This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ 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: [Patch] First bits of the algo merge


Gabriel Dos Reis wrote:

>Let me clarify.  I'm not dismisive of improvements.  I'm all for it.
>I'm just worried that things get complicated in the implementation of
>the initial idea of simplifyig the current implementation.
>  
>
And I totally agree about this. All in all, I think we are *still*
gaining from the idea, plus, as Howard showed yesyterday, there are
interesting cases where the use of the various __ops:: actually improves
the behavior with proxied iterators. Also, one important thing to
consider, in my opinion, is that even if predefined_ops.h is slighlty
bigger than we wanted it to be, is really mechanical, very simple.

That said, let's use v7 for what is meant: experiments, tests. If you
agree, we can re-evaluate the situation in, say, a couple of months, or
something. Same wrt move semantics, of course. Maybe in Berlin, if, as I
hope, we can meet face to face all of us.

>I hope that in the next years, we'll have more V3 maintainers -- or
>some "old timers" like me would retire and leave the place for
>"youngers".  We don't want them to spend hours after hours figuring
>why the simplest ideas weren't expressed simply, when attempting
>to fix a bug.
>  
>
This is certainly true. We (I) should improve the comments. Well, I must
say, it's true in very general terms, not only as far as your very ;) ;)
far retirement is concerned and our library.

Paolo.


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