Open bayandin opened 1 year ago
I support a separate GH org for forks
I support a separate GH org for forks
Sounds nice, we can even add patches if required (for example for pgvector).
The only downside I can see is that if, for some reason or by mistake, GitHub decide to block any of the repos, this affects all the forks, and we won't be able to build the image. So we might be
And, to not to forget in the future:
We download postgis
from their host, the tarball comes with autogen.sh
, if we switch it to the build from a source, then we need to add autoconf/automake to the image (and run autoreconf).
We need to consider how and when to merge changes from upstream into our fork/repo/copy. Also how notifications about patches and releases reach us. Not every project does timely releases, especially not for bugfixes.
The idea with additional patches is good, we can show what we change when building extensions. That enables transparency.
One idea to work around the blocking problem is to run a separate GitLab instance instead.
Motivation
If we can't download a package (for example for extension) due to third-party problems, we can't build a compute-node image, which is bad.
Ref:
DoD
Implementation ideas
Tasks
Other related tasks and Epics