Adjust pre-commit/pre-push hooks to be compatible with git-worktree

[cecemei]

Description

Updates pre-commit/pre-push hooks to work with git-worktree.

git-worktree doesn't have .git folder. The .git in a git worktree directory is a file, e.g. this is the content of the hooks-config worktree in my local env:

gitdir: /Users/cecemei/codespace/oss-druid/.git/worktrees/hooks-config

The content in this gitdir is less compared with a regular git directory.

❯ ls -l /Users/cecemei/codespace/oss-druid/.git/worktrees/hooks-config
total 4560
-rw-r--r--@ 1 cecemei  staff      403 Feb  7 14:40 COMMIT_EDITMSG
-rw-r--r--@ 1 cecemei  staff       29 Feb  7 14:35 HEAD
-rw-r--r--@ 1 cecemei  staff       41 Feb  7 14:35 ORIG_HEAD
-rw-r--r--@ 1 cecemei  staff        6 Feb  7 14:35 commondir
drwxr-xr-x@ 3 cecemei  staff       96 Feb  7 14:35 fsmonitor--daemon
srwxr-xr-x@ 1 cecemei  staff        0 Feb  7 14:35 fsmonitor--daemon.ipc
-rw-r--r--@ 1 cecemei  staff       43 Feb  7 14:35 gitdir
-rw-r--r--@ 1 cecemei  staff  2313883 Feb  7 14:40 index
drwxr-xr-x@ 3 cecemei  staff       96 Feb  7 14:35 logs
❯ ls -l /Users/cecemei/codespace/oss-druid/.git
total 4728
-rw-r--r--@   1 cecemei  staff       99 Jan 24 16:25 COMMIT_EDITMSG
-rw-r--r--@   1 cecemei  staff    14333 Feb  7 13:56 FETCH_HEAD
-rw-r--r--@   1 cecemei  staff       23 Feb  7 13:56 HEAD
-rw-r--r--@   1 cecemei  staff       41 Feb  7 13:57 ORIG_HEAD
-rw-r--r--@   1 cecemei  staff     1623 Feb  7 14:36 config
-rw-r--r--@   1 cecemei  staff       73 Dec 11 11:28 description
drwxr-xr-x@   3 cecemei  staff       96 Jan 26 19:45 fsmonitor--daemon
srwxr-xr-x@   1 cecemei  staff        0 Feb  7 13:56 fsmonitor--daemon.ipc
drwxr-xr-x@   7 cecemei  staff      224 Feb  7 14:40 hooks
-rw-r--r--@   1 cecemei  staff  2313883 Feb  7 14:20 index
drwxr-xr-x@   4 cecemei  staff      128 Feb  6 14:30 info
drwxr-xr-x@   4 cecemei  staff      128 Dec 11 11:34 logs
drwxr-xr-x@ 260 cecemei  staff     8320 Feb  7 14:40 objects
-rw-r--r--@   1 cecemei  staff    62249 Feb  7 13:57 packed-refs
drwxr-xr-x@   6 cecemei  staff      192 Jan 26 20:51 refs
-rw-r--r--@   1 cecemei  staff      174 Jan 23 08:21 sourcetreeconfig
drwxr-xr-x@  13 cecemei  staff      416 Feb  7 14:35 worktrees

The git layout in a worktree doesn't have refs, objects, hooks. The hooks would still be triggered when working in a git-worktree, since it looks up the GIT_COMMON_DIR, which is /Users/cecemei/codespace/oss-druid/.git.

Note that for a regular git directory, GIT_COMMON_DIR is not set, thus git rev-parse --path-format=absolute --git-common-dir just returns GIT_DIR (e.x. /Users/cecemei/codespace/oss-druid/.git).

This PR has:

• been self-reviewed.
  • using the concurrency checklist (Remove this item if the PR doesn't have any relation to concurrency.)
• added documentation for new or modified features or behaviors.
• a release note entry in the PR description.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• added integration tests.
• been tested in a test Druid cluster.

[kgyrtkirk]

is the usage of eval is necessary here?

[cecemei]

it's probably not necessary, but while i was working on this i realized that this would change some files (run-all-in-dir.py, pre-commit, pre-push), and our script does cp_if_not_exist and throws if file is different from the ones in .git dir, so maybe we'd need to add a force flag to installing hook script to overwrite those. that might cause minor disruption to ppl's workflow.

[kgyrtkirk]

cript does cp_if_not_exist and throws if file is different from the ones in .git dir, so maybe we'd need to add a force flag to installing hook script to overwrite those.

that would be what a plain cp does - but I doubt the initial intention was that.
I don't think these scripts change often; but people may modify it locally...

I think cp_if_not_exist should have nothing to do with the usage of eval

there is nothing in this which would make eval necessary; please remove it.