This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc 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: Update on freeze status of glibc 2.18?


On Fri, Jun 21, 2013 at 01:14:15AM +0200, Torvald Riegel wrote:
> On Fri, 2013-06-21 at 00:24 +0200, Andi Kleen wrote:
> > > And it will not disallow elision for these cases.  As Roland points out,
> > 
> > It would for existing binaries.
> 
> That's not the concern you raised initially.  So, can we conclude that

I'm pretty sure I mentioned binaries.

> for the subset of newly compiled binaries, you wouldn't have concerns
> anymore about the DEFAULT != NORMAL split approach?

I have concerns about changes that do not allow elision in existing
binaries (no subset)

> If you were talking about potential users, then which scenarios do you
> have in mind?  If people use the stock glibc that their distribution
> provides, then the applications provided by the distribution will also
> have been recompiled.  That already gives you a large group of users and
> applications, and good testing.  If you're thinking third-party
> binaries, then they'll have to be recompiled to make use of new features
> -- but that's hardly unusual.

AFAIK Fedora doesn't even recompile every release (I may be wrong).

But it's definitely also about third part programs. Linux has
a lot of existing programs and libraries and they do not get recompiled
all the time.

So all the existing binaries. I want to see them all tested with
elision, and if everything goes well deployed widely with elision.

-Andi

-- 
ak@linux.intel.com -- Speaking for myself only.


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