mirror of
https://github.com/hanzoai/insights.git
synced 2026-10-30 16:15:31 +00:00
zeekayandHanzo Dev
18d44c1ec0
Each of these was checked against the live database before it moved,
because a member name is free to rename and a stored value is not.
Members only -- the value never said hog, so nothing in Postgres changes:
TaggerType.HOG / EvaluationType.HOG the column already holds `script`;
`HOG = "script", "Script"` was a
member disagreeing with its own value
FEATURE_FLAGS.HOG the flag key `script` is read by the
flag service and does not move
HEDGEFN_SKIN_SPIDERHOG same, key `mascot-skin-spiderscript`
_HOGLAND_MARKER already `/var/lib/script`; renamed to
_BOX_MARKER, which is what the
function returns and what the comment
beside it already called it
Value too, having established nothing stores it: `InsightType.HOG = 'HOG'`
is derived from `query.kind === 'ScriptQuery'`, never written -- the one
writer, queryNodeToFilter, has no ScriptQuery entry, and zero of the 28
insight rows carry it in `filters.insight`.
Two readers keyed on that value had to move with it, and neither would
have been caught by renaming the enum alone: INSIGHT_TYPE_URLS spells
`HOG:` as a plain key while its siblings are computed, and the Slack
unfurl map keys the same string in Python.
The rest is copy a person reads -- "HOG code exceeds maximum size" is an
API error message, not an identifier.
Left alone deliberately: `ReviewHog` and `HogDesk`, which survive only in
applied migrations' `choices=` lists. The live members are already
`REVIEW = "review"` and `Desk = "desk"`, and Django compares choices in
Python without ever writing them, so those lists are a snapshot of a
field that has not looked like that for some time.
Co-authored-by: Hanzo Dev <[email protected]>