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

[Robinson, Paul via cfe-dev]

Sylvestre Ledru reports that changing the default to C++11 or C++14 "would fix more than 70 issues" in building Debian with Clang.  See the "[cfe-dev] Using 5.0rc2 to rebuild Debian" thread.

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)?
--paulr

From: Martin J. O'Riordan [mailto:martin.oriordan at movidius.com]
Sent: Sunday, May 21, 2017 1:09 PM
To: 'Richard Smith'
Cc: Robinson, Paul; 'clang developer list'
Subject: RE: [cfe-dev] Setting default dialect to C++11

Thanks Richard,

Any chance the “Subject” could change to “RE: [cfe-dev] Setting default dialect to C++14”, simply to make it clear what is being discussed (I didn’t want to pre-emptively make such a change since I don’t own the discussion)?  I’m quite happy with the default being C++14, and for that matter, soon defaulting to C++17, but the discussion should be clear and unambiguous - IMO defaulting to tracking the current Standard is the better choice, with ‘-std=’ being used to specify an earlier Standard, or a proposed and experimental future standard.

            MartinO

From: metafoo at gmail.com<mailto:metafoo at gmail.com> [mailto:metafoo at gmail.com] On Behalf Of Richard Smith
Sent: 21 May 2017 20:23
To: Martin J. O'Riordan <martin.oriordan at movidius.com<mailto:martin.oriordan at movidius.com>>
Cc: Robinson, Paul <paul.robinson at sony.com<mailto:paul.robinson at sony.com>>; clang developer list <cfe-dev at lists.llvm.org<mailto:cfe-dev at lists.llvm.org>>
Subject: Re: [cfe-dev] Setting default dialect to C++11

The proposal is to go straight to C++14. That both matches GCC's current default and the suggested policy of setting the default to the most recent published standard for which we have sufficiently mature support.

On 21 May 2017 at 03:54, Martin J. O'Riordan <martin.oriordan at movidius.com<mailto:martin.oriordan at movidius.com>> wrote:
From the “Subject” and earlier messages in this thread, it would appear that the proposed default Standard is C++11; but the words in this message and a few other messages seem to suggest that the proposed default Standard is C++14.

I’m just curious about the proposed default, is it to be C++11 or C++14?  I don’t have any argument against changing the default to either, but would like to know what is really being proposed.

Thanks,

            MartinO

From: cfe-dev [mailto:cfe-dev-bounces at lists.llvm.org<mailto:cfe-dev-bounces at lists.llvm.org>] On Behalf Of Richard Smith via cfe-dev
Sent: 19 May 2017 18:12
To: Robinson, Paul <paul.robinson at sony.com<mailto:paul.robinson at sony.com>>
Cc: Clang Dev <cfe-dev at lists.llvm.org<mailto:cfe-dev at lists.llvm.org>>
Subject: Re: [cfe-dev] Setting default dialect to C++11

On 18 May 2017 11:18 pm, "Robinson, Paul" <paul.robinson at sony.com<mailto:paul.robinson at sony.com>> wrote:
Not to forget that Charles Li did a pile of work over the past year or so to tidy up a lot of the tests.  Given that the PS4 target already defaults to C++11, I should hope there would be little left to do.
Right, and indeed we have buildbots that test that a C++11 default works. So the remaining cleanups would be for C++14-specific issues, which should be minimal in comparison to the great work Charles did.
--paulr

From: cfe-dev [mailto:cfe-dev-bounces at lists.llvm.org<mailto:cfe-dev-bounces at lists.llvm.org>] On Behalf Of Richard Smith via cfe-dev
Sent: Thursday, May 18, 2017 8:30 PM
To: Tim Northover
Cc: cfe-dev at lists.llvm.org<mailto:cfe-dev at lists.llvm.org>

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

On 18 May 2017 at 19:38, Tim Northover <t.p.northover at gmail.com<mailto:t.p.northover at gmail.com>> wrote:
On 18 May 2017 at 18:56, Richard Smith <richard at metafoo.co.uk<mailto:richard at metafoo.co.uk>> wrote:
> Agreed. I think there's sufficient (although not unanimous) agreement that
> we should increase our default language standard; what we need is for
> someone to work through the immediate fallout from doing so (particularly in
> the test suite).

I'll see what I can do tomorrow. Hopefully I can at least get some
numbers here by the end of the day. Possibly patches; what's our
preferred solution to failures: fix the code (within reason), or
-std=c++98?

If the difference is interesting, test it in both modes. If the test is C++98-specific or has uninteresting diagnostic differences, -std=c++98.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20170824/7fe7ee52/attachment.html>