On Tue, 1 Oct 2019 at 14:59, Vladimir Sitnikov <[email protected]> wrote: > > >part of the release process > >should become tagging the voted upon commit SHA under rel/ to make it > >indelible. ('# git tag rel/v15.4.2 ' or something similar.) > > We have "publishRelase" Gradle task that creates "release tag" and moves > the release from dev to release area on dist.apahce.org. > We could teach it to create rel/v5.2.0 tag as well. > > So the question is: should we create both "v5.2.0 and rel/v5.2.0" tags? > Should we create just "rel/v5.2.0"? > Even though I don't like "rel/v5.2.0" naming for the release tag, it > becomes mandatory, so plain v5.2.0 will be excessive. > > >We should be very careful about changing anything in the POM, lest it > >causes the rel/tag to be created before the RC has been approved. > > Technically speaking, we never edit POM files directly. > JMeter-generated POM files do not contain tag/branch information. > > Here's how current POM files look like: > https://repository.apache.org/content/repositories/orgapachejmeter-1042/org/apache/jmeter/ > (e.g. > https://repository.apache.org/content/repositories/orgapachejmeter-1042/org/apache/jmeter/ApacheJMeter_core/5.2.0/ApacheJMeter_core-5.2.0.pom > ) > > > >Following direction from the Board, Infrastructure has modified git to > >permit force pushes, and branch/tag deletion > > This is super great news. +100500. > Now we can force-push small changes. What a time. > For instance: > https://github.com/apache/jmeter/commit/00fd8f09d578157cc9db5eb6c5e084c89a475d3f > and > https://github.com/apache/jmeter/commit/972a0d1c317934127f9f10ba2821a22d1d8e1723 > should > be just a single commit instead of two separate ones. > > Just in case, I suggest to use push --force-with-lease rather than simple > --force > > >It looks like JMeter Git may have been set up to protect all tags. > > My recent conversation with Infra re Git cleanup and the removal of the > stale jars concluded that > Infra validates **no** Git tags. > They validated branches only.
What do you mean by validate? > I've removed v5.2.0_RC1 from Git today and it went just fine. OK, good. I had thought that deletion of the tag had been tried, but failed. > However, the mail refers to refs/tags/rel, which indeed means "tags", so > the current protection seems to be related to rel/* tags only. That was certainly how I read the Infra email.
