loupekind init¶
What loupekind init does¶
Write LoupeKind's configuration file and run the first-run bootstrap. Every other command reads that file.
The file holds only log, server and update settings at first. The
wedding-default cull profile and the neutral look are built into LoupeKind, not
written by init: an unset cull.profile.* key takes its built-in value, and a
shoot with no look catalog of its own uses neutral. See
Cull profile.
You rarely need to run it. On a fresh install, the first LoupeKind command of any
kind writes the same default config before it runs, with the credential wizards
suppressed. Run init to get those wizards from a terminal, or with --clean to
reset the config to the defaults.
Usage¶
Flags¶
| Flag | Short | Default | What it does |
|---|---|---|---|
--clean |
-c |
false |
Discard the existing configuration and replace it with the defaults |
--dir |
-d |
~/.loupekind |
Directory to write the config into |
--skip-gitlab |
true under CI, else false |
Skip the GitLab credential wizard | |
--skip-key |
-k |
true under CI, else false |
Skip the SSH key wizard |
The two --skip-* flags default themselves from the environment: each is true
when the CI environment variable is exactly true, and false otherwise. So a
pipeline skips the wizards without passing anything.
loupekind init gitlab runs the GitLab wizard on its own.
Plus the global flags on every command.
Running it again is safe¶
init is idempotent. Re-running it will not overwrite settings you have changed;
it fills in anything missing and leaves the rest alone. --clean is the explicit
opt-out: that does discard your settings and rewrite the defaults.
LoupeKind also re-runs init for you after a successful loupekind update,
so your config stays compatible across upgrades.
What it does without a terminal¶
With piped stdin, in CI, or under a test harness, the credential wizards are
skipped silently and only the base configuration is written. That isn't a
failure, but it does mean a scripted loupekind init never configures GitLab.
Run loupekind init gitlab from a real terminal later if you want it.
The two credential wizards it offers (GitLab and the SSH key) come from the
shared framework, and LoupeKind runs without either. The AI-provider key that
loupekind review needs is entered in the studio and stored in the OS
keychain, never through init.
The macOS app does not need it¶
The packaged macOS app ensures its configuration in-process at launch, before it
starts the studio. A fresh install and every upgrade are self-configuring with no
terminal step. loupekind init is for CLI installs.
Where the config file lands¶
~/.loupekind/config.yaml by default, or <--dir>/config.yaml. Confirm it with:
A per-shoot <shoot>/.loupekind/config.yaml is a separate, much narrower file. It
carries only cull.profile.* overrides, and init neither creates nor touches
it. See Configuration.