sadpandajoe commented on PR #39925:
URL: https://github.com/apache/superset/pull/39925#issuecomment-4733729646
Thanks for the thorough pass — point by point:
**1. PR size / can i18n come out cleanly?** Yes — almost all of it. The
catalog regeneration touched 27 files (~1.8k lines), but only **2 strings are
genuinely new in this PR**, both in the new `RedirectWarning` page (`"Unsafe
link blocked"` and `"This link cannot be followed because its address is
unsafe."`). The catalog adds 14 new `msgid`s total; the other 12 ("Copy query
URL", "Delete query", "Connection looks good!", "There was an error generating
the permalink.", etc.) are pre-existing untranslated strings from components
this PR doesn't touch (`SavedQueryList`, `CRUD/hooks`) that a full
`extract-translations` run pulled in — they don't belong here. So I can either
trim the catalog delta to just those 2 strings, or drop the catalog changes
entirely and let the normal translation sync handle extraction. Either removes
the bulk of the size concern in-place. Which do you prefer? (The nav-core /
backend / normalizer pieces are harder to peel cleanly — they're interdependen
t with the `route_base` move — but the i18n churn is genuinely separable.)
**2. Internal shorthand + PLAN.md refs in comments** — Done. Stripped the
review-process shorthand and `PLAN.md` pointers across comments and test names;
rationale preserved inline (e.g. the WSGI layering-invariant explanation now
lives in the middleware module docstring). Comment-and-name-only — no logic or
assertion changes.
**3. Normalizer too heavy for one field** — Agreed, reworked. Replaced the
recursive walker (cycle-guard `WeakSet`, multi-field config) with a targeted
single-string strip wired at the one real consumer — the `explore_url`
double-prefix on the dataset list. The dead cycle-handling code and the
inconsistent shared-reference branch are gone with it.
**4. `getUrl` auto-dedupe undocumented** — Documented in UPDATING.md:
`SupersetClient.getUrl()` silently dedupes a single leading application root in
one pass, so the behavior (and its interaction with the static-scan layer) is
discoverable.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
