[Bug c++/64867] split warning for passing non-POD to varargs function from -Wconditionally-supported into new warning flag, -Wnon-pod-varargs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64867 --- Comment #32 from CVS Commits --- The master branch has been updated by Jason Merrill : https://gcc.gnu.org/g:263c22a95bc9a0d80c4873c0291b0f938cea7310 commit r13-4795-g263c22a95bc9a0d80c4873c0291b0f938cea7310 Author: Jason Merrill Date: Mon Dec 5 10:00:31 2022 -0500 build: add -Wconditionally-supported to strict_warn [PR64867] The PR (which isn't resolved by this commit) pointed out to me that GCC should build with -Wconditionally-supported to support bootstrapping with a C++11 compiler that makes different choices. PR c++/64867 gcc/ChangeLog: * configure.ac (strict_warn): Add -Wconditionally-supported. * configure: Regenerate.
[Bug c++/64867] split warning for passing non-POD to varargs function from -Wconditionally-supported into new warning flag, -Wnon-pod-varargs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64867 Jason Merrill changed: What|Removed |Added CC||jason at gcc dot gnu.org --- Comment #31 from Jason Merrill --- (In reply to David Binderman from comment #30) > gcc looks to be a trailer on this issue. It's not about standards conformance, > its about making it easy for users to find bugs. But under GCC it's not a bug, it just works. Clang is welcome to adopt the same semantics. But I'm also happy to accept the patch in comment #17 once it includes testing. And GCC should build with -Wconditionally-supported so we support bootstrapping with compilers that make different choices.
[Bug c++/64867] split warning for passing non-POD to varargs function from -Wconditionally-supported into new warning flag, -Wnon-pod-varargs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64867
David Binderman changed:
What|Removed |Added
CC||dcb314 at hotmail dot com
--- Comment #30 from David Binderman ---
(In reply to Trass3r from comment #27)
> This should really be enabled by -Wall or -Wextra as it generates code that
> may easily crash or corrupt something.
+1.
Clang errors, intel warns and for gcc you have to enable the somewhat
obscure flag -Wconditionally-supported to see anything at all.
gcc looks to be a trailer on this issue. It's not about standards conformance,
its about making it easy for users to find bugs.
Source code I used is:
struct S
{
S();
S( const S &);
~S();
char * p;
};
void f( int, ...);
void g()
{
S s1;
f( 3, s1);
};
$ /home/dcb/llvm/results/bin/clang++ sep3f.cc
sep3f.cc:19:8: error: cannot pass object of non-trivial type 'S' through
variadic function; call will abort at runtime [-Wnon-pod-varargs]
f( 3, s1);
^
1 error generated.
$ /home/dcb34/intel/oneapi/compiler/2021.3.0/linux/bin/intel64/icpc sep3f.cc
sep3f.cc(19): warning #1595: a class type that is not trivially copyable passed
through ellipsis
f( 3, s1);
^
$ /home/dcb/gcc/results/bin/g++ -c -g -O2 -Wall -Wextra sep3f.cc
$ /home/dcb/gcc/results/bin/g++ -c -g -O2 -Wall -Wextra
-Wconditionally-supported sep3f.cc
sep3f.cc: In function ‘void g()’:
sep3f.cc:19:10: warning: passing objects of non-trivially-copyable type ‘struct
S’ through ‘...’ is conditionally supported [-Wconditionally-supported]
19 | f( 3, s1);
| ~^~~~
I'll add flag -Wconditionally-supported to a build of Fedora and see
what happens.
[Bug c++/64867] split warning for passing non-POD to varargs function from -Wconditionally-supported into new warning flag, -Wnon-pod-varargs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64867 --- Comment #29 from Trass3r --- Ah I see, ok.
[Bug c++/64867] split warning for passing non-POD to varargs function from -Wconditionally-supported into new warning flag, -Wnon-pod-varargs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64867 --- Comment #28 from Andrew Pinski --- (In reply to Trass3r from comment #27) > This should really be enabled by -Wall or -Wextra as it generates code that > may easily crash or corrupt something. But GCC supports this fully. That is why it is/was part of -Wconditionally-supported . The code if used correctly does not cause crashes or corrupts anything. In fact GCC documents this behavior: https://gcc.gnu.org/onlinedocs/gcc-9.2.0/gcc/Conditionally-supported-behavior.html#Conditionally-supported-behavior :) > Clang even turns it into an error by default and inserts a trap to abort at > runtime: https://godbolt.org/z/8DAUAr But clang does not support this. Both compilers are correct; in that this feature is considered as conditionally supported with implementation defined behavior. So both compilers are prefectly ok in having their own choices.
[Bug c++/64867] split warning for passing non-POD to varargs function from -Wconditionally-supported into new warning flag, -Wnon-pod-varargs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64867 Trass3r changed: What|Removed |Added CC||trass3r at gmail dot com --- Comment #27 from Trass3r --- This should really be enabled by -Wall or -Wextra as it generates code that may easily crash or corrupt something. Clang even turns it into an error by default and inserts a trap to abort at runtime: https://godbolt.org/z/8DAUAr
[Bug c++/64867] split warning for passing non-POD to varargs function from -Wconditionally-supported into new warning flag, -Wnon-pod-varargs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64867 Eric Gallager changed: What|Removed |Added CC||msebor at gcc dot gnu.org --- Comment #26 from Eric Gallager --- Martin Sebor has been doing stuff related to warnings about POD-ness lately; cc-ing him
[Bug c++/64867] split warning for passing non-POD to varargs function from -Wconditionally-supported into new warning flag, -Wnon-pod-varargs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64867 Eric Gallager changed: What|Removed |Added Blocks||87403 Summary|warning for passing non-POD |split warning for passing |to varargs function |non-POD to varargs function ||from ||-Wconditionally-supported ||into new warning flag, ||-Wnon-pod-varargs --- Comment #25 from Eric Gallager --- retitling Referenced Bugs: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87403 [Bug 87403] [Meta-bug] Issues that suggest a new warning
