Use a proper C tokenizer to implement the obsolete typedefs test.
The test for obsolete typedefs in installed headers was implemented
using grep, and could therefore get false positives on e.g. “ulong”
in a comment. It was also scanning all of the headers included by
our headers, and therefore testing headers we don’t control, e.g.
Linux kernel headers.
This patch splits the obsolete-typedef test from
scripts/check-installed-headers.sh to a separate program,
scripts/check-obsolete-constructs.py. Being implemented in Python,
it is feasible to make it tokenize C accurately enough to avoid false
positives on the contents of comments and strings. It also only
examines $(headers) in each subdirectory--all the headers we install,
but not any external dependencies of those headers. Headers whose
installed name starts with finclude/ are ignored, on the assumption
that they contain Fortran.
It is also feasible to make the new test understand the difference
between _defining_ the obsolete typedefs and _using_ the obsolete
typedefs, which means posix/{bits,sys}/types.h no longer need to be
exempted. This uncovered an actual bug in bits/types.h: __quad_t and
__u_quad_t were being used to define __S64_TYPE, __U64_TYPE,
__SQUAD_TYPE and __UQUAD_TYPE. These are changed to __int64_t and
__uint64_t respectively. This is a safe change, despite the comments
in bits/types.h claiming a difference between __quad_t and __int64_t,
because those comments are incorrect. In all current ABIs, both
__quad_t and __int64_t are ‘long’ when ‘long’ is a 64-bit type, and
‘long long’ when ‘long’ is a 32-bit type, and similarly for __u_quad_t
and __uint64_t. (Changing the types to be what the comments say they
are would be an ABI break, as it affects C++ name mangling.) This
patch includes a minimal change to make the comments not completely
wrong.
sys/types.h was defining the legacy BSD u_intN_t typedefs using a
construct that was not necessarily consistent with how the C99 uintN_t
typedefs are defined, and is also too complicated for the new script to
understand (it lexes C relatively accurately, but it does not attempt
to expand preprocessor macros, nor does it do any actual parsing).
This patch cuts all of that out and uses bits/types.h's __uintN_t typedefs
to define u_intN_t instead. This is verified to not change the ABI on
any supported architecture, via the c++-types test, which means u_intN_t
and uintN_t were, in fact, consistent on all supported architectures.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
* scripts/check-obsolete-constructs.py: New test script.
* scripts/check-installed-headers.sh: Remove tests for
obsolete typedefs, superseded by check-obsolete-constructs.py.
* Rules: Run scripts/check-obsolete-constructs.py over $(headers)
as a special test. Update commentary.
* posix/bits/types.h (__SQUAD_TYPE, __S64_TYPE): Define as __int64_t.
(__UQUAD_TYPE, __U64_TYPE): Define as __uint64_t.
Update commentary.
* posix/sys/types.h (__u_intN_t): Remove.
(u_int8_t): Typedef using __uint8_t.
(u_int16_t): Typedef using __uint16_t.
(u_int32_t): Typedef using __uint32_t.
(u_int64_t): Typedef using __uint64_t.
2019-03-11 22:59:27 +08:00
|
|
|
#! /usr/bin/python3
|
2024-01-02 02:12:26 +08:00
|
|
|
# Copyright (C) 2019-2024 Free Software Foundation, Inc.
|
Use a proper C tokenizer to implement the obsolete typedefs test.
The test for obsolete typedefs in installed headers was implemented
using grep, and could therefore get false positives on e.g. “ulong”
in a comment. It was also scanning all of the headers included by
our headers, and therefore testing headers we don’t control, e.g.
Linux kernel headers.
This patch splits the obsolete-typedef test from
scripts/check-installed-headers.sh to a separate program,
scripts/check-obsolete-constructs.py. Being implemented in Python,
it is feasible to make it tokenize C accurately enough to avoid false
positives on the contents of comments and strings. It also only
examines $(headers) in each subdirectory--all the headers we install,
but not any external dependencies of those headers. Headers whose
installed name starts with finclude/ are ignored, on the assumption
that they contain Fortran.
It is also feasible to make the new test understand the difference
between _defining_ the obsolete typedefs and _using_ the obsolete
typedefs, which means posix/{bits,sys}/types.h no longer need to be
exempted. This uncovered an actual bug in bits/types.h: __quad_t and
__u_quad_t were being used to define __S64_TYPE, __U64_TYPE,
__SQUAD_TYPE and __UQUAD_TYPE. These are changed to __int64_t and
__uint64_t respectively. This is a safe change, despite the comments
in bits/types.h claiming a difference between __quad_t and __int64_t,
because those comments are incorrect. In all current ABIs, both
__quad_t and __int64_t are ‘long’ when ‘long’ is a 64-bit type, and
‘long long’ when ‘long’ is a 32-bit type, and similarly for __u_quad_t
and __uint64_t. (Changing the types to be what the comments say they
are would be an ABI break, as it affects C++ name mangling.) This
patch includes a minimal change to make the comments not completely
wrong.
sys/types.h was defining the legacy BSD u_intN_t typedefs using a
construct that was not necessarily consistent with how the C99 uintN_t
typedefs are defined, and is also too complicated for the new script to
understand (it lexes C relatively accurately, but it does not attempt
to expand preprocessor macros, nor does it do any actual parsing).
This patch cuts all of that out and uses bits/types.h's __uintN_t typedefs
to define u_intN_t instead. This is verified to not change the ABI on
any supported architecture, via the c++-types test, which means u_intN_t
and uintN_t were, in fact, consistent on all supported architectures.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
* scripts/check-obsolete-constructs.py: New test script.
* scripts/check-installed-headers.sh: Remove tests for
obsolete typedefs, superseded by check-obsolete-constructs.py.
* Rules: Run scripts/check-obsolete-constructs.py over $(headers)
as a special test. Update commentary.
* posix/bits/types.h (__SQUAD_TYPE, __S64_TYPE): Define as __int64_t.
(__UQUAD_TYPE, __U64_TYPE): Define as __uint64_t.
Update commentary.
* posix/sys/types.h (__u_intN_t): Remove.
(u_int8_t): Typedef using __uint8_t.
(u_int16_t): Typedef using __uint16_t.
(u_int32_t): Typedef using __uint32_t.
(u_int64_t): Typedef using __uint64_t.
2019-03-11 22:59:27 +08:00
|
|
|
# This file is part of the GNU C Library.
|
|
|
|
#
|
|
|
|
# The GNU C Library is free software; you can redistribute it and/or
|
|
|
|
# modify it under the terms of the GNU Lesser General Public
|
|
|
|
# License as published by the Free Software Foundation; either
|
|
|
|
# version 2.1 of the License, or (at your option) any later version.
|
|
|
|
#
|
|
|
|
# The GNU C Library is distributed in the hope that it will be useful,
|
|
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
# Lesser General Public License for more details.
|
|
|
|
#
|
|
|
|
# You should have received a copy of the GNU Lesser General Public
|
|
|
|
# License along with the GNU C Library; if not, see
|
Prefer https to http for gnu.org and fsf.org URLs
Also, change sources.redhat.com to sourceware.org.
This patch was automatically generated by running the following shell
script, which uses GNU sed, and which avoids modifying files imported
from upstream:
sed -ri '
s,(http|ftp)(://(.*\.)?(gnu|fsf|sourceware)\.org($|[^.]|\.[^a-z])),https\2,g
s,(http|ftp)(://(.*\.)?)sources\.redhat\.com($|[^.]|\.[^a-z]),https\2sourceware.org\4,g
' \
$(find $(git ls-files) -prune -type f \
! -name '*.po' \
! -name 'ChangeLog*' \
! -path COPYING ! -path COPYING.LIB \
! -path manual/fdl-1.3.texi ! -path manual/lgpl-2.1.texi \
! -path manual/texinfo.tex ! -path scripts/config.guess \
! -path scripts/config.sub ! -path scripts/install-sh \
! -path scripts/mkinstalldirs ! -path scripts/move-if-change \
! -path INSTALL ! -path locale/programs/charmap-kw.h \
! -path po/libc.pot ! -path sysdeps/gnu/errlist.c \
! '(' -name configure \
-execdir test -f configure.ac -o -f configure.in ';' ')' \
! '(' -name preconfigure \
-execdir test -f preconfigure.ac ';' ')' \
-print)
and then by running 'make dist-prepare' to regenerate files built
from the altered files, and then executing the following to cleanup:
chmod a+x sysdeps/unix/sysv/linux/riscv/configure
# Omit irrelevant whitespace and comment-only changes,
# perhaps from a slightly-different Autoconf version.
git checkout -f \
sysdeps/csky/configure \
sysdeps/hppa/configure \
sysdeps/riscv/configure \
sysdeps/unix/sysv/linux/csky/configure
# Omit changes that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/powerpc/powerpc64/ppc-mcount.S: trailing lines
git checkout -f \
sysdeps/powerpc/powerpc64/ppc-mcount.S \
sysdeps/unix/sysv/linux/s390/s390-64/syscall.S
# Omit change that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S: last line does not end in newline
git checkout -f sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S
2019-09-07 13:40:42 +08:00
|
|
|
# <https://www.gnu.org/licenses/>.
|
Use a proper C tokenizer to implement the obsolete typedefs test.
The test for obsolete typedefs in installed headers was implemented
using grep, and could therefore get false positives on e.g. “ulong”
in a comment. It was also scanning all of the headers included by
our headers, and therefore testing headers we don’t control, e.g.
Linux kernel headers.
This patch splits the obsolete-typedef test from
scripts/check-installed-headers.sh to a separate program,
scripts/check-obsolete-constructs.py. Being implemented in Python,
it is feasible to make it tokenize C accurately enough to avoid false
positives on the contents of comments and strings. It also only
examines $(headers) in each subdirectory--all the headers we install,
but not any external dependencies of those headers. Headers whose
installed name starts with finclude/ are ignored, on the assumption
that they contain Fortran.
It is also feasible to make the new test understand the difference
between _defining_ the obsolete typedefs and _using_ the obsolete
typedefs, which means posix/{bits,sys}/types.h no longer need to be
exempted. This uncovered an actual bug in bits/types.h: __quad_t and
__u_quad_t were being used to define __S64_TYPE, __U64_TYPE,
__SQUAD_TYPE and __UQUAD_TYPE. These are changed to __int64_t and
__uint64_t respectively. This is a safe change, despite the comments
in bits/types.h claiming a difference between __quad_t and __int64_t,
because those comments are incorrect. In all current ABIs, both
__quad_t and __int64_t are ‘long’ when ‘long’ is a 64-bit type, and
‘long long’ when ‘long’ is a 32-bit type, and similarly for __u_quad_t
and __uint64_t. (Changing the types to be what the comments say they
are would be an ABI break, as it affects C++ name mangling.) This
patch includes a minimal change to make the comments not completely
wrong.
sys/types.h was defining the legacy BSD u_intN_t typedefs using a
construct that was not necessarily consistent with how the C99 uintN_t
typedefs are defined, and is also too complicated for the new script to
understand (it lexes C relatively accurately, but it does not attempt
to expand preprocessor macros, nor does it do any actual parsing).
This patch cuts all of that out and uses bits/types.h's __uintN_t typedefs
to define u_intN_t instead. This is verified to not change the ABI on
any supported architecture, via the c++-types test, which means u_intN_t
and uintN_t were, in fact, consistent on all supported architectures.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
* scripts/check-obsolete-constructs.py: New test script.
* scripts/check-installed-headers.sh: Remove tests for
obsolete typedefs, superseded by check-obsolete-constructs.py.
* Rules: Run scripts/check-obsolete-constructs.py over $(headers)
as a special test. Update commentary.
* posix/bits/types.h (__SQUAD_TYPE, __S64_TYPE): Define as __int64_t.
(__UQUAD_TYPE, __U64_TYPE): Define as __uint64_t.
Update commentary.
* posix/sys/types.h (__u_intN_t): Remove.
(u_int8_t): Typedef using __uint8_t.
(u_int16_t): Typedef using __uint16_t.
(u_int32_t): Typedef using __uint32_t.
(u_int64_t): Typedef using __uint64_t.
2019-03-11 22:59:27 +08:00
|
|
|
|
|
|
|
"""Verifies that installed headers do not use any obsolete constructs:
|
|
|
|
* legacy BSD typedefs superseded by <stdint.h>:
|
|
|
|
ushort uint ulong u_char u_short u_int u_long u_intNN_t quad_t u_quad_t
|
|
|
|
(sys/types.h is allowed to _define_ these types, but not to use them
|
|
|
|
to define anything else).
|
|
|
|
"""
|
|
|
|
|
|
|
|
import argparse
|
2022-09-22 18:10:41 +08:00
|
|
|
import os
|
Use a proper C tokenizer to implement the obsolete typedefs test.
The test for obsolete typedefs in installed headers was implemented
using grep, and could therefore get false positives on e.g. “ulong”
in a comment. It was also scanning all of the headers included by
our headers, and therefore testing headers we don’t control, e.g.
Linux kernel headers.
This patch splits the obsolete-typedef test from
scripts/check-installed-headers.sh to a separate program,
scripts/check-obsolete-constructs.py. Being implemented in Python,
it is feasible to make it tokenize C accurately enough to avoid false
positives on the contents of comments and strings. It also only
examines $(headers) in each subdirectory--all the headers we install,
but not any external dependencies of those headers. Headers whose
installed name starts with finclude/ are ignored, on the assumption
that they contain Fortran.
It is also feasible to make the new test understand the difference
between _defining_ the obsolete typedefs and _using_ the obsolete
typedefs, which means posix/{bits,sys}/types.h no longer need to be
exempted. This uncovered an actual bug in bits/types.h: __quad_t and
__u_quad_t were being used to define __S64_TYPE, __U64_TYPE,
__SQUAD_TYPE and __UQUAD_TYPE. These are changed to __int64_t and
__uint64_t respectively. This is a safe change, despite the comments
in bits/types.h claiming a difference between __quad_t and __int64_t,
because those comments are incorrect. In all current ABIs, both
__quad_t and __int64_t are ‘long’ when ‘long’ is a 64-bit type, and
‘long long’ when ‘long’ is a 32-bit type, and similarly for __u_quad_t
and __uint64_t. (Changing the types to be what the comments say they
are would be an ABI break, as it affects C++ name mangling.) This
patch includes a minimal change to make the comments not completely
wrong.
sys/types.h was defining the legacy BSD u_intN_t typedefs using a
construct that was not necessarily consistent with how the C99 uintN_t
typedefs are defined, and is also too complicated for the new script to
understand (it lexes C relatively accurately, but it does not attempt
to expand preprocessor macros, nor does it do any actual parsing).
This patch cuts all of that out and uses bits/types.h's __uintN_t typedefs
to define u_intN_t instead. This is verified to not change the ABI on
any supported architecture, via the c++-types test, which means u_intN_t
and uintN_t were, in fact, consistent on all supported architectures.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
* scripts/check-obsolete-constructs.py: New test script.
* scripts/check-installed-headers.sh: Remove tests for
obsolete typedefs, superseded by check-obsolete-constructs.py.
* Rules: Run scripts/check-obsolete-constructs.py over $(headers)
as a special test. Update commentary.
* posix/bits/types.h (__SQUAD_TYPE, __S64_TYPE): Define as __int64_t.
(__UQUAD_TYPE, __U64_TYPE): Define as __uint64_t.
Update commentary.
* posix/sys/types.h (__u_intN_t): Remove.
(u_int8_t): Typedef using __uint8_t.
(u_int16_t): Typedef using __uint16_t.
(u_int32_t): Typedef using __uint32_t.
(u_int64_t): Typedef using __uint64_t.
2019-03-11 22:59:27 +08:00
|
|
|
import re
|
|
|
|
import sys
|
|
|
|
|
2022-09-22 18:10:41 +08:00
|
|
|
# Make available glibc Python modules.
|
|
|
|
sys.path.append(os.path.dirname(os.path.realpath(__file__)))
|
Use a proper C tokenizer to implement the obsolete typedefs test.
The test for obsolete typedefs in installed headers was implemented
using grep, and could therefore get false positives on e.g. “ulong”
in a comment. It was also scanning all of the headers included by
our headers, and therefore testing headers we don’t control, e.g.
Linux kernel headers.
This patch splits the obsolete-typedef test from
scripts/check-installed-headers.sh to a separate program,
scripts/check-obsolete-constructs.py. Being implemented in Python,
it is feasible to make it tokenize C accurately enough to avoid false
positives on the contents of comments and strings. It also only
examines $(headers) in each subdirectory--all the headers we install,
but not any external dependencies of those headers. Headers whose
installed name starts with finclude/ are ignored, on the assumption
that they contain Fortran.
It is also feasible to make the new test understand the difference
between _defining_ the obsolete typedefs and _using_ the obsolete
typedefs, which means posix/{bits,sys}/types.h no longer need to be
exempted. This uncovered an actual bug in bits/types.h: __quad_t and
__u_quad_t were being used to define __S64_TYPE, __U64_TYPE,
__SQUAD_TYPE and __UQUAD_TYPE. These are changed to __int64_t and
__uint64_t respectively. This is a safe change, despite the comments
in bits/types.h claiming a difference between __quad_t and __int64_t,
because those comments are incorrect. In all current ABIs, both
__quad_t and __int64_t are ‘long’ when ‘long’ is a 64-bit type, and
‘long long’ when ‘long’ is a 32-bit type, and similarly for __u_quad_t
and __uint64_t. (Changing the types to be what the comments say they
are would be an ABI break, as it affects C++ name mangling.) This
patch includes a minimal change to make the comments not completely
wrong.
sys/types.h was defining the legacy BSD u_intN_t typedefs using a
construct that was not necessarily consistent with how the C99 uintN_t
typedefs are defined, and is also too complicated for the new script to
understand (it lexes C relatively accurately, but it does not attempt
to expand preprocessor macros, nor does it do any actual parsing).
This patch cuts all of that out and uses bits/types.h's __uintN_t typedefs
to define u_intN_t instead. This is verified to not change the ABI on
any supported architecture, via the c++-types test, which means u_intN_t
and uintN_t were, in fact, consistent on all supported architectures.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
* scripts/check-obsolete-constructs.py: New test script.
* scripts/check-installed-headers.sh: Remove tests for
obsolete typedefs, superseded by check-obsolete-constructs.py.
* Rules: Run scripts/check-obsolete-constructs.py over $(headers)
as a special test. Update commentary.
* posix/bits/types.h (__SQUAD_TYPE, __S64_TYPE): Define as __int64_t.
(__UQUAD_TYPE, __U64_TYPE): Define as __uint64_t.
Update commentary.
* posix/sys/types.h (__u_intN_t): Remove.
(u_int8_t): Typedef using __uint8_t.
(u_int16_t): Typedef using __uint16_t.
(u_int32_t): Typedef using __uint32_t.
(u_int64_t): Typedef using __uint64_t.
2019-03-11 22:59:27 +08:00
|
|
|
|
2022-09-22 18:10:41 +08:00
|
|
|
import glibcpp
|
Use a proper C tokenizer to implement the obsolete typedefs test.
The test for obsolete typedefs in installed headers was implemented
using grep, and could therefore get false positives on e.g. “ulong”
in a comment. It was also scanning all of the headers included by
our headers, and therefore testing headers we don’t control, e.g.
Linux kernel headers.
This patch splits the obsolete-typedef test from
scripts/check-installed-headers.sh to a separate program,
scripts/check-obsolete-constructs.py. Being implemented in Python,
it is feasible to make it tokenize C accurately enough to avoid false
positives on the contents of comments and strings. It also only
examines $(headers) in each subdirectory--all the headers we install,
but not any external dependencies of those headers. Headers whose
installed name starts with finclude/ are ignored, on the assumption
that they contain Fortran.
It is also feasible to make the new test understand the difference
between _defining_ the obsolete typedefs and _using_ the obsolete
typedefs, which means posix/{bits,sys}/types.h no longer need to be
exempted. This uncovered an actual bug in bits/types.h: __quad_t and
__u_quad_t were being used to define __S64_TYPE, __U64_TYPE,
__SQUAD_TYPE and __UQUAD_TYPE. These are changed to __int64_t and
__uint64_t respectively. This is a safe change, despite the comments
in bits/types.h claiming a difference between __quad_t and __int64_t,
because those comments are incorrect. In all current ABIs, both
__quad_t and __int64_t are ‘long’ when ‘long’ is a 64-bit type, and
‘long long’ when ‘long’ is a 32-bit type, and similarly for __u_quad_t
and __uint64_t. (Changing the types to be what the comments say they
are would be an ABI break, as it affects C++ name mangling.) This
patch includes a minimal change to make the comments not completely
wrong.
sys/types.h was defining the legacy BSD u_intN_t typedefs using a
construct that was not necessarily consistent with how the C99 uintN_t
typedefs are defined, and is also too complicated for the new script to
understand (it lexes C relatively accurately, but it does not attempt
to expand preprocessor macros, nor does it do any actual parsing).
This patch cuts all of that out and uses bits/types.h's __uintN_t typedefs
to define u_intN_t instead. This is verified to not change the ABI on
any supported architecture, via the c++-types test, which means u_intN_t
and uintN_t were, in fact, consistent on all supported architectures.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
* scripts/check-obsolete-constructs.py: New test script.
* scripts/check-installed-headers.sh: Remove tests for
obsolete typedefs, superseded by check-obsolete-constructs.py.
* Rules: Run scripts/check-obsolete-constructs.py over $(headers)
as a special test. Update commentary.
* posix/bits/types.h (__SQUAD_TYPE, __S64_TYPE): Define as __int64_t.
(__UQUAD_TYPE, __U64_TYPE): Define as __uint64_t.
Update commentary.
* posix/sys/types.h (__u_intN_t): Remove.
(u_int8_t): Typedef using __uint8_t.
(u_int16_t): Typedef using __uint16_t.
(u_int32_t): Typedef using __uint32_t.
(u_int64_t): Typedef using __uint64_t.
2019-03-11 22:59:27 +08:00
|
|
|
|
|
|
|
#
|
|
|
|
# Base and generic classes for individual checks.
|
|
|
|
#
|
|
|
|
|
|
|
|
class ConstructChecker:
|
|
|
|
"""Scan a stream of C preprocessing tokens and possibly report
|
|
|
|
problems with them. The REPORTER object passed to __init__ has
|
|
|
|
one method, reporter.error(token, message), which should be
|
|
|
|
called to indicate a problem detected at the position of TOKEN.
|
|
|
|
If MESSAGE contains the four-character sequence '{!r}' then that
|
|
|
|
will be replaced with a textual representation of TOKEN.
|
|
|
|
"""
|
|
|
|
def __init__(self, reporter):
|
|
|
|
self.reporter = reporter
|
|
|
|
|
|
|
|
def examine(self, tok):
|
|
|
|
"""Called once for each token in a header file.
|
|
|
|
Call self.reporter.error if a problem is detected.
|
|
|
|
"""
|
|
|
|
raise NotImplementedError
|
|
|
|
|
|
|
|
def eof(self):
|
|
|
|
"""Called once at the end of the stream. Subclasses need only
|
|
|
|
override this if it might have something to do."""
|
|
|
|
pass
|
|
|
|
|
|
|
|
class NoCheck(ConstructChecker):
|
|
|
|
"""Generic checker class which doesn't do anything. Substitute this
|
|
|
|
class for a real checker when a particular check should be skipped
|
|
|
|
for some file."""
|
|
|
|
|
|
|
|
def examine(self, tok):
|
|
|
|
pass
|
|
|
|
|
|
|
|
#
|
|
|
|
# Check for obsolete type names.
|
|
|
|
#
|
|
|
|
|
|
|
|
# The obsolete type names we're looking for:
|
|
|
|
OBSOLETE_TYPE_RE_ = re.compile(r"""\A
|
|
|
|
(__)?
|
|
|
|
( quad_t
|
|
|
|
| u(?: short | int | long
|
|
|
|
| _(?: char | short | int(?:[0-9]+_t)? | long | quad_t )))
|
|
|
|
\Z""", re.VERBOSE)
|
|
|
|
|
|
|
|
class ObsoleteNotAllowed(ConstructChecker):
|
|
|
|
"""Don't allow any use of the obsolete typedefs."""
|
|
|
|
def examine(self, tok):
|
|
|
|
if OBSOLETE_TYPE_RE_.match(tok.text):
|
|
|
|
self.reporter.error(tok, "use of {!r}")
|
|
|
|
|
|
|
|
class ObsoletePrivateDefinitionsAllowed(ConstructChecker):
|
|
|
|
"""Allow definitions of the private versions of the
|
|
|
|
obsolete typedefs; that is, 'typedef [anything] __obsolete;'
|
|
|
|
"""
|
|
|
|
def __init__(self, reporter):
|
|
|
|
super().__init__(reporter)
|
|
|
|
self.in_typedef = False
|
|
|
|
self.prev_token = None
|
|
|
|
|
|
|
|
def examine(self, tok):
|
|
|
|
# bits/types.h hides 'typedef' in a macro sometimes.
|
|
|
|
if (tok.kind == "IDENT"
|
|
|
|
and tok.text in ("typedef", "__STD_TYPE")
|
|
|
|
and tok.context is None):
|
|
|
|
self.in_typedef = True
|
|
|
|
elif tok.kind == "PUNCTUATOR" and tok.text == ";" and self.in_typedef:
|
|
|
|
self.in_typedef = False
|
|
|
|
if self.prev_token.kind == "IDENT":
|
|
|
|
m = OBSOLETE_TYPE_RE_.match(self.prev_token.text)
|
|
|
|
if m and m.group(1) != "__":
|
|
|
|
self.reporter.error(self.prev_token, "use of {!r}")
|
|
|
|
self.prev_token = None
|
|
|
|
else:
|
|
|
|
self._check_prev()
|
|
|
|
|
|
|
|
self.prev_token = tok
|
|
|
|
|
|
|
|
def eof(self):
|
|
|
|
self._check_prev()
|
|
|
|
|
|
|
|
def _check_prev(self):
|
|
|
|
if (self.prev_token is not None
|
|
|
|
and self.prev_token.kind == "IDENT"
|
|
|
|
and OBSOLETE_TYPE_RE_.match(self.prev_token.text)):
|
|
|
|
self.reporter.error(self.prev_token, "use of {!r}")
|
|
|
|
|
|
|
|
class ObsoletePublicDefinitionsAllowed(ConstructChecker):
|
|
|
|
"""Allow definitions of the public versions of the obsolete
|
|
|
|
typedefs. Only specific forms of definition are allowed:
|
|
|
|
|
|
|
|
typedef __obsolete obsolete; // identifiers must agree
|
|
|
|
typedef __uintN_t u_intN_t; // N must agree
|
|
|
|
typedef unsigned long int ulong;
|
|
|
|
typedef unsigned short int ushort;
|
|
|
|
typedef unsigned int uint;
|
|
|
|
"""
|
|
|
|
def __init__(self, reporter):
|
|
|
|
super().__init__(reporter)
|
|
|
|
self.typedef_tokens = []
|
|
|
|
|
|
|
|
def examine(self, tok):
|
|
|
|
if tok.kind in ("WHITESPACE", "BLOCK_COMMENT",
|
|
|
|
"LINE_COMMENT", "NL", "ESCNL"):
|
|
|
|
pass
|
|
|
|
|
|
|
|
elif (tok.kind == "IDENT" and tok.text == "typedef"
|
|
|
|
and tok.context is None):
|
|
|
|
if self.typedef_tokens:
|
|
|
|
self.reporter.error(tok, "typedef inside typedef")
|
|
|
|
self._reset()
|
|
|
|
self.typedef_tokens.append(tok)
|
|
|
|
|
|
|
|
elif tok.kind == "PUNCTUATOR" and tok.text == ";":
|
|
|
|
self._finish()
|
|
|
|
|
|
|
|
elif self.typedef_tokens:
|
|
|
|
self.typedef_tokens.append(tok)
|
|
|
|
|
|
|
|
def eof(self):
|
|
|
|
self._reset()
|
|
|
|
|
|
|
|
def _reset(self):
|
|
|
|
while self.typedef_tokens:
|
|
|
|
tok = self.typedef_tokens.pop(0)
|
|
|
|
if tok.kind == "IDENT" and OBSOLETE_TYPE_RE_.match(tok.text):
|
|
|
|
self.reporter.error(tok, "use of {!r}")
|
|
|
|
|
|
|
|
def _finish(self):
|
|
|
|
if not self.typedef_tokens: return
|
|
|
|
if self.typedef_tokens[-1].kind == "IDENT":
|
|
|
|
m = OBSOLETE_TYPE_RE_.match(self.typedef_tokens[-1].text)
|
|
|
|
if m:
|
|
|
|
if self._permissible_public_definition(m):
|
|
|
|
self.typedef_tokens.clear()
|
|
|
|
self._reset()
|
|
|
|
|
|
|
|
def _permissible_public_definition(self, m):
|
|
|
|
if m.group(1) == "__": return False
|
|
|
|
name = m.group(2)
|
|
|
|
toks = self.typedef_tokens
|
|
|
|
ntok = len(toks)
|
|
|
|
if ntok == 3 and toks[1].kind == "IDENT":
|
|
|
|
defn = toks[1].text
|
|
|
|
n = OBSOLETE_TYPE_RE_.match(defn)
|
|
|
|
if n and n.group(1) == "__" and n.group(2) == name:
|
|
|
|
return True
|
|
|
|
|
|
|
|
if (name[:5] == "u_int" and name[-2:] == "_t"
|
|
|
|
and defn[:6] == "__uint" and defn[-2:] == "_t"
|
|
|
|
and name[5:-2] == defn[6:-2]):
|
|
|
|
return True
|
|
|
|
|
|
|
|
return False
|
|
|
|
|
|
|
|
if (name == "ulong" and ntok == 5
|
|
|
|
and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
|
|
|
|
and toks[2].kind == "IDENT" and toks[2].text == "long"
|
|
|
|
and toks[3].kind == "IDENT" and toks[3].text == "int"):
|
|
|
|
return True
|
|
|
|
|
|
|
|
if (name == "ushort" and ntok == 5
|
|
|
|
and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
|
|
|
|
and toks[2].kind == "IDENT" and toks[2].text == "short"
|
|
|
|
and toks[3].kind == "IDENT" and toks[3].text == "int"):
|
|
|
|
return True
|
|
|
|
|
|
|
|
if (name == "uint" and ntok == 4
|
|
|
|
and toks[1].kind == "IDENT" and toks[1].text == "unsigned"
|
|
|
|
and toks[2].kind == "IDENT" and toks[2].text == "int"):
|
|
|
|
return True
|
|
|
|
|
|
|
|
return False
|
|
|
|
|
|
|
|
def ObsoleteTypedefChecker(reporter, fname):
|
|
|
|
"""Factory: produce an instance of the appropriate
|
|
|
|
obsolete-typedef checker for FNAME."""
|
|
|
|
|
|
|
|
# The obsolete rpc/ and rpcsvc/ headers are allowed to use the
|
|
|
|
# obsolete types, because it would be more trouble than it's
|
|
|
|
# worth to remove them from headers that we intend to stop
|
|
|
|
# installing eventually anyway.
|
|
|
|
if (fname.startswith("rpc/")
|
|
|
|
or fname.startswith("rpcsvc/")
|
|
|
|
or "/rpc/" in fname
|
|
|
|
or "/rpcsvc/" in fname):
|
|
|
|
return NoCheck(reporter)
|
|
|
|
|
|
|
|
# bits/types.h is allowed to define the __-versions of the
|
|
|
|
# obsolete types.
|
|
|
|
if (fname == "bits/types.h"
|
|
|
|
or fname.endswith("/bits/types.h")):
|
|
|
|
return ObsoletePrivateDefinitionsAllowed(reporter)
|
|
|
|
|
|
|
|
# sys/types.h is allowed to use the __-versions of the
|
|
|
|
# obsolete types, but only to define the unprefixed versions.
|
|
|
|
if (fname == "sys/types.h"
|
|
|
|
or fname.endswith("/sys/types.h")):
|
|
|
|
return ObsoletePublicDefinitionsAllowed(reporter)
|
|
|
|
|
|
|
|
return ObsoleteNotAllowed(reporter)
|
|
|
|
|
|
|
|
#
|
|
|
|
# Master control
|
|
|
|
#
|
|
|
|
|
|
|
|
class HeaderChecker:
|
|
|
|
"""Perform all of the checks on each header. This is also the
|
|
|
|
"reporter" object expected by tokenize_c and ConstructChecker.
|
|
|
|
"""
|
|
|
|
def __init__(self):
|
|
|
|
self.fname = None
|
|
|
|
self.status = 0
|
|
|
|
|
|
|
|
def error(self, tok, message):
|
|
|
|
self.status = 1
|
|
|
|
if '{!r}' in message:
|
|
|
|
message = message.format(tok.text)
|
|
|
|
sys.stderr.write("{}:{}:{}: error: {}\n".format(
|
|
|
|
self.fname, tok.line, tok.column, message))
|
|
|
|
|
|
|
|
def check(self, fname):
|
|
|
|
self.fname = fname
|
|
|
|
try:
|
2019-03-14 21:44:22 +08:00
|
|
|
with open(fname, "rt", encoding="utf-8") as fp:
|
Use a proper C tokenizer to implement the obsolete typedefs test.
The test for obsolete typedefs in installed headers was implemented
using grep, and could therefore get false positives on e.g. “ulong”
in a comment. It was also scanning all of the headers included by
our headers, and therefore testing headers we don’t control, e.g.
Linux kernel headers.
This patch splits the obsolete-typedef test from
scripts/check-installed-headers.sh to a separate program,
scripts/check-obsolete-constructs.py. Being implemented in Python,
it is feasible to make it tokenize C accurately enough to avoid false
positives on the contents of comments and strings. It also only
examines $(headers) in each subdirectory--all the headers we install,
but not any external dependencies of those headers. Headers whose
installed name starts with finclude/ are ignored, on the assumption
that they contain Fortran.
It is also feasible to make the new test understand the difference
between _defining_ the obsolete typedefs and _using_ the obsolete
typedefs, which means posix/{bits,sys}/types.h no longer need to be
exempted. This uncovered an actual bug in bits/types.h: __quad_t and
__u_quad_t were being used to define __S64_TYPE, __U64_TYPE,
__SQUAD_TYPE and __UQUAD_TYPE. These are changed to __int64_t and
__uint64_t respectively. This is a safe change, despite the comments
in bits/types.h claiming a difference between __quad_t and __int64_t,
because those comments are incorrect. In all current ABIs, both
__quad_t and __int64_t are ‘long’ when ‘long’ is a 64-bit type, and
‘long long’ when ‘long’ is a 32-bit type, and similarly for __u_quad_t
and __uint64_t. (Changing the types to be what the comments say they
are would be an ABI break, as it affects C++ name mangling.) This
patch includes a minimal change to make the comments not completely
wrong.
sys/types.h was defining the legacy BSD u_intN_t typedefs using a
construct that was not necessarily consistent with how the C99 uintN_t
typedefs are defined, and is also too complicated for the new script to
understand (it lexes C relatively accurately, but it does not attempt
to expand preprocessor macros, nor does it do any actual parsing).
This patch cuts all of that out and uses bits/types.h's __uintN_t typedefs
to define u_intN_t instead. This is verified to not change the ABI on
any supported architecture, via the c++-types test, which means u_intN_t
and uintN_t were, in fact, consistent on all supported architectures.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
* scripts/check-obsolete-constructs.py: New test script.
* scripts/check-installed-headers.sh: Remove tests for
obsolete typedefs, superseded by check-obsolete-constructs.py.
* Rules: Run scripts/check-obsolete-constructs.py over $(headers)
as a special test. Update commentary.
* posix/bits/types.h (__SQUAD_TYPE, __S64_TYPE): Define as __int64_t.
(__UQUAD_TYPE, __U64_TYPE): Define as __uint64_t.
Update commentary.
* posix/sys/types.h (__u_intN_t): Remove.
(u_int8_t): Typedef using __uint8_t.
(u_int16_t): Typedef using __uint16_t.
(u_int32_t): Typedef using __uint32_t.
(u_int64_t): Typedef using __uint64_t.
2019-03-11 22:59:27 +08:00
|
|
|
contents = fp.read()
|
|
|
|
except OSError as e:
|
|
|
|
sys.stderr.write("{}: {}\n".format(fname, e.strerror))
|
|
|
|
self.status = 1
|
|
|
|
return
|
|
|
|
|
|
|
|
typedef_checker = ObsoleteTypedefChecker(self, self.fname)
|
|
|
|
|
2022-09-22 18:10:41 +08:00
|
|
|
for tok in glibcpp.tokenize_c(contents, self):
|
Use a proper C tokenizer to implement the obsolete typedefs test.
The test for obsolete typedefs in installed headers was implemented
using grep, and could therefore get false positives on e.g. “ulong”
in a comment. It was also scanning all of the headers included by
our headers, and therefore testing headers we don’t control, e.g.
Linux kernel headers.
This patch splits the obsolete-typedef test from
scripts/check-installed-headers.sh to a separate program,
scripts/check-obsolete-constructs.py. Being implemented in Python,
it is feasible to make it tokenize C accurately enough to avoid false
positives on the contents of comments and strings. It also only
examines $(headers) in each subdirectory--all the headers we install,
but not any external dependencies of those headers. Headers whose
installed name starts with finclude/ are ignored, on the assumption
that they contain Fortran.
It is also feasible to make the new test understand the difference
between _defining_ the obsolete typedefs and _using_ the obsolete
typedefs, which means posix/{bits,sys}/types.h no longer need to be
exempted. This uncovered an actual bug in bits/types.h: __quad_t and
__u_quad_t were being used to define __S64_TYPE, __U64_TYPE,
__SQUAD_TYPE and __UQUAD_TYPE. These are changed to __int64_t and
__uint64_t respectively. This is a safe change, despite the comments
in bits/types.h claiming a difference between __quad_t and __int64_t,
because those comments are incorrect. In all current ABIs, both
__quad_t and __int64_t are ‘long’ when ‘long’ is a 64-bit type, and
‘long long’ when ‘long’ is a 32-bit type, and similarly for __u_quad_t
and __uint64_t. (Changing the types to be what the comments say they
are would be an ABI break, as it affects C++ name mangling.) This
patch includes a minimal change to make the comments not completely
wrong.
sys/types.h was defining the legacy BSD u_intN_t typedefs using a
construct that was not necessarily consistent with how the C99 uintN_t
typedefs are defined, and is also too complicated for the new script to
understand (it lexes C relatively accurately, but it does not attempt
to expand preprocessor macros, nor does it do any actual parsing).
This patch cuts all of that out and uses bits/types.h's __uintN_t typedefs
to define u_intN_t instead. This is verified to not change the ABI on
any supported architecture, via the c++-types test, which means u_intN_t
and uintN_t were, in fact, consistent on all supported architectures.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
* scripts/check-obsolete-constructs.py: New test script.
* scripts/check-installed-headers.sh: Remove tests for
obsolete typedefs, superseded by check-obsolete-constructs.py.
* Rules: Run scripts/check-obsolete-constructs.py over $(headers)
as a special test. Update commentary.
* posix/bits/types.h (__SQUAD_TYPE, __S64_TYPE): Define as __int64_t.
(__UQUAD_TYPE, __U64_TYPE): Define as __uint64_t.
Update commentary.
* posix/sys/types.h (__u_intN_t): Remove.
(u_int8_t): Typedef using __uint8_t.
(u_int16_t): Typedef using __uint16_t.
(u_int32_t): Typedef using __uint32_t.
(u_int64_t): Typedef using __uint64_t.
2019-03-11 22:59:27 +08:00
|
|
|
typedef_checker.examine(tok)
|
|
|
|
|
|
|
|
def main():
|
|
|
|
ap = argparse.ArgumentParser(description=__doc__)
|
|
|
|
ap.add_argument("headers", metavar="header", nargs="+",
|
|
|
|
help="one or more headers to scan for obsolete constructs")
|
|
|
|
args = ap.parse_args()
|
|
|
|
|
|
|
|
checker = HeaderChecker()
|
|
|
|
for fname in args.headers:
|
|
|
|
# Headers whose installed name begins with "finclude/" contain
|
|
|
|
# Fortran, not C, and this program should completely ignore them.
|
|
|
|
if not (fname.startswith("finclude/") or "/finclude/" in fname):
|
|
|
|
checker.check(fname)
|
|
|
|
sys.exit(checker.status)
|
|
|
|
|
|
|
|
main()
|