Re: [PATCHv2] PR libstdc++/84654 Do not use __float128 if it is disabled by the compiler

[Jonathan Wakely <jwakely@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1244 bytes --]

On 01/05/18 15:34 +0100, Jonathan Wakely wrote:
>On 12/03/18 16:02 -0300, Tulio Magno Quites Machado Filho wrote:
>>Changes since v1:
>>- Completely rewrite of the patch to set ENABLE_FLOAT128 at libstdc++
>>  build time and undef _GLIBCXX_USE_FLOAT128 when building user code.
>>
>>--- 8< ---
>>
>>In order to use __float128 in C++ it's necessary to check if the
>>compiler enabled its support too when building user code.
>>This patch changes the behavior at libstdc++ build by setting
>>ENABLE_FLOAT128, which is used to set the value of the exported macro
>>_GLIBCXX_USE_FLOAT128.
>
>Now that we're in stage 1 we can make this change. The patch looks
>right so I'll go ahead and commit it to trunk (it shouldn't interfere
>with testing any last-minute fixes needed for gcc-8).
>
>Thanks for fixing this, Tulio.

Previously _GLIBCXX_USE_FLOAT128 would either be defined to 1 or
undefined, but after this change it can be defined to 0, and the
checks in <type_traits> and <bits/std_abs.h> were not adjusted to
account for that.

This patch restores the previous behaviour, so that when __float128
support is not found by configure the macro will be undefined instead
of defined to 0.

I plan to commit this tomorrow after some more testing.


[-- Attachment #2: patch.txt --]
[-- Type: text/x-patch, Size: 2471 bytes --]

commit aed5af13903fa21632919269c9e7f7c4e02e8887
Author: Jonathan Wakely <jwakely@redhat.com>
Date:   Mon May 7 20:46:21 2018 +0100

    PR libstdc++/85672 #undef _GLIBCXX_USE_FLOAT128 when not supported
    
    Restore the behaviour in GCC 8 and earlier where _GLIBCXX_USE_FLOAT128
    is not defined when configure detects support is missing. This avoids
    having three states where the macro is either 1, 0, or undefined.
    
            PR libstdc++/85672
            * include/Makefile.am [!ENABLE_FLOAT128]: Change c++config.h entry
            to #undef _GLIBCXX_USE_FLOAT128 instead of defining it to zero.
            * include/Makefile.in: Regenerate.
            * include/bits/c++config (_GLIBCXX_USE_FLOAT128): Move definition
            within conditional block.

diff --git a/libstdc++-v3/include/Makefile.am b/libstdc++-v3/include/Makefile.am
index ba5adda95bd..77e6dc2f6be 100644
--- a/libstdc++-v3/include/Makefile.am
+++ b/libstdc++-v3/include/Makefile.am
@@ -1232,10 +1232,10 @@ endif
 
 if ENABLE_FLOAT128
 stamp-float128:
-	echo 1 > stamp-float128
+	echo 'define _GLIBCXX_USE_FLOAT128 1' > stamp-float128
 else
 stamp-float128:
-	echo 0 > stamp-float128
+	echo 'undef _GLIBCXX_USE_FLOAT128' > stamp-float128
 endif
 
 # NB: The non-empty default ldbl_compat works around an AIX sed
@@ -1272,7 +1272,7 @@ ${host_builddir}/c++config.h: ${CONFIG_HEADER} \
 	-e "s,define _GLIBCXX_USE_DUAL_ABI, define _GLIBCXX_USE_DUAL_ABI $$dualabi," \
 	-e "s,define _GLIBCXX_USE_CXX11_ABI, define _GLIBCXX_USE_CXX11_ABI $$cxx11abi," \
 	-e "s,define _GLIBCXX_USE_ALLOCATOR_NEW, define _GLIBCXX_USE_ALLOCATOR_NEW $$allocatornew," \
-	-e "s,define _GLIBCXX_USE_FLOAT128, define _GLIBCXX_USE_FLOAT128 $$float128," \
+	-e "s,define _GLIBCXX_USE_FLOAT128,$$float128," \
 	-e "$$ldbl_compat" \
 	    < ${glibcxx_srcdir}/include/bits/c++config > $@ ;\
 	sed -e 's/HAVE_/_GLIBCXX_HAVE_/g' \
diff --git a/libstdc++-v3/include/bits/c++config b/libstdc++-v3/include/bits/c++config
index e34524117d6..bfe268da825 100644
--- a/libstdc++-v3/include/bits/c++config
+++ b/libstdc++-v3/include/bits/c++config
@@ -610,9 +610,8 @@ namespace std
 #endif
 
 /* Define if __float128 is supported on this host. */
+#if defined(__FLOAT128__) || defined(__SIZEOF_FLOAT128__)
 #define _GLIBCXX_USE_FLOAT128
-#if !defined(__FLOAT128__) && !defined(__SIZEOF_FLOAT128__)
-#undef _GLIBCXX_USE_FLOAT128
 #endif
 
 // End of prewritten config; the settings discovered at configure time follow.