87b493f	Omit redundant semicolon after namespace. This gets rid of compiler warnings on some older toolchains (e.g. GCC-10 emits pedantic diagnostic about it, which was fixed in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96068).	2023-07-08 14:17:31 +01:00
Ulya Trofimovich	d87f86ed71	Correctly parse raw UTF-8 characters in Flex mode. With `--flex-syntax` option re2c accepts raw (unquoted) strings. If `--input-encoding utf8` is also enabled, re2c should parse raw UTF-8 characters in such strings. Previously it failed to do so because the characters in raw string literals were limited to alphanumeric ASCII and underscore (those allowed in indentifiers, as in normal mode unquoted strings would be parsed as identifiers). This commit fixes it: now in Flex mode and with UTF-8 input, if all previous rules failed, re2c tries to lex the input as a UTF-8 character. This fixes bug #447 "Re2go fail when compiling expression containing combining character".	2023-06-22 07:53:03 +01:00
Ulya Trofimovich	d7c697138a	Convert `GorPass` from enum class to plain enum to fix compiler error. GCC-5.4 is unhappy that `GorPass` which is `enum class: uint32_t` is stored in a bitfield of size 2 (2 bits). In practice it is safe since the enum has only three values (0, 1, 2). Reported in: - #445 "make failure for re2c" - #446 "make failure for re2c on centos7"	2023-06-18 07:41:04 +01:00
Ulya Trofimovich	580ab173dd	run_tests.py: remove exception for overflow-1.re test on Windows. The test no longer causes stack overflow, so it can now run on Windows which has a smaller stack than Linux and macOS.	2023-04-27 22:15:20 +01:00
Ulya Trofimovich	6e844850bb	run_tests.py: patch backslashes in depfile to fix tests on Windows. Change file context definition to cover more scenarios that include filenames, e.g. the case when there is a depfile generated and appended to the output.	2023-04-27 22:15:20 +01:00
helly25	3205c867bd	Provide a min viable http://bazel.build integration. This allows other projects that use bazel to utilize re2c as a tool. Limitations: . Input files (docs, lexer, parser) are not being re-generated. . No tests are executed. . Only the main binary is being build as nothing else is needed. A user will load the package using bazel's 'http_archive' in their WORKSPACE file. In their BUILD files they can then use: ``` load("@re2c//bazel:re2c.bzl", "re2c_gen") re2c_gen( name = "file_re2c", srcs = ["file.re2c"], outs = [ "file.cc", "file.h", ] ) cc_library( name = "file_cc", srcs = ["file.cc"], hdrs = ["file.h"], ) ``` The approach here is to automate as much as possible. In that way, all source and header files will be found automatically (using glob). Also, the provided bazel integration is kept to an absolute minimum. An alternative is to use a `rules_foreign_cc` based approach. But that would require much more bazel work and knowledge. Plus the result would end up being a non-hermetic build, thereby bypassing bazel's advantages. Further, using a rules_foreign_cc approach is much slower (for me ~x10).	2023-04-27 13:22:59 +01:00
Ulya Trofimovich	2dda36aa66	Header file should be a dependency, not a target. So move it before the colon in the dependency file generated with --depfile option. This fixes bug #440.	2023-04-26 00:04:33 +01:00
Ulya Trofimovich	f807f7636e	Add header file to the dependencies generated with --depfile option. This fixes bug #440 "Generated header file is not present in depfile".	2023-04-25 23:44:14 +01:00
Ulya Trofimovich	ce75619557	Add configuration `re2c:invert_captures`. Make option --invert-captures configurable. This requires postponing the decision whether a group is capturing or not to the after-parsing stage when the configurations for a block are known and the immutable AST is converted to a regexp. Add a test for the new configuration.	2023-04-24 23:55:49 +01:00
Ulya Trofimovich	dedd72dd66	Add tests for --invert-captures option.	2023-04-24 23:55:49 +01:00
Ulya Trofimovich	bf3ff8f3ab	Merge AST nodes CAP and REF for capturing and non-capturing groups. These nodes are similar enough so that it is easier to handle them as one node with a boolean parameter.	2023-04-21 22:12:40 +01:00
Ulya Trofimovich	20030ff194	Add option --invert-captures. The new option inverts the meaning on capturing and non-capturing parentheses: `(! ...)` becomes capturing, and `( ... )` becomes non-capturing, including the implicit whole-regexp group, which also becomes non-capturing. The idea of the new option is to let the user decide which gropus should be capturing and which are purely structural, and explicitly mark only the capturing ones for submatch extraction.	2023-04-21 22:10:14 +01:00
Ulya Trofimovich	338806b96c	Change syntax for explicit non-capturing groups. The old syntax was `(? ...)`, the new syntax is `(! ...)`. This is in preparation to add --invert-captures option that will flip the meaning of capturing and non-capturing parentheses.	2023-04-19 07:20:42 +01:00