joshuathemiller / hookboy Goto Github PK

Folks delete repositories periodically. As such, the cache directory should get periodically flushed.

This is useful if caches are placed in a location outside of the repository, like in the user's home directory, and/or an application directory..

Helps with debugging and gives end users a better experience if something funky happens

Hookboy should display help text after every hook.
In light gray so as to not be too distracting.
Should be able to be turned off by running hookboy config set ShowHelp false (ShowHelp must be uppercase)
Should be able to be turned off when first installing hooks by running hookboy install --do-not-show-help
Subsequent runs of hookboy install should not modify any past set config values

Grapple is taken by a few other packages/tools, and it doesn't really feel right as far as gut checks go...

Wrangler may work, but it is also fairly generic.
HookWrangler seems to long.
Hrangler seems like a decent concatenation... The 'H' being silent unless said with a Hank Hill access, of course 🤪

Cowboys wrangle cows. How about hookboy? Seems less confusing to look at the hrangler.

Considered hooksmith, but this tool does not help with creating hooks, just wrangling them

The following seems like a decent starting point: https://paritylicense.com/versions/7.0.0.html

However, I would like to be able to support close source usage of the binaries of this project.

Add telemetry collection at various points:

• install
• apply
• uninstall
• global exception

hookboy.dev

• Docs explaining usage
• Hookboy Config Simple
• Hookboy Config Advanced

Some hooks are unable to be skipped with --verify=false. To account for this, all hookfiles should first check for an env var named hookboy_skip with a value set to true. This will allow hooks to be skipped via command line modification of env vars.

Once public, add Sonar scans. Can always prepare pipelines beforehand though

Not necessarily a bug... as it should always be present. I'm going to treat this as a feature 😁

The issue was originally occurring when trying to write the hookfiles themselves to the .git/hooks folder, and the folder itself not being present which occurred during manual testing (I deleted the folder).

MVP: Start with the ability to simply resolve URLs.
Better: dictionary to support shorter weblinks.

• GitHub
• Git in general (w/ easy commit hash support for versioning)

Cache location should be configurable.

Install should, if hookboy has not yet been used in the git repository, take a snapshot of the current git hooks folder. It should then delete all files besides the snapshot and place a file called .hookboy-conf to denote that hookboy has been installed.

Uninstall (new command) should check for the existence of a .hookboy-conf file and return gracefully if none is found. If one is found, all hookboy files and folders should be removed from the .git/hooks folder. If a snapshot folder is present, all files and folders from within should be moved back to the .git/hooks folder.

Clean (new command) should ask for confirmation before it is fully ran. When ran, it should delete all files and folders from within the .git/hooks directory.

• "Snapshot" could mean move all existing files to a folder called snapshot-b4-hookboy

Cache directory should not insert another path. What would be better would be if the default cache directory is simply set to .git/hooks/.hookboy-cache

To support better testing and future improvements

The default hook files that are created should check for the existence of the cache location upon every run, and fail if it does not exist. This is to provide a better notification experience for the end using after deletion of the cache folder (such as when git clean -xdf is ran). The failure message should be helpful, and state that running the install command again should resolve the failure.

And enhancement to #5

Better: dictionary to support shorter weblinks.

GitHub
Git in general (w/ easy commit hash support for versioning)