[cfe-dev] [libc++][patch] Improving Linux support

[Howard Hinnant]

On Jun 11, 2011, at 1:42 PM, Petteri Räty wrote:

> http://llvm.org/bugs/show_bug.cgi?id=8992#c17
> 
> Can someone comment on the patch here? It would be nice if at least
> Hello World worked on Linux.

Bugzilla does not appear to be fully functional at the moment.  Putting my comment here:

Howard

(In reply to comment #17)
> Created an attachment (id=6685) [details]
> Updated patch tested to fix std::cout << "Hello World" << std::endl;
> 
> (In reply to comment #16)
> > 
> > I haven't done any further testing besides the fact that it compiles. Any
> > comments before I proceed further?
> >
> 
> It seems supporting _DefaultRuneLocale is not needed to fix this bug now that I
> tested my patch with a hello world example. testit of course still has a lot of
> failures until everything is supported on Linux.

Sorry for the delay in reviewing this.

Comments:

I'm not comfortable with moving the __APPLE__ code away from the *_l functions.  I appreciate that the functionality has been replicated within the with_locale function.  But I am nervous about the performance implications of reinventing the *_l for __APPLE__.

On the positive side, you've achieved good code reuse with with_locale.  Perhaps it would be good to reinvent the *_l functions using with_locale *only* for those platforms which don't already have *_l and do have the per-thread uselocale.

I realize that this suggestion results in less elegant code.  But our customers care much more about performance than code elegance.