This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: Remove algo logic duplication Round 3
- From: Christopher Jefferson <chris at bubblescope dot net>
- To: "libstdc++" <libstdc++ at gcc dot gnu dot org>
- Date: Tue, 24 Sep 2013 15:01:28 +0100
- Subject: Re: Remove algo logic duplication Round 3
- Authentication-results: sourceware.org; auth=none
- References: <52409F6F dot 7040609 at gmail dot com> <alpine dot DEB dot 2 dot 10 dot 1309232317070 dot 4088 at laptop-mg dot saclay dot inria dot fr> <5240B539 dot 1050202 at oracle dot com> <alpine dot DEB dot 2 dot 10 dot 1309232345420 dot 4088 at laptop-mg dot saclay dot inria dot fr> <5240BA10 dot 4090400 at oracle dot com> <alpine dot DEB dot 2 dot 10 dot 1309240028450 dot 4088 at laptop-mg dot saclay dot inria dot fr> <5240C28A dot 4040402 at oracle dot com> <5240C3C4 dot 8020305 at oracle dot com> <alpine dot DEB dot 2 dot 10 dot 1309240135390 dot 4088 at laptop-mg dot saclay dot inria dot fr>
I have read over this patch, and think it is good.
Now I can see the patch in a more readable form (and have applied it),
I now agree with this method of combining the algorithms. I did some
simple benchmarking and could not measure a difference with and
without this patch.
If you want to be 100% sure the patch works correctly, I would like to
see the boost test suite run with and without this patch, as boost is
traditionally one of the major users, and abusers, of the C++ standard
library. However, I wouldn't make that a requirement.
Chris
On 24 September 2013 00:37, Marc Glisse <marc.glisse@inria.fr> wrote:
> On Mon, 23 Sep 2013, Paolo Carlini wrote:
>
>> Hi again,
>>
>> On 9/23/13 5:36 PM, Paolo Carlini wrote:
>>>
>>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51680
>>> Funny, I don't remember making that comment at all, and isn't that old ;)
>>
>> ... but wait, that bug report is *not* about an inline in a member
>> function declaration. Is about an inline in a free standing function. I
>> definitely agree with myself about that comment.
>>
>> Did you read all the comments in the audit trail? Do you think the bug
>> report is relevant for the specific issue we are discussing today, that is
>> *member* functions defined in the body with/without an explicit inline?
>
>
> Maybe not. Even for free function templates as in the PR, I don't know if
> adding inline still makes any difference. As I said, let's forget that.
>
> --
> Marc Glisse