Repository navigation
Support tapbacks on older messages and exact emoji on macOS 27 - #159
Conversation
|
Thank you, Marcus. Reacting to older and off-screen messages through Tapback Details, with exact custom emoji, is the biggest parity gap on the macOS 27 gateway, and the discipline is good: emoji on bounded stdin, the Messages UI lock, one press and never a retry after an uncertain result. The generated Blocking
Should fix
Smaller notes: the What I will do. I will run this on the real macOS 27.2 gateway, on the self-thread only, once the above is in, because the open questions need a live Mac: whether Tapback Details opens the picker on 27.0 and 27.1, the accessibility identifier of the classic buttons, and how the 26 catalogue is laid out. Replies (#160) and this branch share Generated by Claude Code |
…ll 4x faster (CHANGELOG, CHANGE.log, roster, weekly notes) WHAT: CHANGELOG entries for #158 (Delivered until Read) and #161 (composer emoji picker) with the real end-to-end numbers for the poll speed-up (shallow 0.77 s to 0.2 s, deep 2.1 s to 1.0 s on gus); CHANGE.log entries for 401b99d, #158 and #161; Marcus Vorwaller added to the no-X table in CONTRIBUTORS-X.md (nothing on his GitHub profile or social accounts, his site marcus.pictures lists none, checked 10 Nov); WEEKLY-NOTES W46 section for 10 Nov with Marcus's own words about the rebases and the reasons #159 #160 #162 #163 were sent back. WHY: the PR-run checklist records every merge the day it lands, with the contributor's real name and a verified X handle or a visible "none found" plus asking Fred the same day. CONTEXT: independent review of #159 #160 #162 #163 (read-only, nothing sent to the Mac) found: #162 ffmpeg guessed the format from sender bytes (reproduced: a two-line ffconcat file delivered as a video rendered a different cached picture), #163 Qt plays bytes the safety check never saw and playback stops on any list rebuild, #160 three failure paths after Return report "nothing sent" and the recipient is never checked before Return (both confirmed against the code), #159 the shipped macOS 27 path has no test. Comments with the specifics are on each PR. Deployed 10 Nov to the gateway Mac (read_state.py and imsg, backups in ~/.blip/backups-20261110), gus and vic, all healthy, no log errors. TAGS: docs, changelog, change-log, weekly-notes, roster, marcus, pr-run, review LR-T: pr-run, contributor-pr, bookkeeping LR-D: blip, docs LR-K: pr-158-161-merged, pr-159-160-162-163-sent-back, marcus-no-x LR-R: 401b99d, ea64c98, 0d302b1 Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BXN1v67b9f9VHGe3FhaQuD
dad4c85 to
cb1972e
Compare
|
Addressed the review and rebased onto current main. The native classic-button path now has wrong-label, duplicate-button, lost-focus and changed-origin tests. Removed the unused JXA picker/catalogue implementation, picture flag and unused native opener seam. Below macOS 27, custom emoji use direct bubble actions only; the unproven catalogue path is gone. Picker-based production routes are gated to 27.2+ rather than assuming 27.0/27.1 support. Validation: 827 Bun tests, 347 synthetic Mac Python tests, and 7 Qt picker checks passed after rebase. No real conversations were opened or reactions sent. CI is running on the updated head. The live AX identifiers and older macOS behavior remain for your planned self-thread gateway validation; the version gate deliberately fails closed in the meantime. |
1cbae30 to
47874b6
Compare
… (CHANGELOG, CHANGE.log, roster, weekly notes, CLAUDE.md) WHAT: CHANGELOG and CHANGE.log entries for #162 (native stickers) and the reaction-scan speed-up (9b4f8d9) with the real numbers; CONTRIBUTORS-X.md moves Marcus Vorwaller to 3 merged (#158, #161, #162), still no X handle found and asked Fred; WEEKLY-NOTES W46 gets the #162 and speed-up bullets plus the note that #159, #160 and #163 got a second independent review of their rebased heads; CLAUDE.md's measured-latency paragraph gains the lesson from the reaction scan (the 90-day margin is real data: 3 of 28,305 reactions are stamped before their target, the worst by 0.83 days; an ORDER BY in SQL brings the full scan back; the date index reorders reactions unless sorted by ROWID; do not probe the guid index by shape because _strip_assoc is lenient on purpose). WHY: the PR-run checklist records every merge the day it lands, and the next agent that "optimises" enrich() must find the traps written down instead of rediscovering them. CONTEXT: Fred approved the #162 merge and the speed-up from the literal diffs on 10 Nov. Merge commit aee3394, speed-up 9b4f8d9. TAGS: docs, changelog, change-log, weekly-notes, roster, claude-md, pr-run, stickers, latency LR-T: pr-run, bookkeeping, latency LR-D: blip, docs LR-K: pr-162-merged, tapback-window-bound, marcus-no-x LR-R: aee3394, 9b4f8d9 Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BXN1v67b9f9VHGe3FhaQuD
|
Second-round review of 47874b6, and thank you: the original blocker is properly closed. An independent reviewer went through it read-only (nothing sent to a Mac); I re-checked the first item myself. Fixed and confirmed
Should-fix
Nits: Still to prove live on the 27.2 gateway, self-thread only, once this and the rebase are done: whether Heads-up for the rebase: #162 is merged and already carries the |
Use each bubble action and verify the exact picker target and database receipt. Keep payloads on bounded stdin, serialize Messages UI actions, press once and never retry uncertain outcomes. Include synthetic AX and QtTest coverage without changing the app style or version.
Cover classic native picker context and uniqueness guards, count reaction-label lookalikes, and preserve custom emoji kind on reads. Fail closed for unproven picker versions, preflight synthetic event access, and close both picker layers.
A custom emoji that resembles a classic tapback must retain kind=emoji so selecting it again removes it and settles pending state.
Exercise the synthetic bridge fixture alongside native reaction tests.
47874b6 to
5fc0454
Compare
…mission probe addresses System Events by PID and is bounded
WHAT: imsg-react's ae_permission() (AEDeterminePermissionToAutomateTarget,
askUserIfNeeded=false, the prompt-free "may sshd automate System Events?"
check that system_events_consent() asks before every tapback on macOS 27)
now builds a typeKernelProcessID ("kpid") descriptor from System Events'
pid (pgrep -x, no LaunchServices) instead of a bundle-id ("bund") one, and
runs the call on a daemon thread bounded at AE_PROBE_SECONDS = 5. An answer
that does not come is None -> "cannot confirm ... nothing was done", the
same refusal as any other unknown answer.
WHY: over ssh on macOS 27.2 the bundle-id form never returns (LaunchServices
is not available to ssh sessions on 27.2, the same fault the Pin helper
works around with pgrep). Every tapback then sat inside the Messages UI lock
for minutes before touching Messages, and the Mac-side process outlived the
caller. Found during the #159 live test (self-thread only, Fred approved):
faulthandler placed both hung runs in ae_permission <- system_events_consent
<- react, before any UI action, so nothing was pressed. On main since
5658760 (2026-10-07); tapbacks are off in this install's bridge.conf, which
is why nobody saw it.
VERIFIED on the gateway, read-only: bundle id still blocked at 15 s, kpid
answered 0 in 0.04 s; the fixed probe from this file answered granted in
0.15 s three times. With the same fix in #159's copy, a self-thread Heart on
a month-old message landed in 12.6 s (chat.db type 2000). Tests:
test_react.py 63 pass (3 new: a probe that never answers refuses in under
5 s; _bounded answers and errors; kpid, never bund); all 28 CI Python/Qt
suites pass; bun test 872 pass. Door one-way (bridge/mac/imsg-react), the
literal diff approved by Fred.
NOT YET: installed on the Mac. fnix went offline at 15:12 EST right after the
Heart (Tailscale offline, ssh times out); it gets this file once it is back.
LR-T: blip, bridge, imsg-react, tapbacks, macos-27, appleevents, tcc, hang, fix
LR-D: blip, bridge, macos
LR-K: ae-permission-kpid-not-bund, bounded-tcc-probe, launchservices-over-ssh-27-2
LR-R: 5658760
LR-L: law-3, law-10, law-11
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BXN1v67b9f9VHGe3FhaQuD
|
Round three on 5fc0454, and this time with the real macOS 27.2 gateway (self-thread only, with Fred's explicit OK). Thank you: every round-2 point is fixed in code, and I checked each one by breaking it. Round-2 points, fixed and guarded
Live on the 27.2 gateway
Small things, none blocking on their own
Once the reacted-bubble case works, I will run the remaining live steps (custom emoji on a fresh bubble, remove, a second reaction) and land it ahead of #160. Generated by Claude Code |
…s live test What the #159 self-thread test on the real gateway proved (Heart on an old message works, event-posting granted) and found (main's permission probe hung over ssh on 27.2, fixed in 33aab7b + 1a51bbe; #159 cannot add or remove a second reaction on a reacted bubble, sent back). LR-T: blip, weekly-notes, pr-159, tapbacks LR-D: blip, docs LR-K: weekly-w46-wed Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BXN1v67b9f9VHGe3FhaQuD
WHAT: the same-turn bookkeeping for the #159 merge (6a4fbce, 56bc294). - CHANGELOG.md: a proper Unreleased entry; the one-line note Marcus's rebase left above the "## Unreleased" heading is removed (it was orphaned there). - CHANGE.log: #159 entry with the live 27.2 proof and what Larry added. - CONTRIBUTORS-X.md: Marcus 5 merged (#158 #161 #162 #163 #159), #160 under review, X handle still none found (rechecked 2026-11-11, asked Fred). - WEEKLY-NOTES.md: W46 Wed 11 Nov, #159 bullet and the two Blip-side findings. NOT staged on purpose: bridge/mac/findmy_city.py and test_findmy_city.py are someone else's uncommitted work in this tree. LR-T: blip, pr-159, changelog, contributors, weekly-notes LR-D: blip, docs LR-K: pr-run-bookkeeping LR-R: 6a4fbce, 56bc294 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Merged as 6a4fbce. Thank you, Marcus: reacting to any message with any emoji on 27.2 is a real gap closed, and you fixed every point of my second round in the same evening. What I checked on 5fc0454, merged onto current main:
What I changed on top, in the open:
Two things I found on the way that were Blip's, not yours, both fixed on main: the permission check before any tapback hung forever over ssh on 27.2 when System Events was addressed by bundle id (33aab7b, now by pid and bounded), and the first reaction after a Mac reboot was refused because a just-launched System Events answers "not running" once (ef9a1ab). Small things I left, none blocking, yours if you want them: pressing Enter in the picker's search field sends the first match as a reaction ( #160 is next. Correction to a line I first wrote here: I said its two blockers still applied, and I had not re-checked its new head. I have now: Generated by Claude Code |
Adds older/off-screen message reactions and exact custom emoji, with a searchable reaction picker. Classic and custom reactions preserve their identity on reads, including custom emoji that resemble a classic tapback. The generated QML module matches its TypeScript source.
The Mac resolves target ordinals including literal reaction-annotation lookalikes, settles the bubble, validates picker context/focus and button uniqueness, and performs at most one selection. chat.db decides the result; uncertain gestures are never retried. Custom input is bounded stdin. Picker routes require macOS 27.2+; earlier versions may use direct bubble actions but never the unproven emoji catalogue path. Synthetic event access is preflighted without prompting, and cleanup dismisses both picker layers.
Based on current upstream main, including #158/#161, with upstream theme, controls and manifest preserved. The separate composer emoji feature is already upstream. The maintainer's live self-thread validation remains necessary for AX identifiers and platform behavior;
docs/reactions.mdrecords the limits.Validation: 827 Bun tests, 347 synthetic Mac Python tests and 7 Qt picker checks pass. The native classic path covers wrong labels, duplicate buttons, focus loss and changed origin; the custom read-path fixture runs in CI. GitHub CI passes. No real conversations opened or reactions sent.