diff mbox

Fix the C++ version of issignaling when __NO_LONG_DOUBLE_MATH is defined

Message ID 20170823152817.6184-1-gftg@linux.vnet.ibm.com
State Committed
Delegated to: Joseph Myers
Headers

Commit Message

Gabriel F T Gomes Aug. 23, 2017, 3:28 p.m. UTC 
  On Wed, 23 Aug 2017 00:41:23 +0000
Joseph Myers <joseph@codesourcery.com> wrote:

>This patch breaks building tests for long double = double configurations.
>__issignalingl does not exist in that case; the long double version of
>issignaling has to call __issignaling if __NO_LONG_DOUBLE_MATH is defined.

The following patch should fix this, however I do not have access to a
system where __NO_LONG_DOUBLE_MATH is defined.  I'm building this with
build-many-glibcs.py, but that will take some time.

OK for master?

-- 8< --
When __NO_LONG_DOUBLE_MATH is defined, __issignalingl is not available,
thus issignaling with long double argument should call __issignaling,
instead.

Tested for powerpc64le.

	* math/math.h [defined __cplusplus] (issignaling): In the long
	double case, call __issignalingl only if __NO_LONG_DOUBLE_MATH
	is not defined.  Call __issignaling, otherwise.
---
 math/math.h | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

Comments

Joseph Myers Aug. 24, 2017, 12:16 p.m. UTC | #1 
On Wed, 23 Aug 2017, Gabriel F. T. Gomes wrote:

> On Wed, 23 Aug 2017 00:41:23 +0000
> Joseph Myers <joseph@codesourcery.com> wrote:
> 
> >This patch breaks building tests for long double = double configurations.
> >__issignalingl does not exist in that case; the long double version of
> >issignaling has to call __issignaling if __NO_LONG_DOUBLE_MATH is defined.
> 
> The following patch should fix this, however I do not have access to a
> system where __NO_LONG_DOUBLE_MATH is defined.  I'm building this with
> build-many-glibcs.py, but that will take some time.
> 
> OK for master?

OK.
Gabriel F T Gomes Aug. 24, 2017, 7:18 p.m. UTC | #2 
On Thu, 24 Aug 2017 12:16:40 +0000
Joseph Myers <joseph@codesourcery.com> wrote:

>On Wed, 23 Aug 2017, Gabriel F. T. Gomes wrote:
>
>> The following patch should fix this, however I do not have access to a
>> system where __NO_LONG_DOUBLE_MATH is defined.  I'm building this with
>> build-many-glibcs.py, but that will take some time.
>> 
>> OK for master?  
>
>OK.

Thanks.  Pushed as 3d7b66f66cb2.

Running this with build-many-glibcs.py fixed the problem for the
mipsel-linux-gnu-soft configuration (which failed without the patch).

The other configurations are still running.  If other errors show up, I'll
come back to this.
diff mbox

Patch

diff --git a/math/math.h b/math/math.h
index b5d6c43fcf..5acbe88906 100644
--- a/math/math.h
+++ b/math/math.h
@@ -486,7 +486,15 @@  enum
 extern "C++" {
 inline int issignaling (float __val) { return __issignalingf (__val); }
 inline int issignaling (double __val) { return __issignaling (__val); }
-inline int issignaling (long double __val) { return __issignalingl (__val); }
+inline int
+issignaling (long double __val)
+{
+#  ifdef __NO_LONG_DOUBLE_MATH
+  return __issignaling (__val);
+#  else
+  return __issignalingl (__val);
+#  endif
+}
 #  if __HAVE_DISTINCT_FLOAT128
 inline int issignaling (_Float128 __val) { return __issignalingf128 (__val); }
 #  endif