-
Notifications
You must be signed in to change notification settings - Fork 1
Clarify baseline command: always reset to current state #37
Copy link
Copy link
Closed
Description
Problem
baseline and relax are confusing. Their names don't clearly communicate their behavior:
baselinecurrently does "ratcheting" - it tries to be smart about tightening limits over time, rejecting growth by defaultrelaxis a quick fix that adds a buffer to failing files
The ratcheting behavior in baseline is clever, but the command name suggests something simpler: "set the baseline to the current state."
Proposal
Change baseline so its primary purpose is "always reset to the current state of the world":
- Initialization:
loq baselineshould be the natural first command when adopting loq on an existing codebase - Re-baselining: When you've made deliberate changes and want to accept the new reality,
loq baselineshould capture that
The current ratcheting logic (reject growth, only tighten) is useful but belongs elsewhere or behind a flag.
Questions to resolve
-
What happens to the ratcheting functionality? Options:
- Move to a separate
loq ratchetcommand - Make it a flag like
--ratchetor--no-growth - Remove it entirely (YAGNI?)
- Move to a separate
-
How does this affect
relax? Ifbaselinebecomes "reset to current state",relaxmight be redundant or could be simplified to just "add buffer to current limits" -
Should
baselinecreate config if it doesn't exist? (currently requires existing config, unlikerelax)
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels