Open danlamanna opened 2 months ago
It appears some queries may fail with CACHALOT_FINAL_SQL_CHECK=True
; see #268.
I think it's currently a bit misleading that the default behavior forgoes correctness for performance.
Primarily, still looking for feedback as sdolemelipone has found. The error found in #268 isn't really a surprise; I've seen it before multiple times in Django cachalot when updating for the latest Django version.
Description
I'm struggling to understand the exact implications of
CACHALOT_FINAL_SQL_CHECK
introduced in https://github.com/noripyt/django-cachalot/pull/199. It seems like by default omitting this check can cause invalidations to be missed. It also appears (at least from looking at the code) that it could cache results that are part ofCACHALOT_UNCACHABLE_TABLES
if it misses them.In https://github.com/noripyt/django-cachalot/pull/199#issuecomment-979440732, @Andrew-Chen-Wang said it should be a flag because of its experimental nature. Is it still considered experimental, or can it be promoted to the default? I think it's currently a bit misleading that the default behavior forgoes correctness for performance.