[cfe-dev] Setting default dialect to C++11

[Robinson, Paul via cfe-dev]

Hi Mehdi, my understanding is that Tim was on the hook for making the tests C++14 clean.  I'm not aware that the work is done yet.
@Tim, if somebody else should pick this up, please let us know.
--paulr

From: cfe-dev [mailto:cfe-dev-bounces at lists.llvm.org] On Behalf Of Mehdi AMINI via cfe-dev
Sent: Sunday, October 15, 2017 12:29 PM
To: Richard Smith
Cc: clang developer list; Martin J. O'Riordan
Subject: Re: [cfe-dev] Setting default dialect to C++11

Hi,

This thread died 6 weeks ago and we're still using gnu++98 as a default as far as I can tell, what's the plan? Shall we make the switch in trunk? I suspect we would want this to happen early in the release cycle rather than late?

--
Mehdi


2017-08-24 17:33 GMT-07:00 Richard Smith via cfe-dev <cfe-dev at lists.llvm.org<mailto:cfe-dev at lists.llvm.org>>:
On 24 August 2017 at 15:18, Tim Northover via cfe-dev <cfe-dev at lists.llvm.org<mailto:cfe-dev at lists.llvm.org>> wrote:
Hi Paul,

On 24 August 2017 at 14:10, Robinson, Paul via cfe-dev
<cfe-dev at lists.llvm.org<mailto:cfe-dev at lists.llvm.org>> wrote:
> It's clearly too late to make a change like this for 5.0.  Is there any
> remaining roadblock to making the change for 6.0 (trunk)?

I've been slowly trying to make Clang C++14 clean. I think most of the
actual changes needed for Clang are in place now (the big one was
making C++14 mode friendlier to imaginary constants) and the rest is
fixing up tests.

I think most of the test changes are because the wording on how the
arguments to "operator new" get treated (promotion etc) was changed.
I've attached a WiP diff here for reference. Unfortunately it's not
even complete and these 4 tests still fail:

    Clang :: CodeGenCXX/vtable-available-externally.cpp
    Clang :: Lexer/cxx-features.cpp
    Clang :: OpenMP/taskloop_reduction_codegen.cpp
    Clang :: OpenMP/taskloop_simd_reduction_codegen.cpp

I don't even think they're particularly difficult ones, just where I
happened to stop when I last looked at it. I've also got to address
some review comments from Richard (I haven't forgotten) to make sure
we haven't decreased C-mode OpenMP coverage.

There are also a handful of bits of the test-suite that need to
explicitly specify c++98.

So I think the actual work needed is pretty small (and I can find time
to do it, probably even this weekend or something). We just need to
make the decision: C++11 or C++14 and go or no-go. My vote is go and
C++14.

+1. I think it would also make sense to include in the Clang 5 release notes that we intend for this to be the last release that defaults to -std=gnu++98.

_______________________________________________
cfe-dev mailing list
cfe-dev at lists.llvm.org<mailto:cfe-dev at lists.llvm.org>
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20171016/4b4a31cb/attachment.html>