Skip to content

Held the EMBER browse list back until the scan confirms a dataset holds video - #52

Merged
CodyCBakerPhD merged 1 commit into
mainfrom
claude/browse-ember-filtering-bug-gvlzpt
Aug 28, 2026
Merged

Held the EMBER browse list back until the scan confirms a dataset holds video#52
CodyCBakerPhD merged 1 commit into
mainfrom
claude/browse-ember-filtering-bug-gvlzpt

Conversation

@CodyCBakerPhD

Copy link
Copy Markdown
Member

Browse EMBER painted every dataset from the bucket listing the moment that listing landed, and only cut the list down to the datasets holding video once the archive sweep had read their file lists. The result is a flash of a list nobody asked for, most of which then disappears.

The list is now confirmed-only wherever a sweep is going to run: rows appear as the sweep lands each dataset that holds video, already carrying their video count, and until the first one arrives the list says it is still looking. An archive too large to sweep (SWEEP_BUDGET_BYTES) is unchanged, since nothing about it is confirmable up front. Sweep-driven redraws are coalesced to one a frame.

Adds an integration test that serves the manifests slowly so the window is wide enough to look inside; it fails on main.

Original prompt

when pressing browse EMBER button there is a brief moment where the loaded datasets are completely incorrect and they flash in that card content; by incorrect I mean they don't seem to filter empty or embargoed ahead of time


Generated by Claude Code

…ds video

Browse EMBER painted every dataset the bucket listing turned up as soon as
the listing landed, then cut the list down to the datasets holding video
once the archive sweep had read their file lists. On a warm cache that
whole window is short, so what a visitor sees is a flash of a list they
never asked for followed by most of it disappearing.

The list is now confirmed-only wherever a sweep is going to run: rows
appear as the sweep lands each dataset that actually holds video, already
carrying their video count, and until the first one arrives the list says
it is still looking. An archive too large to sweep is unchanged, since
nothing about it is confirmable up front. Redraws are coalesced to one a
frame so a scan of hundreds of datasets does not rebuild the list hundreds
of times.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CFBzpN6VGNZbWev9Ken8c6
@github-actions

github-actions Bot commented Aug 28, 2026

Copy link
Copy Markdown
PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-08-28 19:14 UTC

@CodyCBakerPhD CodyCBakerPhD self-assigned this Aug 28, 2026
@CodyCBakerPhD
CodyCBakerPhD marked this pull request as ready for review August 28, 2026 19:13
@CodyCBakerPhD
CodyCBakerPhD merged commit 5143eea into main Aug 28, 2026
15 checks passed
@CodyCBakerPhD
CodyCBakerPhD deleted the claude/browse-ember-filtering-bug-gvlzpt branch August 28, 2026 19:14
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants