[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Can we revert "avoid 64-bit intrinsics on 32-bit Windows"?
From: |
Pádraig Brady |
Subject: |
Re: Can we revert "avoid 64-bit intrinsics on 32-bit Windows"? |
Date: |
Tue, 06 Jan 2015 17:37:43 +0000 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.3.0 |
On 06/01/15 17:07, Paul Eggert wrote:
> On 01/06/2015 08:51 AM, Paul Eggert wrote:
>> -# pragma intrinsic _BitReverse
>> -# pragma intrinsic _BitReverse64
>> +# pragma intrinsic _BitScanReverse
>> +# pragma intrinsic _BitScanReverse64
>>
>
> Come to think of it, now that this typo is fixed, we should be able to
> revert the "avoid 64-bit intrinsics on 32-bit Windows" patch, right?
> That patch added complexity that I'd rather avoid. I think the previous
> diagnosis that the intrinsic is missing on 32-bit Windows is incorrect.
>
> If we really need something like that patch, let's put it into the
> "#elif _MSC_VER" section so that it's more easily separable from the
> mainline code. For now I reverted it by using the attached.
Absolutely revert it if it's not needed.
Though from a very quick search I thought the pragmas
only control optimization rather than availability.
Gisle, how does this this version work for you?
http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=blob_plain;f=lib/count-leading-zeros.h;h=5904eb7f
thanks,
Pádraig