diff --git a/biome.json b/biome.json index 118207a2a..3b776b535 100644 --- a/biome.json +++ b/biome.json @@ -10,6 +10,7 @@ "includes": [ "**", "!samples", + "!vendor", "!.local-registry", "!packages/protocol/spec", "!docs/design/async-origin-capture/**" diff --git a/docs/PROGRESS.md b/docs/PROGRESS.md index ffe4b9bda..443e7af0f 100644 --- a/docs/PROGRESS.md +++ b/docs/PROGRESS.md @@ -1310,6 +1310,31 @@ recorded" was false. Still open from #120: WebKit/Firefox projects, the 22-secret smoke, `document.open`. #119 (other masking-option gaps on re-observed observers) is untouched. +### 2026-11-09 — Replay moves onto upstream rrweb 2.2.1 → `krassx/rrweb-2.2.1-a625` + +Fork `bugsee-port-2.2.1` (merge of rrweb-io/rrweb tag `rrweb@2.2.1` into `bugsee-port-2.2.0-iframe-block`; one +textual conflict in `rrweb-snapshot/src/snapshot.ts`: upstream #1964 gated the iframe-load branch with +`_isBlockedElement(n, blockClass, blockSelector)`; P4 already uses `!isBlockedNode` plus an +`_isBlockedInContext` re-check, so P4 was kept) → record-only bundle 198322 B / 62407 B gzip (2.2.0-iframe-block +987d0765: 188191 / 59348). Recipe unchanged: vite `dist/rrweb-record.js` via `vite.config.entries.js`, then +`esbuild --bundle --minify --format=esm`; two clean rebuilds byte-identical. + +2.2.1 record/snapshot notes and how they land here: + +- Media play/pause/seek on non-media (#1673): upstream used `instanceof HTMLMediaElement`, which is false for a + `