On Thu, Aug 17, 2017 at 7:18 AM, Klaim - Joël Lamotte via Boost < boost@lists.boost.org> wrote:

On 17 August 2017 at 13:10, Tom Kent via Boost wrote:

...

msvc-14.1 32 X X X msvc-14.1 64 X X X

Could you clarify which version of the compiler this was? 1900 or 1901? (or the Visual Studio 2017 update version?)

The full versions for VS and tools is: Python 2: 2.7.13 Python 2: 2.7.13 amd64 Python 3: 3.6.0 Python 3: 3.6.0 amd64 zlib: 1.2.8 bzip2: 1.0.6 Microsoft Visual Studio 2005 - msvc-8.0 - Service Pack 1 Microsoft Visual Studio 2008 - msvc-9.0 - Service Pack 1 Microsoft Visual Studio 2010 - msvc-10.0 - Service Pack 1 Microsoft Visual Studio 2012 - msvc-11.0 - Update 4 Microsoft Visual Studio 2013 - msvc-12.0 - Update 5 Microsoft Visual Studio 2015 - msvc-14.0 - Update 3 Microsoft Visual Studio 2017 - msvc-14.1 - RTW Tom

-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Olaf van der Spek via Boost Sent: Donnerstag, 17. August 2017 14:53 To: boost@lists.boost.org Cc: Olaf van der Spek Subject: Re: [boost] Boost 1.65.0 Release Candidate 4

On Thu, Aug 17, 2017 at 1:10 PM, Tom Kent via Boost wrote:

...

On Wed, Aug 16, 2017 at 10:33 PM, Marshall Clow via Boost < boost@lists.boost.org> wrote: msvc-14.1 32 X X X msvc-14.1 64 X X X

Do you do any runs with /std:c++latest?

... and /permissive-? Marcel

On Thu, Aug 17, 2017 at 5:06 PM, Tom Kent via Boost wrote:

On Thu, Aug 17, 2017 at 7:57 AM, Marcel Raad via Boost < boost@lists.boost.org> wrote:

...

...

-----Original Message----- From: Boost [mailto:boost-bounces@lists.boost.org] On Behalf Of Olaf van der Spek via Boost Sent: Donnerstag, 17. August 2017 14:53 To: boost@lists.boost.org Cc: Olaf van der Spek Subject: Re: [boost] Boost 1.65.0 Release Candidate 4

On Thu, Aug 17, 2017 at 1:10 PM, Tom Kent via Boost wrote:

...

On Wed, Aug 16, 2017 at 10:33 PM, Marshall Clow via Boost < boost@lists.boost.org> wrote: msvc-14.1 32 X X X msvc-14.1 64 X X X

Do you do any runs with /std:c++latest?

... and /permissive-?

No and no, I run with the default, out of the box, settings. I'm confident the default settings are the most common in the wild, so that is what I build the windows binaries to target.

Would it be easy to do extra runs with those settings? -- Olaf

On 17/08/2017 14:01, Olaf van der Spek via Boost wrote:

On Thu, Aug 17, 2017 at 5:33 AM, Marshall Clow via Boost wrote:

...

As always, the release managers would appreciate it if you download the candidate of your choice and give building it a try. Please report both success and failure, and anything else that is noteworthy. The fix for "Unknown compiler version - please run the configure tests and report the results" doesn't appear to be included..

It's not, it's too late. I've just filed a PR to improve the messages and provide a way to turn them off: https://github.com/boostorg/config/pull/178 --- This email has been checked for viruses by AVG. http://www.avg.com

On 18 August 2017 at 10:37, Olaf van der Spek via Boost wrote:

On Thu, Aug 17, 2017 at 6:54 PM, John Maddock via Boost wrote:

...

It's not, it's too late.

I've just filed a PR to improve the messages and provide a way to turn them off: https://github.com/boostorg/config/pull/178

Isn't it just in time for RC6? :p

As it's just a message, not a compile error, and it's not even in develop yet, I'd say no. We can't delay forever.

On 18 August 2017 at 15:24, Egor Pugin via Boost wrote:

Maybe it's possible to cherry-pick required changes and release 1.65.1 a bit later with full VS2017.3/clang 5.0/gcc 7.2 support?

Both clang-5.0 and VS2017.3 are major updates (no clue about GCC 7.2), getting things right certainly must be more important than just getting release 1.65 half-broken out of the door (quality over speed). I know a number of people are working over-time to make things happen ASAP, but sometimes ASAP is not the most important. A delay of a week or 2 don't make any difference. desgki -- "*Ihre sogenannte Religion wirkt bloß wie ein Opiat reizend, betäubend, Schmerzen aus Schwäche stillend.*" - Novalis 1798

On 18 August 2017 at 16:46, Rene Rivera via Boost wrote:

It makes a difference for some of us.

You mean "us" devs, no? What about users?

As some of us are waiting for the release to happen so that we can merge fixes and improvements to master for the next release cycle.

Relax, time's on our side! There are fixes (PR's) out there that have not been merged for years and nobody does anything about those, even though they cause plenty of people major head-aches, but, yes, those people are not part of "us". degski -- "*Ihre sogenannte Religion wirkt bloß wie ein Opiat reizend, betäubend, Schmerzen aus Schwäche stillend.*" - Novalis 1798

On 18 August 2017 at 17:53, Daniel James via Boost wrote:

We used to delay releases until every little thing was sorted out. And that lead to releases being heavily delayed, which is good for no one, users or developers.

Me thinks that for users, the only thing that counts is that a new release is rock-solid (i.e. at least as good as the previous one). We see plenty of people on the list that are still on, just for arguments' sake 1.47, there must be a reason for that. They (must) feel un-comfortable with newer releases and the older stuff (the releases that got delayed maybe?) just works for them (the bit they are using in any case). degski -- "*Ihre sogenannte Religion wirkt bloß wie ein Opiat reizend, betäubend, Schmerzen aus Schwäche stillend.*" - Novalis 1798

On 18 August 2017 at 18:22, Daniel James via Boost wrote:

Regardless, we're not going to change our release procedure at the very last minute.

There does not seem to be much of a procedure from where I'm sitting. Either one just pushes the (release )button every 6 weeks (like Rust), 3 months or whatever, no matter what issues are lingering: "We''ll fix it in the next release approach", or one releases the latest product when it's ready. We seem to have a half-way house here. BUT, I'll rest my case here, life's too short and the next release is only 4 months away (pre-cluding delays). degski -- "*Ihre sogenannte Religion wirkt bloß wie ein Opiat reizend, betäubend, Schmerzen aus Schwäche stillend.*" - Novalis 1798