Robert P. J. DAY says:
apparently, the patchset for gcc 4.2.1 applies properly to the
source for gcc 4.2.2 and gcc 4.2.3. so, if you want, you can simply
add support for those last two just by augmenting menuconfig and
adding a couple symlinks for those two directories. seems like a
cheap way to add a couple new versions.
1 Workaround for buglet in std::vector etc. when compiling
2 with gcc-4.0.1 -Wall -O -fno-exceptions
5 .../include/c++/4.0.0/bits/vector.tcc: In member function 'void std::vector<_Tp,
6 _Alloc>::reserve(size_t) [with _Tp = int, _Alloc = std::allocator<int>]':
7 .../include/c++/4.0.0/bits/vector.tcc:78: warning: control may reach end of
8 non-void function 'typename _Alloc::pointer std::vector<_Tp,
9 _Alloc>::_M_allocate_and_copy(size_t, _ForwardIterator, _ForwardIterator) [with
10 _ForwardIterator = int*, _Tp = int, _Alloc = std::allocator<int>]' being inlined
12 See http://gcc.gnu.org/PR21951
14 To: gcc-patches at gcc dot gnu dot org
15 Subject: [4.0.x] may reach end warning in system headers
16 Message-Id: <20050701183024.E138714C16A9@geoffk5.apple.com>
17 Date: Fri, 1 Jul 2005 11:30:24 -0700 (PDT)
18 From: gkeating at apple dot com (Geoffrey Keating)
21 One of our users was getting
23 /usr/include/gcc/darwin/4.0/c++/bits/stl_uninitialized.h:113: warning:
24 control may reach end of non-void function '_ForwardIterator
25 std::__uninitialized_copy_aux(_InputIterator, _InputIterator,
26 _ForwardIterator, __false_type) [with _InputIterator =
27 __gnu_cxx::__normal_iterator<TPoolAllocator::tAllocState*,
28 std::vector<TPoolAllocator::tAllocState,
29 std::allocator<TPoolAllocator::tAllocState> > >, _ForwardIterator =
30 __gnu_cxx::__normal_iterator<TPoolAllocator::tAllocState*,
31 std::vector<TPoolAllocator::tAllocState,
32 std::allocator<TPoolAllocator::tAllocState> > >]' being inlined
34 which shouldn't be happening, he has no way to change a standard C++
35 header. The warning is bogus anyway, but it's fixed in 4.1 through
36 the CFG changes, which I don't really want to backport to the 4.0
37 branch, so instead I'll add this patch. Other warnings generated from
38 tree-inline.c check for DECL_SYSTEM_HEADER like this.
40 Bootstrapped & tested on powerpc-darwin8, I'll commit when the branch
44 - Geoffrey Keating <geoffk@apple.com>
46 ===File ~/patches/gcc-40-4121982.patch======================
48 2005-06-28 Geoffrey Keating <geoffk@apple.com>
50 * tree-inline.c (expand_call_inline): Prevent 'may reach end'
51 warning in system headers.
54 ===================================================================
55 RCS file: /cvs/gcc/gcc/gcc/tree-inline.c,v
56 retrieving revision 1.170.8.4
57 diff -u -p -u -p -r1.170.8.4 tree-inline.c
58 --- gcc-4.0.1/gcc/tree-inline.c.old 6 Jun 2005 19:20:32 -0000 1.170.8.4
59 +++ gcc-4.0.1/gcc/tree-inline.c 1 Jul 2005 18:27:26 -0000
60 @@ -1693,7 +1693,8 @@ expand_call_inline (tree *tp, int *walk_
61 && !TREE_NO_WARNING (fn)
62 && !VOID_TYPE_P (TREE_TYPE (TREE_TYPE (fn)))
63 && return_slot_addr == NULL_TREE
64 - && block_may_fallthru (copy))
65 + && block_may_fallthru (copy)
66 + && !DECL_IN_SYSTEM_HEADER (fn))
68 warning ("control may reach end of non-void function %qD being inlined",
70 ============================================================