[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Test tstdiomisc fails when multiplication of NAN by -1 results in NAN

On Mon, May 24, 2010 at 12:09 PM, John David Anglin
<dave@hiauly1.hia.nrc.ca> wrote:
> On Mon, 24 May 2010, Carlos O'Donell wrote:
>> On Sun, May 23, 2010 at 12:37 PM, John David Anglin
>> <dave@hiauly1.hia.nrc.ca> wrote:
>> > Carlos,
>> >
>> > You have found a bug.  The code should xor the sign bit when doing
>> > negation.  The existing code doesn't work for NANs.  I'll try to
>> > fix negdf2 and negsf2.
>> Should I file a bug?
> I'm testing a fix.  I'll install to 4.6 and backport later.
> A bug report would provide a record of the problem and fix.


>> You are correct, I've checked for 64-bit, and this is correct. I was
>> accidentally using the 32-bit float formats.
>> However, there is a bug in gdb then, which shows the NaN value as a
>> 32-bit value and interprets only the top 32-bits of the 64-bit double
>> -1.0 and displays -1.875 as if it were a 32-bit float.
> Yah, I have never liked gdb's handling of floating point values.
> For example, does "p/x val" interpret `val' as a float or double
> when val is a hex constant string.  It seems to depend on the
> length of the string.  You can't specify a size with `p'.  Values
> in registers are another story.

We'll fix this later... which I expect is exactly what I say every
time I come across this bug :-)


Reply to: