On 11/11/22 19:59, Rainer Deyke via Boost wrote:
On 10.11.22 18:00, Marshall Clow via Boost wrote:
Available at: https://boostorg.jfrog.io/artifactory/main/beta/1.81.0.beta1/source/
The SHA256 checksums are as follows:
6b6b95fde0ec9c6c0e1584f35c6c2dc4eb6c8629c9635e844b8e6a9ad9594da1 boost_1_81_0_b1_rc1.tar.gz 9f2ce5d1cefb8f6b7e8cfcabc73632f348958e7cba2604dcb3688fecf6463d67 boost_1_81_0_b1_rc1.tar.bz2 e7547a789339499c6fc582a96ffb8989d97074f107520ddebc1ed5f30a0d6964 boost_1_81_0_b1_rc1.zip 758af4ae4f4224b6f2bd8f05fe0109d1f248c0377ee2dce03ccbeb30d82fb8b4 boost_1_81_0_b1_rc1.7z
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 am getting compile errors when using the beta. It seems that Boost.Json uses the identifier Opposite as a template argument, but X.h (from X11) contains the following line:
#define Opposite 4
This is a regression from Boost version 1.80.0.
X.org headers define quite a few macros like this. I did have name clashes between X.org headers and other libraries (non-Boost) before, and I'm of the opinion it is X.org that needs to be fixed. Unfortunately, given that it is barely maintained, people will likely have to work around this until it dies out. Bottom line, this problem isn't new, and IMHO is not worth fixing in Boost. Of course, it is up to Boost.JSON maintainers to decide.