[Issue 4191] New: [FreeBSD] real constants are rounded to double precision

[d-bugmail at puremagic.com]

http://d.puremagic.com/issues/show_bug.cgi?id=4191

           Summary: [FreeBSD] real constants are rounded to double
                    precision
           Product: D
           Version: unspecified
          Platform: x86
        OS/Version: FreeBSD
            Status: NEW
          Keywords: patch, wrong-code
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: nobody at puremagic.com
        ReportedBy: rsinfu at gmail.com


--- Comment #0 from Shin Fujishiro <rsinfu at gmail.com> 2010-05-15 01:44:46 PDT ---
Created an attachment (id=629)
Patch for DMD (svn r487)

This code fails with dmd built on FreeBSD/i386:
--------------------
static assert(real.max < real.infinity); // fails on FreeBSD
--------------------
Some static assertions in std.math are blocked by this problem.

There is a bug in GCC on FreeBSD/i386 and long double constants are rounded to
double precision(*).  A workaround is already there in DMD as Port::ldbl_max,
but it seems it's forgotten. :)  The attached patch enables this workaround,
and adds necessary fpsetprec.

[*] freebsd-standards: long double broken on i386?
http://groups.google.co.jp/group/lucky.freebsd.standards/browse_thread/thread/5d7b42dac8d29461

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------