This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [v3] More of condition_variable_any

On 2 February 2010 22:29, Paolo Carlini wrote:
> Great!
> I'm really curious about this kind of tweak:
> - ? ? ?{ return __wait_until_impl(__lock, __atime); }
> + ? ? ?{ return this->__wait_until_impl(__lock, __atime); }
> for sure, unless there are very good reasons (i.e, a template base
> class, etc), we are not consistent in the library. Is there a good
> technical reason, or it's just a kind of stylistic choice? Because if
> it's just matter of style, and there are no subtle lookup issues (I'm
> not aware of any for member functions) I would be in favor of not adding
> more qualifications... In fact, I think we briefly discussed that with
> Benjamin a few years ago and in that occasion we *removed* many...

I was thinking ADL could find another function, because __atime can
contain user-defined types from arbitrary namespaces, but I should
have thought about it properly -- there are no lookup issues.

Here's a new patch, without 'this' qualifiers.  Tested x86_64/Linux again.

Attachment: cv_any.2.txt
Description: Text document

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]