Master branch open for merging

The master branch is is now open for post-beta merges, but only as described in the Post-Beta Merge Policy. See < https://github.com/boostorg/boost/wiki/Releases%3A-Beta-Merge-Policy> -- The Boost Release Team

Marshall Clow wrote:
I ask permission to merge the following changes to master: # Boost.Endian: https://github.com/boostorg/endian/commit/2bd479044b796eaa9bff27ee019688fa55... This fixes a regression reported in https://github.com/boostorg/endian/issues/33. # Boost.Iostreams: https://github.com/boostorg/iostreams/commit/703e9a077b899417ead9fb6c35db8f0... When Iostreams builds zlib or bzip2 from source, install CMake configuration files for these two libraries (the libraries themselves are already being installed.) # tools/boost_install: https://github.com/boostorg/boost_install/commit/68c1506bd46dd75bc428210021c... When generating the CMake configuration files, strip suffix from `zlib.1` and `bzip2.1`, to enable Iostreams to find the dependencies from the previous commit. https://github.com/boostorg/boost_install/commit/096f05db519ae787ad9f6481d6a... https://github.com/boostorg/boost_install/commit/cf4b78006e8f22fb883ad56ce58... https://github.com/boostorg/boost_install/commit/ba82fdf6c14e9c111c12c921cdd... Add a test for using Iostreams from CMake. These commits are large, but they only affect testing on Travis and Appveyor and do not change the functionality of either boost_install or Iostreams. They just enable the previous two changes to be tested. # Superproject: https://github.com/boostorg/boost/commit/73410010da498717461a755dbeae3ab7daa... This omits the suggestion to run "b2 headers" when bootstrap.bat is run from a release.

I ask for permission to merge this change to the master of Boost.Histogram: https://github.com/boostorg/histogram/commit/90882ac086e8cec2a369c77c78f38d6... boost::histogram::axis::variant is using boost::variant internally as an implementation detail (we want to be able to change it in the future). While working on the Python wrapper for Boost.Histogram, an ambiguity was discovered in the visit function in this design. The ambiguity is resolved by using composition instead of private inheritance, which also simplifies the code. Best regards, Hans

On Wed, Mar 13, 2019 at 7:47 AM Marshall Clow via Boost
The master branch is is now open for post-beta merges
I'd like to merge everything from 229 to the latest: https://github.com/vinniefalco/beast/commits/v233 However, there are build failures on Travis due to changes in Boost.Context: https://github.com/boostorg/context/issues/103 Thanks

On Thu, Mar 14, 2019 at 12:41 PM Vinnie Falco via Boost < boost@lists.boost.org> wrote:
Odd; the test runners seem reasonably happy: https://www.boost.org/development/tests/master/developer/context.html -- Marshall

On Fri, Mar 15, 2019 at 7:30 AM Marshall Clow via Boost
Odd; the test runners seem reasonably happy: https://www.boost.org/development/tests/master/developer/context.html
There was a Mac OS fix to Boost.Context: https://github.com/boostorg/context/commit/42ac5103195e2f52a0b4a8a52d267fd11... Beast builds on Mac OS Travis now, but the clang configuration still has the error: https://travis-ci.org/boostorg/beast/jobs/506539071 Thanks

Dear Marshal, On Wed, Mar 13, 2019 at 3:47 PM Marshall Clow via Boost < boost@lists.boost.org> wrote:
I ask for permission to merge the following commits into master for Boost.Histogram: https://github.com/boostorg/histogram/commit/ae85b2a6b73bf2dfb6690d292fecad8... https://github.com/boostorg/histogram/commit/6391709d54b9866d4945ec44f145d6a... Improves and simplifies sanitized tests in the CI and fixes the following issues in the test matrix https://www.boost.org/development/tests/master/developer/Marshall-UBSAN-hist... https://www.boost.org/development/tests/master/developer/Marshall-UBSAN-hist... https://www.boost.org/development/tests/master/developer/Marshall-UBSAN-hist... https://github.com/boostorg/histogram/commit/fbc24390f9eacf13c2e6d11e2d8446e... This fixes an embarrassing spelling mistake in the public interface. I am very glad that this was found before the first release of Histogram. This is technically a breaking change, but I think an exception can be made here, because Boost.Histogram is a new library and doesn't have an established user base yet. https://github.com/boostorg/histogram/commit/e48be4c90da41472cda93a5a9587be6... This fixes an oversight, where a member type that should be private was public. This should be fixed before the release so that users will not rely on this. Best regards, Hans

On 13.03.19 15:46, Marshall Clow via Boost wrote:
Hi, I would like to merge this to master: https://github.com/boostorg/test/commit/d44509b4e4d25afb32e8832e404b38e95681... Mostly documentation changes. All tests green on my side. Thanks! Raffi

On 19.03.19 15:26, Marshall Clow via Boost wrote:
Thanks, I would like to merge this additional changes/fixes. The diff is visible here: https://github.com/boostorg/test/compare/master..b9d74d1f55be2d70cdaee9bdb3e... Changes: - Windows.h -> windows.h - comment unused variable and remove dead code - the rest is about trying to fix a test that should not be run on GCC4.8 and causing noise on Travis. It will be certainly another permission to merge later this week, only about documentation. Raffi

On 21.03.19 14:27, Marshall Clow via Boost wrote:
I would like to merge the following to master: * https://github.com/boostorg/test/commit/6a8773f62693877cfc9ac19df9990445a4ff... : documentation changes only * https://github.com/boostorg/test/commit/9ffe07ebd0fc1684b5f0fa8890872c91952b... : change in a cast Tests are green on my CI on develop. Thanks! Raffi PS: those are the last changes for 1.70 on my side.

On 25.03.19 14:48, Marshall Clow via Boost wrote:
Thanks! Now ... :) I would like to merge this to master: https://github.com/boostorg/test/commit/ec3aeac46b7f3ec287e1655c63cb55314504... This is to fix the warning issues mentioned in another thread. Tests green on my side (VS2008/2013/2015/2017, appleclang6 C++03/11, gcc5 C++03/11, UBSAN gcc8/C++14, UBSAN clang). Warning seems to be solved until the next visibility issue comes up. Thanks, Raffi

On Wed, Mar 13, 2019, 17:47 Marshall Clow via Boost
Requesting permission to merge https://github.com/boostorg/variant/commit/7ecf721f2b20df142d73d0dc1fe0677ab... (along with https://github.com/boostorg/variant/commit/90e7fbae4848dff3f3132e5056168cee7... that does nothing docs, b2 build or code related). Tests are green.

On 19/03/2019 16:16, Antony Polukhin via Boost wrote:
If it's permissible, I have two linked changes to Math and multiprecision libraries that would be good to merge: https://github.com/boostorg/multiprecision/commit/3772428c1bdacbd1ddb30dce7e... https://github.com/boostorg/math/commit/08b7a6145f90eff611a1afbdd21e9c54faeb... Plus one minor commit for Math that addresses some numeric_limits<__float128> issues: https://github.com/boostorg/math/commit/072ebaac2506876e10cd1b52cde2210eb5ce... A more comprehensive fix for the issues raised on the list about this will have to wait till after the next release. Thanks, John. --- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

Merged, thanks! --- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

On Wed, Mar 13, 2019 at 3:47 PM Marshall Clow via Boost < boost@lists.boost.org> wrote:
I ask for permission to merge a few more commits to the master for Histogram: https://github.com/HDembinski/histogram/commit/43f88383acffa380aba03d7cb31e4... This is an important bug-fix. The bug caused an assert to fail in debug mode and a segfault in release mode. A missing unit test was added to test the bug condition. https://github.com/HDembinski/histogram/commit/706fa31ca2a6dfb5ef0004904146d... Fixes all warnings on msvc that come from histogram code. https://github.com/HDembinski/histogram/commit/0e9231a24fd8d070aa85b925f610d... Restore building coverage reports on Travis (accidentally broken in a previous commit). https://github.com/HDembinski/histogram/commit/95ef98091e1022c35de59ebbb400a... https://github.com/HDembinski/histogram/commit/558907eca7c03b331185fa1b3eef8... Turn off building branches on Travis and Appveyor (I use PRs now to develop changes and to let CI test them).

On 3/13/19 5:46 PM, Marshall Clow via Boost wrote:
May I merge this change to Boost.Integer master: https://github.com/boostorg/integer/commit/33c165552cf1f9f16720acac43ad5477d... (sorry for the noise from space trimming, the essential part is this: https://github.com/boostorg/integer/commit/33c165552cf1f9f16720acac43ad5477d... ) This fixes lcm_range function returning incorrect result if there is a value of 1 in the input range. CI tests passed, a few testers that picked up the change are also green.

On Wed, Mar 13, 2019 at 7:47 AM Marshall Clow via Boost
The master branch is is now open for post-beta merges
I've got the urge.... The urge - TO MERGE!!! How do you feel about this? https://github.com/boostorg/beast/pull/1546 Thanks

Marshall Clow Via Boost wrote:
I ask permission to merge the following bugfixes to master: https://github.com/boostorg/geometry/compare/master...bg-prepare-include Furthermore I'd like to push new test cases and documentation changes along with these fixes. They are not in the comparison above for convenience. Adam

Marshall Clow Via Boost wrote:
AFAIU they're ok: https://travis-ci.org/boostorg/boost/builds/512048716 https://circleci.com/gh/boostorg/boost/11703 Geometry CI too: https://circleci.com/workflow-run/d25bd8d9-07eb-4ded-9519-cece8dc5fc1d Adam

On 13/03/2019 15:46, Marshall Clow via Boost wrote:
I'd like to merge the following container commit: https://github.com/boostorg/container/commit/0691018321f03aad662d7d2ae5b3e06... which solves clang gnu++98 and c++03 issues in Geometry: https://github.com/boostorg/geometry/issues/577 Best, Ion

I have some documentation tidying: https://github.com/boostorg/beast/pull/1548 OK to merge?

Do I have permission to merge to master a commit which adds [[nodiscard]] to outcome::failure_type and to outcome::success_type? Niall

On Tue, Apr 2, 2019 at 12:01 PM Niall Douglas via Boost < boost@lists.boost.org> wrote:
Do I have permission to merge to master a commit which adds [[nodiscard]] to outcome::failure_type and to outcome::success_type?
Have the tests cycled - do they look OK? When you're happy with the test results, then yes. -- Marshall

On Tue, Apr 2, 2019 at 1:54 PM Marshall Clow via Boost
Have the tests cycled - do they look OK? When you're happy with the test results, then yes.
I need to merge this, what do you think? https://github.com/boostorg/beast/commit/7c084509d680c34089345af7eac90d27eff... I'm still waiting for tests, Travis/Appveyor look good Thanks

On Fri, Mar 29, 2019 at 7:56 AM Marshall Clow via Boost
Not just for Vinnie, but documentation improvements are always OK to merge. You don't need to ask about them.
Well the snippet is in the docs but it is also in the tests so in theory the tests could break so I erred on the side of caution by asking. However, now there's a real bugfix that I would like to merge https://github.com/vinniefalco/beast/commit/51db4d491a027ce151c176b0d5ce3a62... Relevant defect: https://github.com/boostorg/beast/issues/1375 Thanks

On Fri, Mar 29, 2019 at 3:07 PM Vinnie Falco
I have a few more fixes I'd like to merge, here are the individual commits (including the previous one) *basic_parser::content_length is stable (API Change):* This fixes a bug in the value returend by basic_parser::content_length() https://github.com/vinniefalco/beast/commit/51db4d491a027ce151c176b0d5ce3a62... *test::stream has deprecated lowest_layer for ssl* This fixes a compilation error when used with ssl::stream https://github.com/boostorg/beast/pull/1549/commits/3c82717fed6a950dcd87ae26... *MSVC uses ::fopen_s* This fixes the warning cause by using the older std::fopen on MSVC https://github.com/boostorg/beast/pull/1549/commits/01bdea2b4d06c701699ab113...

Marshall Clow Via Boost wrote:
I ask permission to merge the following bugfixes to master: https://github.com/boostorg/geometry/compare/master...bg-prepare-include Adam

On Tue, Apr 2, 2019 at 6:40 AM Marshall Clow via Boost
If you're happy with the test results, then you can merge.
I'd like to merge this branch with a fix for websocket timeouts: https://github.com/boostorg/beast/pull/1553/commits Thanks
participants (11)
-
Adam Wulkiewicz
-
Andrey Semashev
-
Antony Polukhin
-
Hans Dembinski
-
Ion Gaztañaga
-
John Maddock
-
Marshall Clow
-
Niall Douglas
-
Peter Dimov
-
Raffi Enficiaud
-
Vinnie Falco