TODO
Some features we need or want, plus some neat ideas that may not be too feasible to implement.
-
Once "post" method is sure to run, then move logger step from "pre" to "post".
-
Make sure ipc-parse can determine if action was actually performed or else the reason of why not.
-
Add [log.verbosity] 0 or 1 or 2 feature to control level of messaging spewage to the git client.
-
Fix hot-potatoe-grind bug where git-deploy on the same repo coming from the same IP (but running in different local directories) will infinite pummel the client and server, back and forth between the two directories.
- For example, running this on a single machine: "git deploy -C /usr/src/prod master & ; git deploy -C /usr/src/beta dev &"
-
Investigate converting get_fork_hash common fork sniffer scan to use "git merge-base --fork-point " instead of grinding through the logs.
-
Right now, the git server can "choke" if there are many git-deploy clients linked to a single repo.
- Investigate optimization to stagger git-deploy clients with random delays or even abort, if there is already a sufficient git-deploy running locally. If git-deploy crons on all servers are configured to run at the same time, i.e., every 10 minutes, then everything gets sluggish on the server at that time while everything catches up, and thus client operations will also hang for a while.
- Investigate optimization to make push-notify release ONLY relevant deploy clients, i.e., those with activity within past few minutes or those pinned to the specific branch that was pushed. It's difficult to tell which branch the client is pinned to from the server side, especially prior to the pull occurring. So the git-deploy client would probably need to inform the server via ENV which specific branch to be notified about changes on.
-
Provide branch, old hash, and new hash (for every branch updated by the git client) to post-read hook args (Requires man-in-the-middle sniffer)
- If nothing is updated to the client, then there will be no arguments to post-read.
- There will be a multiple of 3 args, each triple being branch, old hash, new hash, (depending on how many branches the client updates).
- In order to facilitate the InterProcessCommunication to post-read, information should be stored in $IPC/info.txt until the post-read completes.
-
Provide branch, old hash, new hash, and common parent "fork" commit hash to pre-write hook and post-write hook args.
- This fork hash can be used for debugging
- or for callback webhook.
- In order to facilitate the InterProcessCommunication between the pre-write and post-write, information should be stored in $IPC/info.txt until the post-write completes.
-
Fix git-deploy to handle split cheese case where git server uses both IPv4 and IPv6.
-
Add Support for HTTP protocol git read and write operations using Basic password Authorization (instead of only pubkeys over SSH protocol).
- Design a way to support "git-deploy" feature via HTTP (through REMOTE_USER or DeployToken or URI flag or Special HTTP Header or PAT [Personal Access Token] or maybe some other mechanism). Allow client to specify max-delay seconds (default 90) in case nothing new is ready since last pull.
-
Integrate or convert to be compatible with Git::Hooks::* plugins.
-
Augment Git::Hooks (maybe Git::Hooks::Server) to provide extra functionality
- Add Drivers to implement missing capabilities required
- GITHOOKS_PRE_READ / PRE_READ
- GITHOOKS_PRE_WRITE / PRE_WRITE
- GITHOOKS_POST_READ / POST_READ
- GITHOOKS_POST_WRITE / POST_WRITE
- Use same general compatible [githooks] syntax
- git config --list | grep 'githooks.plugin'
- git config --add githooks.plugin WebHook
- git config --add githooks."webhook".callbackurl https://website.com/post.cgi
- Use the same general githooks.pl format like: run_hook($0, @ARGV);
- Provide a seemless way to transport information between hooks.
- For example, the ability to export ENV variables from a PRE* hook to a POST* hook.
- Allow data in $git->stash to persist among all hooks where the $git object is the first argument passed to each custom block hook.
- Add Drivers to implement missing capabilities required
-
Monkey the core.hooksPath setting on the git server to point to these hooks provided.
-
[webhook] features for callback:
- Allow for WhiteList or BlackList filters to trigger webhook or ignore webhooks under certain conditions: : When specified branches are involved : When certain KEY users are involved : When coming from a specific IP or Network CIDR : When certain files are affected (tricky for pull reads) : When certain strings exist in any of the commit comments being pushed. (Tricky for pull reads.)
- provide failover queue retry mechanism fibinacci backoff until remote webhook server returns 2xx or 3xx status.
- at least provide when FORCE push destroys branch history : common fork point hash : list of commits that were destroyed