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: C++ ABI RFC [was Re: C++/libiberty PATCH for many mangling fixes (6057, 48051, 50855, 51322, etc)]


> > Note that one of the objectives of this email is to try and get
> > maintainers from thinking there is going to be "a perfect time" to
> > switch. Development history tells us there will always be more
> > changes. We've been sitting on ABI-breaking changes since 2003.
> 
> e.g. http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51823 which I only
> learnt about the other day.
> 
> And they keep on coming too, here's a brand new one:
> http://gcc.gnu.org/ml/libstdc++/2011-12/msg00153.html
> 
> Which is why I asked that aside about starting to fix those things
> when --enable-symvers=gnu-versioned-namespace is used, so that they're
> correct in v7 asap.  If I can't make std::future::wait_for() return
> future_status::deferred without breaking the v6 ABI, I'd like to at
> least get the fix out of my git clone and into libstdc++.so.7

Instead of the compat hacks?

;)
 
If you want to start a separate thread about the wait_for issue I can
help you think through options.

> > So, the smart thing to do is to plan a transition..... and to allow
> > testing in an easy and reproducible manner beforehand. The more time
> > the better.
> 
> I'm sold :)

Yay! TRON fights for the users!

-benjamin


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