* [PATCH] libstdc++: Work around clang misdesign in time_get<>::get [PR104990]
@ 2022-03-21 6:42 Jakub Jelinek
2022-03-21 10:00 ` Jonathan Wakely
0 siblings, 1 reply; 2+ messages in thread
From: Jakub Jelinek @ 2022-03-21 6:42 UTC (permalink / raw)
To: Jonathan Wakely; +Cc: gcc-patches, libstdc++
Hi!
Apparently clang has a -fgnuc-version= option which allows it to pretend
it is any GCC version the user likes. It is already bad that it claims to
be GCC 4.2 compatible by default when it is not (various unimplemented
extensions at least), but this option is a horrible idea.
Anyway, this patch adds a hack for it.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2022-03-21 Jakub Jelinek <jakub@redhat.com>
PR libstdc++/104990
* include/bits/locale_facets_nonio.tcc (get): Don't check if do_get
isn't overloaded if __clang__ is defined.
--- libstdc++-v3/include/bits/locale_facets_nonio.tcc 2022-03-18 10:37:41.176593188 +0100
+++ libstdc++-v3/include/bits/locale_facets_nonio.tcc 2022-03-20 20:28:07.203815325 +0100
@@ -1465,7 +1465,7 @@ _GLIBCXX_END_NAMESPACE_LDBL_OR_CXX11
ctype<_CharT> const& __ctype = use_facet<ctype<_CharT> >(__loc);
__err = ios_base::goodbit;
bool __use_state = false;
-#if __GNUC__ >= 5
+#if __GNUC__ >= 5 && !defined(__clang__)
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wpmf-conversions"
// Nasty hack. The C++ standard mandates that get invokes the do_get
Jakub
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: [PATCH] libstdc++: Work around clang misdesign in time_get<>::get [PR104990]
2022-03-21 6:42 [PATCH] libstdc++: Work around clang misdesign in time_get<>::get [PR104990] Jakub Jelinek
@ 2022-03-21 10:00 ` Jonathan Wakely
0 siblings, 0 replies; 2+ messages in thread
From: Jonathan Wakely @ 2022-03-21 10:00 UTC (permalink / raw)
To: Jakub Jelinek; +Cc: gcc Patches, libstdc++
On Mon, 21 Mar 2022 at 06:42, Jakub Jelinek wrote:
>
> Hi!
>
> Apparently clang has a -fgnuc-version= option which allows it to pretend
> it is any GCC version the user likes. It is already bad that it claims to
> be GCC 4.2 compatible by default when it is not (various unimplemented
> extensions at least), but this option is a horrible idea.
>
> Anyway, this patch adds a hack for it.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
OK for trunk.
In stage 1 we might want to consider removing the __GNUC__ check
entirely. We don't support using old versions of genuine GCC. Intel
icc claims to be the latest GCC. And now Clang can be told to cosplay
as any version of GCC, so we can't trust __GNUC__ at all. So this
check would be just !defined(__clang__).
>
> 2022-03-21 Jakub Jelinek <jakub@redhat.com>
>
> PR libstdc++/104990
> * include/bits/locale_facets_nonio.tcc (get): Don't check if do_get
> isn't overloaded if __clang__ is defined.
>
> --- libstdc++-v3/include/bits/locale_facets_nonio.tcc 2022-03-18 10:37:41.176593188 +0100
> +++ libstdc++-v3/include/bits/locale_facets_nonio.tcc 2022-03-20 20:28:07.203815325 +0100
> @@ -1465,7 +1465,7 @@ _GLIBCXX_END_NAMESPACE_LDBL_OR_CXX11
> ctype<_CharT> const& __ctype = use_facet<ctype<_CharT> >(__loc);
> __err = ios_base::goodbit;
> bool __use_state = false;
> -#if __GNUC__ >= 5
> +#if __GNUC__ >= 5 && !defined(__clang__)
> #pragma GCC diagnostic push
> #pragma GCC diagnostic ignored "-Wpmf-conversions"
> // Nasty hack. The C++ standard mandates that get invokes the do_get
>
> Jakub
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2022-03-21 10:00 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-03-21 6:42 [PATCH] libstdc++: Work around clang misdesign in time_get<>::get [PR104990] Jakub Jelinek
2022-03-21 10:00 ` Jonathan Wakely
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).