1.56.0 beta release candidate 2 available for testing
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/ 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. This helps ensure the candidates build OK before we push them out to SourceForge. The files (and associated md5s) are: MD5 (boost-1_56_0_b1_rc2.7z) = cd6267325a8dadbb82b384c4776ee450 MD5 (boost-1_56_0_b1_rc2.tar.bz2) = b74cd72043b6b2fa16c6690633acb422 MD5 (boost-1_56_0_b1_rc2.tar.gz) = b51810be6737e73a4deee70e84895ca7 MD5 (boost-1_56_0_b1_rc2.zip) = f4ceb73492ae7d9b1843d841498850e5 Since this is the first post-git release, we are expecting some teething pains with this release. Please test extensively. Besides bug fixes the TypeIndex library has been included, and I think that the problems that mjklaim@gmail.com was seeing have been fixed. I would like to post this to SourceForge Thursday night/Friday morning, if possible. Thank you. --The release managers
On Jul 15, 2014, at 9:58 AM, Marshall Clow
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
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.
This helps ensure the candidates build OK before we push them out to SourceForge.
The files (and associated md5s) are: MD5 (boost-1_56_0_b1_rc2.7z) = cd6267325a8dadbb82b384c4776ee450 MD5 (boost-1_56_0_b1_rc2.tar.bz2) = b74cd72043b6b2fa16c6690633acb422 MD5 (boost-1_56_0_b1_rc2.tar.gz) = b51810be6737e73a4deee70e84895ca7 MD5 (boost-1_56_0_b1_rc2.zip) = f4ceb73492ae7d9b1843d841498850e5
Since this is the first post-git release, we are expecting some teething pains with this release. Please test extensively.
Builds successfully with clang-darwin. — Marshall
On 07/15/2014 12:58 PM, Marshall Clow wrote:
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
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.
This helps ensure the candidates build OK before we push them out to SourceForge.
The files (and associated md5s) are: MD5 (boost-1_56_0_b1_rc2.7z) = cd6267325a8dadbb82b384c4776ee450 MD5 (boost-1_56_0_b1_rc2.tar.bz2) = b74cd72043b6b2fa16c6690633acb422 MD5 (boost-1_56_0_b1_rc2.tar.gz) = b51810be6737e73a4deee70e84895ca7 MD5 (boost-1_56_0_b1_rc2.zip) = f4ceb73492ae7d9b1843d841498850e5
Builds successfully on Ubuntu 13.10 using gcc v4.8.1. On a side note: it looks like the v1.56 archives have ballooned in size by ~60% compared to v1.55. Any idea what the reason for the greatly increased size is? Jason
On Jul 15, 2014, at 10:36 AM, Jason Roehm
On 07/15/2014 12:58 PM, Marshall Clow wrote:
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
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.
This helps ensure the candidates build OK before we push them out to SourceForge.
The files (and associated md5s) are: MD5 (boost-1_56_0_b1_rc2.7z) = cd6267325a8dadbb82b384c4776ee450 MD5 (boost-1_56_0_b1_rc2.tar.bz2) = b74cd72043b6b2fa16c6690633acb422 MD5 (boost-1_56_0_b1_rc2.tar.gz) = b51810be6737e73a4deee70e84895ca7 MD5 (boost-1_56_0_b1_rc2.zip) = f4ceb73492ae7d9b1843d841498850e5
Builds successfully on Ubuntu 13.10 using gcc v4.8.1.
On a side note: it looks like the v1.56 archives have ballooned in size by ~60% compared to v1.55. Any idea what the reason for the greatly increased size is?
Yes. There’s a bunch of pre-built documentation checked into git that wasn’t checked into SVN. That’s most of it; we’re looking at doing something about that. libs/numeric/ublas/docs, for example, is almost 120 MB. — Marshall
On 07/15/2014 08:58 PM, Marshall Clow wrote:
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
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.
It's a bad time to ask this question, but I've realized there were are number of recent Boost.Build changes on develop, as summarized below. Maybe the key changes are - Support for Windows Phone with msvc. - Speed optimization for msvc - by avoiding running setup script for every C file. Do we want to merge that into master, and include in the release? Incidentally, it appears that 'merge pull request' button on github UI does indeed produce rather lame commit messages, so we might want to generally avoid it. Edward Diener (8): Add a clang for Windows jam file for use with clang when the target is msvc. Merge branch 'develop' of https://github.com/boostorg/build into develop Added version of common.jam to be used for clang in VC++ target mode. This is just a current example of changes which work when using clang in VC++ target mode in Windows, and should eventually be integrated into common.jam when we figure out how to support clang for both VC++ target mode and mingw/gcc target mode on Windows. This file is essentially clang-linux.jam with changes for clang on Windows in gcc/mingw target mode, which should also use clang-linux.jam. The changes are simply that multithreading should use the -pthread option rather than the -mthreads option on Windows. I could have directly updated clang-linux.jam, since the change only affects clang targeted to Windows but I am being conservative and showing the change in this file for others to look at first. Updated to make the setup threading more robust. Merge branch 'develop' of https://github.com/boostorg/build into develop Changes for clang on Windows using mingw/gcc Remove example file since changes have been integrated and tested in clang-linux.jam. Juraj Ivančić (3): Support more than 64 parallel jobs (-j). Avoid creating a thread if remaining jobs can be placed in the same array as thread handles. add support for MSVC setup script rewrite Jürgen Hunold (2): Merge pull request #16 from jivancic/slow_msvc_setup_scripts_2 Merge pull request #5 from jivancic/more_than_64_jobs Roland Schwarz (1): Handle explicit empty value for property-map.find-replace. Steve Gates (8): Adding a new feature to help support building for the Windows Runtime (Windows store/phone). 1. Adds a new feature <windows-api> for specifying the WINAPI_FAMILY API set to target. Defaults to desktop, which is how builds before my changes. 2. A few small changes for adding ARM support. 3. Changes to call to correct vcvars setup script for Windows Phone 8. Removing use of testing.execute feature until it gets added in separate pull request. Updating to include support for Windows Phone 8.1. Addressing feedback from review: 1. Fixing up a bit how the phone setup script location is calculated. 2. Reusing generate-setup-cmd rule in configure-really to avoid duplication and make cleaner. 3. Adding some more comments explaining. Support building for the Windows Runtime. Fixing build break with Visual Studio if phone scripts aren't installed with VS 2012 or 2013. Fixing problem preventing building on VS2010 due to locate-default-setup rule being passed an undefined value. Use WindowsPhoneCore.lib on Windows Phone Vladimir Prus (2): Merge pull request #18 from MSOpenTech/dev12_phone81_fix Merge pull request #19 from MSOpenTech/fix_dev10_break
On 07/15/2014 06:58 PM, Marshall Clow wrote:
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
...
MD5 (boost-1_56_0_b1_rc2.7z) = cd6267325a8dadbb82b384c4776ee450
build successfully on ubuntu 14-04 with gcc 4.8.2 and clang 3.5 results with gcc for building and running tests in status folder: ...failed updating 33 targets... ...skipped 41 targets... ...updated 50401 targets... see attached file for summary of test failures. Tell me if there is interests in more details. I can also test the beta releases with std=c++11, c++1y and stdlib=libc++ if there are any particular interest for any of that. -- Bjørn
On 07/15/2014 06:58 PM, Marshall Clow wrote:
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
...
MD5 (boost-1_56_0_b1_rc2.7z) = cd6267325a8dadbb82b384c4776ee450
2 failures building lib/log on ubuntu 14.04 with clang 3.5 c++11 libc++ ./b2 -j 6 toolset=clang cxxflags="-std=c++11 -stdlib=libc++" linkflags="-stdlib=libc++" 2 failures building libs/log bin.v2/libs/log/build/clang-linux-3.5/release/log-api-unix/threading-multi/trivial.o... bin.v2/libs/log/build/clang-linux-3.5/release/link-static/log-api-unix/threading-multi/trivial.o... Both failures caused by: ./boost/smart_ptr/make_shared_object.hpp:254:20: error: no matching function for call to 'sp_forward' ::new( pv ) T( boost::detail::sp_forward<Args>( args )... ); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ... ./boost/smart_ptr/detail/sp_forward.hpp:32:25: note: candidate function [with T = const char *] not viable: no known conversion from 'const char *' to 'const char *&&' for 1st argument template< class T > T&& sp_forward( T && t ) BOOST_NOEXCEPT clang -std=c+11 seems to satisfy: #if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 in ./boost/smart_ptr/detail/sp_forward.hpp and thus select sp_forward template that by comments looks like workaround for GCC 4.4 outdated version of rvalue references. The test should should be changed to: #if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && not defined __clang__ or #if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && not defined BOOST_CLANG or use a Boost.Config feature test if it exist for this. -- Bjørn
On Wed, Jul 16, 2014 at 3:44 PM, Bjørn Roald
On 07/15/2014 06:58 PM, Marshall Clow wrote:
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
clang -std=c+11 seems to satisfy:
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404
in ./boost/smart_ptr/detail/sp_forward.hpp and thus select sp_forward template that by comments looks like workaround for GCC 4.4 outdated version of rvalue references.
The test should should be changed to:
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && not defined __clang__
or
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && not defined BOOST_CLANG
or use a Boost.Config feature test if it exist for this.
That was my mistake, I'll prepare a pull request to fix this.
On 07/16/2014 02:24 PM, Andrey Semashev wrote:
On Wed, Jul 16, 2014 at 3:44 PM, Bjørn Roald
wrote: On 07/15/2014 06:58 PM, Marshall Clow wrote:
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
clang -std=c+11 seems to satisfy:
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404
in ./boost/smart_ptr/detail/sp_forward.hpp and thus select sp_forward template that by comments looks like workaround for GCC 4.4 outdated version of rvalue references.
The test should should be changed to:
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && not defined __clang__
or
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && not defined BOOST_CLANG
or use a Boost.Config feature test if it exist for this.
That was my mistake, I'll prepare a pull request to fix this.
thanks, for the sake of consistency, maybe better to use "&& !defined( __clang__ )" I tested both gcc and clang with and without -std=c++11 with the test like this, and boost builds successfully. #if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && !defined( __clang__ ) -- Bjørn
On Jul 16, 2014, at 5:31 AM, Bjørn Roald
On 07/16/2014 02:24 PM, Andrey Semashev wrote:
On Wed, Jul 16, 2014 at 3:44 PM, Bjørn Roald
wrote: On 07/15/2014 06:58 PM, Marshall Clow wrote:
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
clang -std=c+11 seems to satisfy:
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404
in ./boost/smart_ptr/detail/sp_forward.hpp and thus select sp_forward template that by comments looks like workaround for GCC 4.4 outdated version of rvalue references.
The test should should be changed to:
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && not defined __clang__
or
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && not defined BOOST_CLANG
or use a Boost.Config feature test if it exist for this.
That was my mistake, I'll prepare a pull request to fix this.
thanks, for the sake of consistency, maybe better to use "&& !defined( __clang__ )"
I tested both gcc and clang with and without -std=c++11 with the test like this, and boost builds successfully.
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404 && !defined( __clang__ )
I don’t think that this is a big enough problem to hold up the release of the beta. What do you think? [ Note: Beta != final release. That comes in a couple weeks, and we certainly want this fixed for that. ] — Marshall
Bjørn Roald wrote:
2 failures building lib/log on ubuntu 14.04 with clang 3.5 c++11 libc++ ... clang -std=c+11 seems to satisfy:
#if defined( __GNUC__ ) && __GNUC__ * 100 + __GNUC_MINOR__ <= 404
in ./boost/smart_ptr/detail/sp_forward.hpp and thus select sp_forward template that by comments looks like workaround for GCC 4.4 outdated version of rvalue references.
Thank you for the report. I've reverted this change on the master branch for now.
On Tue, Jul 15, 2014 at 6:58 PM, Marshall Clow
The second release candidate files for 1.56.0 beta are available at http://boost.cowic.de/rc/
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.
By the way, could you add tags to mark the revisions corresponding to betas?
participants (7)
-
Andrey Semashev
-
Bjørn Roald
-
Jason Roehm
-
Klaim - Joël Lamotte
-
Marshall Clow
-
Peter Dimov
-
Vladimir Prus