[PATCH] Fix for logb/logbf/logbl (bz 13954/13955/13956)

[Andreas Jaeger]

On 05/04/2012 09:41 AM, Andreas Krebbel wrote:
> On 05/04/2012 09:09 AM, Andreas Jaeger wrote:
>> On 05/03/2012 11:23 PM, Andreas Krebbel wrote:
>>> On 05/03/2012 03:32 PM, Andreas Jaeger wrote:
>>>> On Thursday, May 03, 2012 14:58:44 Andreas Krebbel wrote:
>>>>> On 05/02/2012 02:15 PM, Andreas Jaeger wrote:
>>>>>> So, let's ask Dave and Andreas for testing of this patch,
>>>>>
>>>>> Tested on top of e2dbf201abdfa13fc4035a1a8888ecec91bef44c together with
>>>>> the build fix from David.
>>>>>
>>>>> Bye,
>>>>>
>>>>> -Andreas-
>>>>
>>>> Andreas, could you baseline the files again, please?
>>>
>>> Sure.
>>
>> Thanks.
>>
>> We've made many corrections in the last months, so some ULPs should be
>> smaller as well - did you really start with an empty file? If not,
>> please redo and commit.
>
> Yes it was a fresh start. The values appear to be similiar to e.g. sparc. Where did you
> expect different values?

On x86-64 we had some lower values but I don't remember which ones.

Do you have git access or shall I commit the change for you? This is for 
s390-64, isn't it?

> The only change I did manually was removing the logb values as suggested by Joseph:
>
> <  # logb
> <  Test "logb (-2000) == 10":
> <  ildouble: 2994973666814185929704314548549361968505810854084608
> <  ldouble: 2994973666814185929704314548549361968505810854084608
> <  Test "logb (0x0.0000000000001p-1022) == -1074":
> <  ildouble: 21849734927635462193269144330446690990765983662080
> <  ldouble: 21849734927635462193269144330446690990765983662080
> <  Test "logb (0x0.000011p-1022) == -1042":
> ...

Sure, those were wrong,

Andreas
-- 
  Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi
   SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
    GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg)
     GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126