Bug#748616: gcc-4.8: [armel] float to double conversion does not silence sNaN

2014-05-18 Thread Aurelien Jarno
On Sun, May 18, 2014 at 11:35:04PM +0200, Matthias Klose wrote: > Am 18.05.2014 23:30, schrieb Aurelien Jarno: > > Package: gcc-4.8 > > Version: 4.8.2-21 > > please check with gcc-4.7, gcc-4.9 and gcc-snapshot too. I reported it against gcc 4.8 because it is the default version, but all versions

Bug#748616: gcc-4.8: [armel] float to double conversion does not silence sNaN

2014-05-18 Thread Matthias Klose
Am 18.05.2014 23:30, schrieb Aurelien Jarno: > Package: gcc-4.8 > Version: 4.8.2-21 please check with gcc-4.7, gcc-4.9 and gcc-snapshot too. -- To UNSUBSCRIBE, email to debian-gcc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: htt

Bug#748616: gcc-4.8: [armel] float to double conversion does not silence sNaN

2014-05-18 Thread Aurelien Jarno
Package: gcc-4.8 Version: 4.8.2-21 Severity: normal Consider the following code: #define _GNU_SOURCE #include #include int main (void) { float sNaN = __builtin_nansf (""); double x = (double) sNaN; return issignaling(x); } It returns 1 on armel, but 0 on armel or other architectures. Qu