Enabling merge commits in TCK (#64)

Merge commits allow for well-crafted PRs to be preserved with full
attribution to the contributor and the history intact, should the
commiter decide it is necessary. They also provide clearer traceability
to originating pull request - the final set of commits in the PR are
exactly what gets merged into the target branch.

Without merge commits, the other two options provided by GitHub have
drawbacks. Squash and merge does not allow for histories to be preserved
with complex features. Some changes may be easier to understand as a
series of commits; for example, one of our spec alignment bugfixes in
up-cpp:

    https://github.com/eclipse-uprotocol/up-cpp/pull/240/commits

Squash and merge can also lose clear authorship when the commits are
squashed. Not only are the commits by the true author replaced by a
squash commit generated by the committer, but any PR containing commits
by more than one author will result in a final commit with one listed
author and vague attribution in the commit message to the rest of the
authors.

Rebase and merge brings new problems. While it does preserve the history
and authorship, it makes traceability to the original PR more difficult
since the final commits were not present in the PR and there is no merge
commit generated referencing the PR. Additionally, since the sequence of
commits is altered, the committer will be added to the attribution of
the history for *each* commit instead of just the merge commit and any
additional author metadata (beyond the author field, such as gpg
signatures) will be lost.

otterdog/eclipse-uprotocol.jsonnet

@@ -319,6 +319,7 @@ orgs.newOrg('eclipse-uprotocol') {
     },
     orgs.newRepo('up-tck') {
       allow_update_branch: false,
+      allow_merge_commit: true,
       code_scanning_default_languages+: [
         "python"
       ],