Repository navigation
Open
Labels
app/idPackage/app: id (pkgs/ or apps/ in this repo)Package/app: id (pkgs/ or apps/ in this repo)effort/lLarge: multi-day or open-ended; expect staged check-insLarge: multi-day or open-ended; expect staged check-inskind/featureNew feature or capability requestNew feature or capability requestmeta/enrichedIntake review done: acceptance criteria, links, and normalized labels addedIntake review done: acceptance criteria, links, and normalized labels addedmode/onlineFully doable remotely: design, config, code, cloudFully doable remotely: design, config, code, cloudnextOrg Project gate: queued next — orchestrator adds to the boardOrg Project gate: queued next — orchestrator adds to the boardpriority/lowNice to have; do when convenientNice to have; do when convenientsource/creatorFiled by the new-issue-creator jobFiled by the new-issue-creator jobtask/triageQueue packet kind: verify, reproduce, and route this itemQueue packet kind: verify, reproduce, and route this item
Description
From .opencode/plans/phase-12-platform-flexibility.md (workstream D, F12.11), which has no tracking issue. Phase 12 stack (PR #801).
What and why
Screen patches as cell runs, an optional binary encoding, optional compression, and resume tokens so a reconnecting client receives a patch, not a snapshot. Baseline from F11.9 (bytes per frame). Exit: the bandwidth target (under 10% of full-frame size), with a parity test that a patched client equals a full-frame client.
Dependencies
- F11.9 bytes-per-frame baseline.
- The presentation layer (
pkgs/id/src/world_presentation.rs,frames.rs).
Acceptance
- Cell-run screen patches with an optional binary encoding and compression.
- Resume tokens: reconnect gets a patch, not a snapshot.
- Parity test: patched client equals full-frame client.
Acceptance criteria
- The body's
## Acceptancebullets hold: cell-run screen patches ship with an optional binary encoding and compression, and resume tokens let a reconnecting client receive a patch instead of a snapshot. - The parity test exists and passes: a patched client ends byte-identical to a full-frame client.
- Bandwidth is measured against the F11.9 baseline and meets the under-10%-of-full-frame target.
Metadata
Metadata
Assignees
Labels
app/idPackage/app: id (pkgs/ or apps/ in this repo)Package/app: id (pkgs/ or apps/ in this repo)effort/lLarge: multi-day or open-ended; expect staged check-insLarge: multi-day or open-ended; expect staged check-inskind/featureNew feature or capability requestNew feature or capability requestmeta/enrichedIntake review done: acceptance criteria, links, and normalized labels addedIntake review done: acceptance criteria, links, and normalized labels addedmode/onlineFully doable remotely: design, config, code, cloudFully doable remotely: design, config, code, cloudnextOrg Project gate: queued next — orchestrator adds to the boardOrg Project gate: queued next — orchestrator adds to the boardpriority/lowNice to have; do when convenientNice to have; do when convenientsource/creatorFiled by the new-issue-creator jobFiled by the new-issue-creator jobtask/triageQueue packet kind: verify, reproduce, and route this itemQueue packet kind: verify, reproduce, and route this item
Activity
dezren39 commentedon Oct 26, 2026
Intake review
Related: Phase 12 stack PR #801 (merged); depends on the F11.9 bytes-per-frame baseline; presentation layer in
pkgs/id/src/world_presentation.rs/frames.rs. No merged PR resolves this (verified 2026-10-26).Notes: Performance feature with a measurable exit target (under 10% of full-frame size). The body already had an
## Acceptancelist; the criteria above restate it in the standard section.Risks: Patch/full-frame parity is easy to break subtly; the parity test is the acceptance gate, not an optional extra.
Suggested approach: Implement cell-run patches in the presentation layer first, prove parity, then add the optional binary encoding/compression behind the same tests.
dezren39 commentedon Oct 28, 2026
🤖 Agent note · picked up
dezren39 commentedon Oct 28, 2026
🤖 Agent note · done
e561ab58ff76ea8e84f2b9a1f108770ebe44a734Evidence (fresh GitHub reads at HEAD):
pkgs/id/src/frames.rs(226 lines) is length-prefixed transport framing for QUIC/SSH streams only — grep forpatch|delta|binary|resume|compress|snapshotreturns zero hits.pkgs/id/src/world_presentation.rs(595 lines) validates and emits fullkind: screenJSON documents — the same grep returns zero hits; there is no patch/delta representation..opencode/plans/phase-12-platform-flexibility.md;ResumeTokenappears nowhere.Route: #805 is effort/l overall, so it is staged, not sent as one fix. First slice enqueued as queue task 2659 (
small_fix, deduperepo-health:v1:fix:developing-today/code:issue:805): a transport-neutral cell-runScreenPatchinworld_presentation.rswith a parity test (patched client == full-frame client) and a size test (patch < 10% of full frame for a small change — the F12.11 exit target). Binary encoding, compression, resume tokens, and transport/session wiring remain on this issue as later slices. The issue stays open; this triage task completes astriaged.166 remaining items
dezren39 commentedon Nov 4, 2026
🤖 Agent note · handed off
opencodedezren39 commentedon Nov 4, 2026
🤖 Agent note · handed off
repo-worker-code, lanecode-o)opencodeviarepo-queue.py lane; handing off to the laptop opencode claimant, which launches the bound session.dezren39 commentedon Nov 4, 2026
🤖 Agent note · handed off
ses_e5e5da74c03ffeAlonqyVaaY3Q). Attempt 14/16.dezren39 commentedon Nov 4, 2026
🤖 Agent note · handed off
opencode; handed off to the laptop opencode claimant, which launches the bound session that drives task 5906 to its terminal state.dezren39 commentedon Nov 4, 2026
🤖 Agent note · handed off
developing-today commentedon Nov 4, 2026
🤖 Agent note · handed off
dezren39 commentedon Nov 4, 2026
🤖 Agent note · picked up
queue/5906-805-screenpatch-phase-2, no labels.origin/main, re-run the targeted id tests, then land the PR if verified.dezren39 commentedon Nov 4, 2026
🤖 Agent note · parked
dezren39(Drewry Pope); the runbook forbids touching Drew-owned PRs. No PR edits, pushes, merges, or review actions were made. I merged currentorigin/mainlocally only (fc16562); it is not pushed. Targeted test build is still running locally; prior packet records 66 world_session and 30 world_presentation tests passing on the previous PR head.dezren39 commentedon Nov 4, 2026
🤖 Agent note · correction (supersedes the park note above)
queue/5906-805-screenpatch-phase-2, i.e. this queue task's own branch, and the author/assigneedezren39is the shared fleet GitHub account that authors every queue PR — not evidence of personal Drew ownership. The PR carries noheld,user/owned,do-not-mergeordo-not-auto-editlabels, and this matches the earlier lane determination that the Drew-owned park was a misread. Work continues; no park or escalation.dezren39 commentedon Nov 4, 2026
🤖 Agent note · done (PR open, landing blocked by a red required check on main)
pr_opened— PR feat(id): resume tokens and bounded screen baselines — #805 Phase 2 #5036 is verified and up to date withorigin/mainat heada2933a00a; it is not merged because the merge queue's required check is red on main itself.cargo test -p id --lib world_session— 66 passed, 0 failedcargo test -p id-core world_presentation— 30 passed, 0 failedcargo test -p id --lib shell_contract— 8 passed, 0 failed (main's newdirectoryframe closes the old id(shell-contract): the directory client frame (PR #4954) is missing from the shell contract — contract test red on main #5020 gap)origin/main(d45cf2bd4) merged asfc1656227; shell-contract revision bumped 5 → 6 (a2933a00a) because main independently used revision 5 for thedirectoryframe while this branch used 5 for theresume_tokenfields.basic: proxy port fallbacksfails on clean mainffc19be78—.opencode/opencode.jsoncintentionally dropped theantigravity-directbaseURL after the design(mcpx): resume upstream HTTP sessions after a cold restart — persist the resumable session state #1414/opencode config: antigravity-direct repoint to native in-fork provider (opencode#1414) #5483 native-provider repoint, butcheck-proxy-port-fallbacks.shstill requires it. Filed as basic: proxy port fallbacks is red on main — the antigravity-direct guard row still requires an opencode.jsonc baseURL after the #1414 native-provider repoint #5497 with the one-line guard-row fix.basic: production line limitwas also cancelled runner-less twice (infra).