You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It would be nice to display an evaluation of how close a failed rebuild is to success from the TUI summary view. The two main cases to support would be build failures and comparison failures. Build failures could use a coarse metric of how many steps were observed to run (tricky with nested commands like npx. might want to split these out and use set -eux) possibly with some log length metric. Comparison failures could be evaluated using ContentSummary for now and could be augmented with richer diffs.
The UI could be a two-segment progress bar (with a fancier structure than below), the first being build success and the second being comparison completeness:
npm foo 1.0.0 [o___|____]
npm bar 1.0.0 [oooo|o___]
Assuming this to be a monotonic indicator of success would have issues: With this approach, you could imagine a local maxima of a build that succeeds but is of a substantially lower complexity than is necessary. Still, it could be a good, quick, indicator of progress, especially if we disaggregate results by verdict (or have many verdicts for a single artifact/package).
The text was updated successfully, but these errors were encountered:
It would be nice to display an evaluation of how close a failed rebuild is to success from the TUI summary view. The two main cases to support would be build failures and comparison failures. Build failures could use a coarse metric of how many steps were observed to run (tricky with nested commands like npx. might want to split these out and use
set -eux
) possibly with some log length metric. Comparison failures could be evaluated using ContentSummary for now and could be augmented with richer diffs.The UI could be a two-segment progress bar (with a fancier structure than below), the first being build success and the second being comparison completeness:
Assuming this to be a monotonic indicator of success would have issues: With this approach, you could imagine a local maxima of a build that succeeds but is of a substantially lower complexity than is necessary. Still, it could be a good, quick, indicator of progress, especially if we disaggregate results by verdict (or have many verdicts for a single artifact/package).
The text was updated successfully, but these errors were encountered: