Horton

Re: Bug fix for glibc broke freebsd build in REL_11_STABLE

Andrew Gierth <andrew@tao11.riddles.org.uk>

View thread View on postgr.es
From: Andrew Gierth <andrew@tao11.riddles.org.uk>
To: Andres Freund <andres@anarazel.de>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Victor Wagner <vitus@wagner.pp.ru>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2018-09-04T20:59:59Z
Lists: pgsql-hackers 
>>>>> "Andres" == Andres Freund <andres@anarazel.de> writes:

 Andres> Thomas and I are sitting in a cafe and are trying to figure out
 Andres> what's going on...

I have a standalone test case:

#include <stdio.h>
#include <math.h>

int main(int argc, char **argv)
{
    double d1 = (argc ? 1e180 : 0);
    double d2 = (argv ? 1e200 : 0);
    int r2 = __builtin_isinf(d1 * d2);
    int r1 = isinf(d1 * d2);
    printf("r1 = %d, r2 = %d\n", r1, r2);
    return 0;
}

Note that swapping the r1 and r2 lines makes the problem disappear (!).

on amd64, clang 3.9.1:

cc -O2 -m32 flttst.c && ./a.out
r1 = 1, r2 = 0

cc -O2 flttst.c && ./a.out
r1 = 1, r2 = 1

Can't reproduce on 32-bit arm.

-- 
Andrew (irc:RhodiumToad)

Commits

  1. Error out for clang on x86-32 without SSE2 support, no -fexcess-precision.

  2. Fix 8a934d677 for libc++ and make more include order resistant.