villebro commented on PR #44508:
URL: https://github.com/apache/superset/pull/44508#issuecomment-5769831076
Pushed a fix for the 16 test failures. Root cause: the token→byte rename was
mechanical (\`token_limit\`→\`max_bytes\`) but several tests hardcoded numeric
literals (500, 700, 3000, 200...) that were calibrated as token counts. The old
char-based heuristic (~3 bytes/token for JSON) meant those numbers carried
roughly 3-4x more headroom in actual bytes than their literal value.
Interpreted as literal bytes post-rename, several fixtures no longer fit under
budget after truncation. Fixed by computing the real post-truncation byte size
for each affected fixture directly against the truncation functions and picking
thresholds with comfortable margin.
The `check_pot_drift_test.py` / babel extraction failure is unrelated and
pre-existing on `master`: the missing strings ("Screenshot cache is
unavailable", etc.) come from #44144 (already merged, one commit before this
branch), which added `gettext()` calls in `superset/dashboards/api.py` without
regenerating `messages.pot`. This branch touches neither file — confirmed via
`git diff` showing zero changes to either path between master and this branch.
--
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]
