[Bug c++/78160] New: explicit template instantation with hidden symbols fails
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78160
Bug ID: 78160
Summary: explicit template instantation with hidden symbols
fails
Product: gcc
Version: 6.2.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: gcc at ebasoft dot com.pl
Target Milestone: ---
Created attachment 39922
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=39922&action=edit
full source code with CMake configuration
When some template struct is specialized in shared library on other type that
has hidden visibility it passes compilation but no code is generated.
This causes missing symbols in library. If base type is required to be default
visible it should fail at compilation stage with error instead of causing hard
to understand source of problems.
I reproducted this error on gcc 5.4.1, 6.2.0
//Example code (full cmake project attached)
///LIBRARY HEADER some_shared.h
//--
#include
#include
#define visibility_hidden __attribute__((visibility("hidden")))
#define visibility_default __attribute__((visibility("default")))
struct visibility_default base_spec_char
{
using value_type = char;
char foo[10];
};
//SPECIALIZING ON THIS TYPE WILL CAUSE MISSING CODE AND SYMBOLS
struct visibility_hidden base_spec_uint8
{
using value_type = uint8_t;
uint8_t foo[20];
};
template
struct visibility_default var_t
{
using var_store_t = VarStoreType;
using value_type = typename var_store_t::value_type;
std::unique_ptr alloc();
};
using varchar_t = var_t;
using varbinary_t = var_t;
///LIBRARY SOURCE
//--
#include "some_shared.h"
template
std::unique_ptr::var_store_t>
var_t::alloc( )
{
return std::make_unique();
}
template struct var_t;
template struct var_t;
//MAIN PROGRAM LINKED TO SHARED LIBRARY
#include
#include "some_shared.h"
int main(int argc, char **argv)
{
varchar_t foo;
varbinary_t bar;
auto obj = foo.alloc();
auto obj2 = bar.alloc(); //<- unresolved symbols
std::cout << "Hello, world!"<< obj.get()<< obj2.get() << std::endl;
return 0;
}
main.cpp:10: undefined reference to `var_t::alloc()'
[Bug c++/102286] [constexpr] construct_at incorrectly starts union array lifetime in some cases
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102286
Artur Bać changed:
What|Removed |Added
CC||gcc at ebasoft dot com.pl
--- Comment #4 from Artur Bać ---
This is not legal in c++20 but gcc allows accessing not active member of a
union in constexpr
https://godbolt.org/z/KsEhffeEa
clang refuses and it is right
"construction of subobject of member 'object' of union with active member
'init' is not allowed in a constant expression"
With c++20 there is no way to have aligned_storage for not trivial type in
constexpr (storage without initialization, not std::array that requires
construction of non trivial objects upon member activation)
https://godbolt.org/z/9EGrf8fbr
#include
struct foo
{
int * i;
constexpr foo() { i = new int; }
constexpr ~foo() { delete(i); }
};
union storage
{
constexpr storage()
: init{}
{}
constexpr ~storage() {}
foo object[1];
char init = '\0';
};
consteval bool test()
{
storage u;
auto p = std::addressof(u.object[0]);
std::construct_at(p);
std::destroy_at(p);
return true;
}
int main()
{
static_assert( test() );
return test();
}
[Bug c++/100983] Deduction guide for member template class rejected at class scope
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100983 Artur Bać changed: What|Removed |Added CC||gcc at ebasoft dot com.pl --- Comment #4 from Artur Bać --- trunk at compiler explorer still rejects valid code https://godbolt.org/z/v4ebhj9Gh, only the message of requirement of namespace scope is missing from gcc 11.2, invalid use of template-name without an argument list https://godbolt.org/z/7Wev6saWz "ctad" must be declared at namespace scope + invalid use of template-name without an argument list clang https://godbolt.org/z/vavPTbf36 works as expected
[Bug c++/104636] New: implicit use of explicit constructor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104636
Bug ID: 104636
Summary: implicit use of explicit constructor
Product: gcc
Version: 11.2.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: gcc at ebasoft dot com.pl
Target Milestone: ---
during build of kde-misc/kdeconnect with clang and then gcc I found that gcc
allows implicit use of explicit ctor mentioning even that it would use
explicit, it sounds like it is compiling this as c++98 or so ..
https://godbolt.org/z/WsWPbbKfe
:14:13: warning: converting to 'foo' from initializer list would use
explicit constructor 'foo::foo(const char*, uint8_t)'
14 | return {};
| ^
:14:13: note: in C++11 and above a default constructor can be explicit
similar bug was resolved and fixed
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=40685
[Bug c++/100983] Deduction guide for member template class rejected at class scope
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100983 --- Comment #6 from Artur Bać --- The typename was from my real code by mistake where value_type s template param. But in real code withing template I have to use typename and it doesn't work with trunk too. https://godbolt.org/z/E6Pavhfza
[Bug c++/100983] Deduction guide for member template class rejected at class scope
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100983 --- Comment #7 from Artur Bać --- Do I have to open new bug because of You marked it as fixed while it is not fixed ?
[Bug c++/104641] New: Deduction guide for member template class rejected at class scope when used with typename dependant type
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104641 Bug ID: 104641 Summary: Deduction guide for member template class rejected at class scope when used with typename dependant type Product: gcc Version: 12.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: c++ Assignee: unassigned at gcc dot gnu.org Reporter: gcc at ebasoft dot com.pl Target Milestone: --- Follow up of https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100983 Doesn't work when used with template typename inside template https://godbolt.org/z/E6Pavhfza
[Bug c++/100983] Deduction guide for member template class rejected at class scope
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100983 --- Comment #9 from Artur Bać --- created https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104641
[Bug c++/104636] implicit use of explicit constructor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104636 --- Comment #4 from Artur Bać --- So conclusion from Your arguments is to me that -std=c++20 doesn't guarantee strict c++20 compatibility as another switches are required to force explicit constructor be used only explicitly and disallow implicit construction with explicit constructor.
[Bug c++/58855] Attributes ignored on type alias in template
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58855
gcc at ebasoft dot com.pl changed:
What|Removed |Added
CC||gcc at ebasoft dot com.pl
--- Comment #3 from gcc at ebasoft dot com.pl ---
This bug is still present in gcc version 11.2.1 20220115
namespace concepts
{
template
concept arithmetic = std::integral || std::floating_point;
}
template
using point_vector = T __attribute__((__vector_size__(sizeof(T)*2)));
warning: ignoring attributes applied to dependent type ‘T’ without an
associated declaration [-Wattributes]
70 | using point_vector = T
__attribute__((__vector_size__(sizeof(T)*2)));
|
^
[Bug c++/104386] New: no_unique_address causes invalid member alignment of pod struct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104386
Bug ID: 104386
Summary: no_unique_address causes invalid member alignment of
pod struct
Product: gcc
Version: 11.2.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: gcc at ebasoft dot com.pl
Target Milestone: ---
struct base
{ [[no_unique_address]]
uint32_t x; std::byte v;
};
struct foo : public base
{ std::byte z; };
gcc sizeof(foo) == 8 https://godbolt.org/z/G4Mo3PdKT
clang sizeof(foo) == 12 https://godbolt.org/z/bdzvaMn9c
As I was told it is gcc bug
https://www.reddit.com/r/cpp/comments/sjx2mk/no_unique_addres_where_can_in_c_standard_instead/
[orbital1337]"Your class base is a POD but its not a POD for the purpose of
layout (since it has potentially overlapping data members). Thus the Itanium
ABI specifies that it's size without padding (dsize) should be set to its size
with padding (sizeof). The first datamember of your class foo is put at
dsize(base). clang does the right thing and puts it at an offset of 8 bytes
whereas gcc ignores that one line of the specification and puts it at an offset
of 5 bytes instead."
[Bug c++/104386] no_unique_address causes invalid member alignment of pod struct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104386
--- Comment #2 from Artur Bać ---
I think it is connected with this case
https://godbolt.org/z/cvfs1KqGW
https://godbolt.org/z/vdzTzo7be
//sizeof(foo) ==12
struct base
{
uint32_t x{};
std::byte v{};
base() noexcept = default;
};
struct foo : public base
{
std::byte z;
};
//sizeof(foo) == 8
struct base
{
uint32_t x;
std::byte v;
base() noexcept : x{}, v{} {}
};
struct foo : public base
{
std::byte z;
};
and
https://www.reddit.com/r/cpp/comments/slfugx/comment/hvs4n7j/?utm_source=share&utm_medium=web2x&context=3
"Som1Lse · 5 hr. ago
Interestingly GCC disagrees with itself:
-std=c++17 it considers base to be a POD and doesn't reuse padding.
-std=c++20 it considers base to not be a POD and agrees with Clang, reusing
the padding bytes."
[Bug libstdc++/102358] niter_base and miter_base overloaded for move_iterator missing constexpr
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102358 Artur Bać changed: What|Removed |Added CC||gcc at ebasoft dot com.pl --- Comment #2 from Artur Bać --- I have spend 20min figuring out why std::copy fails compile at constexpr with libstdc++ only ;-) and it is already reported. "* include/bits/stl_iterator.h (__niter_base): Make constexpr for C++20." c++20 ? according to cppreference make_move_iterator is constexpr since c++17 https://en.cppreference.com/w/cpp/iterator/make_move_iterator
[Bug libstdc++/102358] niter_base and miter_base overloaded for move_iterator missing constexpr
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102358 --- Comment #4 from Artur Bać --- In 12 ? In 11.2.1/include/g++-v11/bits/cpp_type_traits.h template _GLIBCXX20_CONSTEXPR inline _Iterator __miter_base(_Iterator __it) and as it was mentioned overload in stl_iterator.h has missing constexpr at all.
[Bug libstdc++/102358] niter_base and miter_base overloaded for move_iterator missing constexpr
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102358 --- Comment #6 from Artur Bać --- My bad. copy itself is constexpr since c++20, I was misleaded by overload used for move_iterator with copy that has missing constexpr template auto __miter_base(move_iterator<_Iterator> __it)
[Bug c++/115335] New: std::span at method is missing feature test macro __cpp_lib_span >= 202311L
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115335 Bug ID: 115335 Summary: std::span at method is missing feature test macro __cpp_lib_span >= 202311L Product: gcc Version: 14.1.1 Status: UNCONFIRMED Severity: normal Priority: P3 Component: c++ Assignee: unassigned at gcc dot gnu.org Reporter: gcc at ebasoft dot com.pl Target Milestone: --- gcc (Gentoo 14.1.1_p20240518 p1) 14.1.1 20240516 In gcc stdlibc++ 14 there is method span<>::at added that should be visible only to standard c++26. https://en.cppreference.com/w/cpp/container/span/at But implementation is missing feature test macro __cpp_lib_span >= 202311L and this method is visible for all standards backward in c++20 and c++23. availability in backward standards makes difficult to refactor code for example from vector to span that will be still portable with older gcc like 13 when it is being done on platform with gcc libstdc++ 14 as there is no compiler error when method at is being used after refactoring vector to span.
