Hi, Otto, Yes, sorry. It was a result of the discussion in https://jira.mariadb.org/browse/MDEV-31230 but it would've been better to have something in the git commit too. Anyway, here's the explanation: this broke compilation on kvm-bintar-freebsd123-amd64, kvm-bintar-freebsd130-amd64, kvm-bintar-trusty-amd64, kvm-bintar-trusty-x86, kvm-fulltest, kvm-fulltest2, kvm-rpm-rhel8-aarch64, kvm-rpm-rocky8-aarch64, winx64-debug. Additionally it broke srpm builds on kvm-rpm-rhel8-amd64, kvm-rpm-rhel8-ppc64le, kvm-rpm-rocky84-amd64, kvm-rpm-centos74-amd64-debug, kvm-rpm-centos74-amd64, kvm-rpm-alma84-amd64. removing FindJava and FindJNI wrappers was wrong too, they have a purpose (fix a specific issue in cmake's files) and this issue wasn't fixed in cmake yet (as of 3.26.3) and cmake was not "printing wrong information about variable CONNECT_JDBC", there is no variable CONNECT_JDBC. There is a user-configurable setting CONNECT_WITH_JDBC which convey's user intention to enable the feature. It is ON by default. And there's a feature tag CONNECT_JDBC that shows in the summary whether the feature was actually enabled. As such there's no bug here, so reverting. Regards, Sergei Chief Architect, MariaDB Server and security@mariadb.org On Oct 11, Otto Kekäläinen via developers wrote:
Hi Sergei and others!
While reviewing changes in MariaDB 10.11.5 I was wondering about this commit:
https://github.com/mariadb/server/commit/dc9498beb6f79f08edd02fb7f7153e5150a...
It has only the default message from issuing 'git revert' and a human written explanation *why* the revert was done is missing.
AFAIK the reverted commit https://github.com/MariaDB/server/commit/1d0e3d80d84c2b375a444cdbb41c28afd78... was correct and good - so why was it reverted?