[release] Boost 1.81.0 Release Candidate 1 is available
The first release candidates for the 1.81.0 release are now available at: <https://boostorg.jfrog.io/artifactory/main/release/1.81.0/source/> The SHA256 checksums are as follows: 205666dea9f6a7cfed87c7a6dfbeb52a2c1b9de55712c9c1a87735d7181452b6 boost_1_81_0_rc1.tar.gz 6e689b266b27d4db57f648b1e5c905c8acd6716b46716a12f6fc73fc80af842e boost_1_81_0_rc1.zip d7bc4e2a2aee374a8c24832cacfed1265680c1e824d09fab3e00f87508155353 boost_1_81_0_rc1.7z 71feeed900fbccca04a3b4f2f84a7c217186f28a940ed8b7ed4725986baf99fa boost_1_81_0_rc1.tar.bz2 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 Release managers
On Dec 8, 2022, at 7:24 AM, Marshall Clow <mclow.lists@gmail.com> wrote:
The first release candidates for the 1.81.0 release are now available at: <https://boostorg.jfrog.io/artifactory/main/release/1.81.0/source/>
The SHA256 checksums are as follows:
205666dea9f6a7cfed87c7a6dfbeb52a2c1b9de55712c9c1a87735d7181452b6 boost_1_81_0_rc1.tar.gz 6e689b266b27d4db57f648b1e5c905c8acd6716b46716a12f6fc73fc80af842e boost_1_81_0_rc1.zip d7bc4e2a2aee374a8c24832cacfed1265680c1e824d09fab3e00f87508155353 boost_1_81_0_rc1.7z 71feeed900fbccca04a3b4f2f84a7c217186f28a940ed8b7ed4725986baf99fa boost_1_81_0_rc1.tar.bz2
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.
I have successfully built the libraries in Mac OS 12.6.1, with "Apple clang version 14.0.0 (clang-1400.0.29.202)” and C++ 03/11/14/17/20/2b, on both x86 and ARM. — Marshall
On Thu, 8 Dec 2022 at 16:25, Marshall Clow via Boost wrote:
The first release candidates for the 1.81.0 release are now available at: <https://boostorg.jfrog.io/artifactory/main/release/1.81.0/source/>
The SHA256 checksums are as follows:
Why is there no tag/release at GitHub? https://github.com/boostorg/boost/tags https://github.com/boostorg/boost/releases Boost is only usable to me when it contains CMake support and I cannot test the above tarballs which completely lack it. Thank you, Mojca
Am 09.12.22 um 05:41 schrieb Mojca Miklavec via Boost:
Boost is only usable to me when it contains CMake support and I cannot test the above tarballs which completely lack it.
You need to install Boost with the official workflow (bootstrap and b2) which will generate CMake config files consumeable via `find_package` Not sure about the reason we don't use tags for RCs though.
On Fri, 9 Dec 2022 at 09:57, Alexander Grund via Boost wrote:
Am 09.12.22 um 05:41 schrieb Mojca Miklavec via Boost:
Boost is only usable to me when it contains CMake support and I cannot test the above tarballs which completely lack it.
You need to install Boost with the official workflow (bootstrap and b2) which will generate CMake config files consumeable via `find_package`
The thing is ... is there an easy way to replace something like set(BOOST_VERSION_SOURCE 1.81.0.beta1) FetchContent_Declare(boost URL https://github.com/boostorg/boost/releases/download/boost-${BOOST_VERSION_SOURCE}/boost-${BOOST_VERSION_SOURCE}.tar.xz) FetchContent_MakeAvailable(boost) with the code that will run and compile bootstrap and b2 automatically and subsequently find the installed version, at least for some typical Linux and MSVC? (I assume that I should be able to come up with some overly complex code to achieve that, but I would really like to avoid going that way.) Asking every user to install a gazillion of dependencies is a support nightmare. Having boost (and many other dependencies) compile automatically as part of the build process (when dependencies are not already found on the system) is an extremely helpful lifesaver. Mojca (PS: When I tried installing a binary version of boost for GitHub actions on Windows via various package managers it sometimes took 11 minutes or more, so I have given up on that idea.)
On Mon, Dec 12, 2022 at 8:50 AM Mojca Miklavec via Boost <boost@lists.boost.org> wrote:
On Fri, 9 Dec 2022 at 09:57, Alexander Grund via Boost wrote:
Am 09.12.22 um 05:41 schrieb Mojca Miklavec via Boost:
Boost is only usable to me when it contains CMake support and I cannot test the above tarballs which completely lack it.
You need to install Boost with the official workflow (bootstrap and b2) which will generate CMake config files consumeable via `find_package`
The thing is ... is there an easy way to replace something like
set(BOOST_VERSION_SOURCE 1.81.0.beta1) FetchContent_Declare(boost URL https://github.com/boostorg/boost/releases/download/boost-${BOOST_VERSION_SOURCE}/boost-${BOOST_VERSION_SOURCE}.tar.xz) FetchContent_MakeAvailable(boost)
with the code that will run and compile bootstrap and b2 automatically and subsequently find the installed version, at least for some typical Linux and MSVC?
(I assume that I should be able to come up with some overly complex code to achieve that, but I would really like to avoid going that way.) Asking every user to install a gazillion of dependencies is a support nightmare. Having boost (and many other dependencies) compile automatically as part of the build process (when dependencies are not already found on the system) is an extremely helpful lifesaver.
This is what package managers were invented to solve. -- -- René Ferdinand Rivera Morell -- Don't Assume Anything -- No Supone Nada -- Robot Dreams - http://robot-dreams.net
пн, 12 дек. 2022 г. в 17:50, Mojca Miklavec via Boost <boost@lists.boost.org>:
(PS: When I tried installing a binary version of boost for GitHub actions on Windows via various package managers it sometimes took 11 minutes or more, so I have given up on that idea.)
I can't imagine package managers pulling and building Boost significantly slower than FetchContent does. That being said, you usually won't be able to install a Boost *release candidate* with a package manager.
On Mon, 12 Dec 2022 at 15:58, René Ferdinand Rivera Morell wrote:
This is what package managers were invented to solve.
Which one? The first google hit for a package manager under Windows points to https://learn.microsoft.com/en-us/windows/package-manager/ and I don't seem to find boost there: https://github.com/microsoft/winget-pkgs/tree/master/manifests/b Sure, there's probably a package manager that provides boost, but maybe another dependency is missing there. Supporting novice users (as well as setting up CI jobs) on tons of operating systems is a nightmare as dependencies grow. And, in fact, I would dare to claim that CMake's FetchContent does in a way count to be "a package manager". Not a fully fledged one, but certainly an extremely useful one. On Mon, 12 Dec 2022 at 16:34, Дмитрий Архипов via Boost wrote:
I can't imagine package managers pulling and building Boost significantly slower than FetchContent does.
I first wanted to write a longer explanation, but thought that perhaps a live example would tell more than thousand words: https://github.com/mojca/cmake-troubleshooting-example/actions/runs/36810987... Results: - build from source: 1:48 for everything: to fetch boost, configure and build the project - install boost via vcpkg: 4:55 (+ whatever it takes to configure and build the project) - install boost via choco: 12:58 to install boost + 0:27 to configure and build the project - install via windows package manager: good luck ;), see above (I'll be grateful for hints about the best way to drop the time even lower. Maybe simply unzipping official windows builds ...) Plus whatever it takes to (learn how to) install the package manager first. (OK, vcpkg is a tiny bit biased. The list above fetches slightly more dependencies than the minimal example needs, but it fetches the real list that I need for a project. For choco I have no choice.) On Ubuntu it's a matter of a couple of seconds. But even on Ubuntu it's a problem that older versions don't provide a sufficiently recent version, so in many cases it needs to be installed "manually" as well. Reasons for longer build times: 1. vcpkg has to build everything from source, and it has to build full package dependencies even when the code itself doesn't need any of those recursive dependencies. 2. choco needs to install absolutely everything, and then I suspect that they might be checksumming files or doing some other magic that makes the installation of existing binaries (with no compilation) painfully, painfully slow.
That being said, you usually won't be able to install a Boost *release candidate* with a package manager.
True. But also: it's both heavily non-trivial to handle if one wants a project to be easy to build for novice users across a wide range of operating systems and compilers. Bottomline: having a CMake-compliant source version of Boost is still extremely, extremely useful. Mojca
On Tue, Dec 13, 2022, at 2:03 AM, Mojca Miklavec via Boost wrote:
Supporting novice users (as well as setting up CI jobs) on tons of operating systems is a nightmare as dependencies grow. Agreed
And, in fact, I would dare to claim that CMake's FetchContent does in a way count to be "a package manager". Not a fully fledged one, but certainly an extremely useful one. This claim seems to be undercut by the implication that you don't know package managers that provide boost. I assume that was merely for the sake of the argument, and you probably know them full well. I, for one, like to build boost manually, but I've used conan, biicode, build2, vcpkg and even nuget in the past. They all "worked".
FetchContent burdens developers with duplicating the logic - specifically platform-specific dependencies as you pointed out - to build dependencies. Of course, this implies high degree of control. However, that does not usually jibe with novice-friendliness or portability.
Bottomline: having a CMake-compliant source version of Boost is still extremely, extremely useful.
Agreed. But not with the goal of being novice friendly or "package-management" (quite the opposite: it's to allow experts to streamline their build scripts and pick bespoke revisions) Just my perspective, Seth
On Mon, Dec 12, 2022 at 5:03 PM Mojca Miklavec via Boost <boost@lists.boost.org> wrote:
...
Once you do figure out how to get Boost 1.81 RC1 (or the release) installed please have a look at Boost.URL, a new library! Check out this BEAUTIFUL Reference Card: <https://www.boost.org/doc/libs/master/libs/url/doc/html/url/ref/helpcard.html> Full documentation here: <https://www.boost.org/doc/libs/master/libs/url/doc/html/index.html> Thanks!
at Boost.URL, a new library! Check out this BEAUTIFUL Reference Card: <https://www.boost.org/doc/libs/master/libs/url/doc/html/url/ref/helpcard.html>
Indeed, it is very nice Vinnie. What's missing though, are URLs inside the SVG to go to matching reference docs :). --DD [1]: https://developer.mozilla.org/en-US/docs/Web/SVG/Element/a
On Dec 8, 2022, at 8:41 PM, Mojca Miklavec <mojca.miklavec.lists@gmail.com> wrote:
On Thu, 8 Dec 2022 at 16:25, Marshall Clow via Boost wrote:
The first release candidates for the 1.81.0 release are now available at: <https://boostorg.jfrog.io/artifactory/main/release/1.81.0/source/>
The SHA256 checksums are as follows:
Why is there no tag/release at GitHub? https://github.com/boostorg/boost/tags https://github.com/boostorg/boost/releases
We don’t tag release candidates; only releases. — Marshall
On Thu, Dec 8, 2022 at 9:24 AM Marshall Clow via Boost < boost@lists.boost.org> wrote:
The first release candidates for the 1.81.0 release are now available at: <https://boostorg.jfrog.io/artifactory/main/release/1.81.0/source/>
The SHA256 checksums are as follows:
205666dea9f6a7cfed87c7a6dfbeb52a2c1b9de55712c9c1a87735d7181452b6 boost_1_81_0_rc1.tar.gz 6e689b266b27d4db57f648b1e5c905c8acd6716b46716a12f6fc73fc80af842e boost_1_81_0_rc1.zip d7bc4e2a2aee374a8c24832cacfed1265680c1e824d09fab3e00f87508155353 boost_1_81_0_rc1.7z 71feeed900fbccca04a3b4f2f84a7c217186f28a940ed8b7ed4725986baf99fa boost_1_81_0_rc1.tar.bz2
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 Release managers
Looks good for Windows/Visual Studio. toolset arch compile Link Execute msvc-10.0 32 X X X msvc-10.0 64 X X X msvc-11.0 32 X X X msvc-11.0 64 X X X msvc-12.0 32 X X X msvc-12.0 64 X X X msvc-14.0 32 X X X msvc-14.0 64 X X X msvc-14.1 32 X X X msvc-14.1 64 X X X msvc-14.2 32 X X X msvc-14.2 64 X X X msvc-14.3 32 X X X msvc-14.3 64 X X X Compile means that the b2 command completed without errors Link means that visual studio was able to link a sample executable to a library (libboost_thread-vcXXX-mt[-gd]-1_XX.lib) generated Execute means that the linked program executed without errors. Full build logs can be found here: https://gist.github.com/teeks99/da27ca59a0b1c7837651b794f19e600b Tom
On Thu, Dec 8, 2022 at 7:25 AM Marshall Clow via Boost <boost@lists.boost.org> wrote:
...
No Asio release notes? This page looks mighty lean: <https://www.boost.org/users/history/version_1_81_0.html> Thanks
On 14. Dec 2022, at 17:48, Vinnie Falco via Boost <boost@lists.boost.org> wrote:
No Asio release notes? This page looks mighty lean:
Many other libraries including Histogram are also missing, but they are listed here: https://github.com/boostorg/website/blob/master/feed/history/boost_1_81_0.qb...
On Dec 14, 2022, at 8:48 AM, Vinnie Falco <vinnie.falco@gmail.com> wrote:
On Thu, Dec 8, 2022 at 7:25 AM Marshall Clow via Boost <boost@lists.boost.org> wrote:
...
No Asio release notes? This page looks mighty lean:
Those are for the beta release. They will be updated with the notes for the actual release (hopefully later today; I’m working on it now) — Marshall
participants (10)
-
Alexander Grund -
Dominique Devienne -
Hans Dembinski -
Marshall Clow -
Mojca Miklavec -
René Ferdinand Rivera Morell -
Seth -
Tom Kent -
Vinnie Falco -
Дмитрий Архипов