We keep the existing release cycle (release please branches into master) w.r.t. docker images, npm, etc.
This is targeted at users who want the latest/greatest but potentially unstable code
At the end of ADT sprints, we merge a regular release into master, e.g. v0.64.0
We cut our release/<adjective>-<animal> from that commit.
Suppose the current release is Sassy Salamander.
We create a release please PR for release/sassy-salamander.
We run our tests on release/sassy-salamander.
If our tests do not pass, we merge a fix into release/sassy-salamander.
When things pass, we merge the release please PR on release/sassy-salamander, which does a normal npm/docker release, and tags (devnet, sassy-salamander).
If there were no hotfixes, then this simply moves the tags forward to v0.64.0, otherwise, it releases v0.64.X (and moves the tags).
We keep the existing release cycle (release please branches into master) w.r.t. docker images, npm, etc.
At the end of ADT sprints, we merge a regular release into master, e.g. v0.64.0
We cut our
release/<adjective>-<animal>from that commit.Suppose the current release is Sassy Salamander.
We create a release please PR for
release/sassy-salamander.We run our tests on
release/sassy-salamander.If our tests do not pass, we merge a fix into
release/sassy-salamander.When things pass, we merge the release please PR on
release/sassy-salamander, which does a normal npm/docker release, and tags (devnet,sassy-salamander).If there were no hotfixes, then this simply moves the tags forward to v0.64.0, otherwise, it releases v0.64.X (and moves the tags).