app
Pup - the AI's own browser
Public Made by Adomby adom
pup gives your AI a real browser on your desktop that it drives completely, in the background, signed in as you: windows that never jump in front of your work, one shared profile that learns your logins, and every window labeled on the taskbar with its AI thread's name.
← Commit history
Fix cross-wired taskbar identity John caught (fusion and nb resolved to the SAME hwnd, so one thread stamped its AUMID onto the other's window and the loser stayed a grouped raw-Chrome button while a pup tile showed no overlay): the birth-time hwnd capture is now a SET-DIFF (snapshot window handles before createTarget, the NEW off-screen handle not present before and not already owned by any session is ours by construction) instead of a coordinate match whose 60px tolerance overlapped the 40px birth spacing; concurrent births disambiguate by the unique birthLeft with an 18px tolerance; and a hard guard rejects any handle already owned by another live session at every resolution path (birth, geometry, off-screen re-capture) so two windows can never share one handle
4 files changed
+47−20
adom-bridge-puppeteer-manifest.json+4−4@@ -1,10 +1,10 @@ { "manifest_version": 1, "name": "puppeteer",- "version": "2.0.33",- "url": "https://wiki.adom.inc/download/adom/pup-bridge/2.0.33/adom-bridge-puppeteer-v2.0.33.zip",- "sha256": "a1251e392468828ab0b55b13437b828836aa35a9d8a4d44db932c910afa3384f",- "size": 642310,+ "version": "2.0.34",+ "url": "https://wiki.adom.inc/download/adom/pup-bridge/2.0.34/adom-bridge-puppeteer-v2.0.34.zip",+ "sha256": "4ac706f11de9f032a54c318a793c9a1c5c120db56233ac08068181f0c28b072f",+ "size": 642959, "verbPrefixes": [ "browser_", "desktop_recorder_",
src/BRIDGE_VERSION+1−1@@ -1 +1 @@-2.0.33+2.0.34
src/bridge.json+1−1@@ -2,7 +2,7 @@ "manifest_version": 1, "name": "puppeteer", "displayName": "Puppeteer (Chrome for Testing + Native)",- "version": "2.0.33",+ "version": "2.0.34", "description": "Drives a browser ALREADY on the machine \u2014 installed Chrome, else Microsoft Edge (every Windows PC ships with Edge) \u2014 launched with a FRESH isolated profile, so the common case has NO ~150 MB Chrome-for-Testing download and it never touches the user's real tabs/logins. Binds its local server to LOOPBACK (127.0.0.1) only, so no Windows Firewall 'allow Node through the firewall' prompt ever pops. Chrome for Testing is fetched only as a LAST RESORT if the machine has no Chromium browser at all. SPAWN-VERIFIED fallback: pup tries each detected browser (Chrome \u2192 Edge \u2192 CfT) and actually launches it, falling through to the next if one won't start (and handling browsers like Edge that relaunch onto another PID), so a corrupt/partial install never hard-fails an open. The browser that launches is cached as the default (persisted across restarts) so it isn't re-probed each time. On spawn the bridge SELF-WARMS: detects browsers + disk, and only if the box has no Chrome/Edge and no cached CfT does it background-fetch CfT so the first open is instant. pup_use pins a specific browser (chrome | edge | cft | auto); {browser:'chrome',install:true} downloads & installs REAL Google Chrome (user-scoped, no admin) on an Edge-only box. pup_readiness (READ-ONLY) reports which browser will be used + the full candidate list + the default + free disk. Every failure returns a specific errorCode + _hint (chrome_install_no_disk, chrome_download_network, chrome_install_permission, chrome_stable_not_detected, \u2026). Launches with FULL browser permissions by default for full debug (strictPermissions:true opts out). Headful automation: open/close windows+tabs, navigate, screenshot, eval JS, record. pup is a GENERIC, non-profile browser only \u2014 every session is a fresh isolated Chrome/Edge process, never the user's real signed-in browser; profile defaults to sessionId. To drive the user's REAL Chrome/Edge (their profile, cookies, SSO), use the Adom browser EXTENSION's nbrowser_* verbs \u2014 a separate bridge, not pup's job. macOS (v1.9.74+): full parity with mac-native mechanics \u2014 background-by-default is the app-level hide (windows stay drivable/screenshottable while hidden), raise/lower map to unhide+activate/hide, pup_alert_window posts a native toast whose click focuses the window, pup_use {browser:'chrome',install:true} installs Chrome from the official dmg, and taskbar identity/badging remains Windows-only.", "homepage": "https://wiki.adom.inc/adom/adom-bridge", "author": "Adom Inc.",
src/server.js+41−14@@ -5348,6 +5348,19 @@ async function hwndOwnerPid(hwnd) { } // Confirm an hwnd is a LIVE window owned by a pup browser process. Used as the hard gate before any // taskbar paint/flash, and to invalidate a stale cached hwnd (recycled numbers).+// v2.0.34 (John caught fusion + nb resolving to the SAME hwnd): a window handle can belong to+// exactly ONE session. This rejects a candidate handle already claimed by another live session's+// _hwnd or birth __pupHwnd, so concurrent births/resolves on the shared profile can't cross-wire+// (the bug where one thread stamped its identity onto another thread's taskbar button).+function hwndOwnedByOtherSession(hwnd, exceptSessionId) {+ const h = Number(hwnd);+ for (const [k, s] of sessions) {+ if (k === exceptSessionId || !s) continue;+ if (Number(s._hwnd) === h) return true;+ for (const t of (s.tabs || [])) { if (t && t.page && Number(t.page.__pupHwnd) === h) return true; }+ }+ return false;+} async function hwndBelongsToPup(hwnd, session) { // v1.9.287 (John caught two sessions on DIFFERENT profiles sharing one hwnd): validate that the // hwnd belongs to THIS session's OWN profile process, not merely "some pup window". The old check@@ -5455,7 +5468,12 @@ async function resolveSessionHwndByBounds(session, sessionId) { try { const p0 = (session.tabs || [])[0] && (session.tabs || [])[0].page; const bh = p0 && p0.__pupHwnd;- if (bh && await hwndBelongsToPup(Number(bh), session)) {+ if (bh && hwndOwnedByOtherSession(Number(bh), sessionId)) {+ // v2.0.34: another live session already owns this handle — our birth capture cross-wired.+ // Drop it and fall through so we resolve a DISTINCT window instead of stamping onto theirs.+ console.log(`[hwnd] "${sessionId}" birth handle ${bh} is owned by ANOTHER session — discarding`);+ try { p0.__pupHwnd = null; } catch (e) {}+ } else if (bh && await hwndBelongsToPup(Number(bh), session)) { session._hwnd = Number(bh); persistSessionHwnd(sessionId, Number(bh)); console.log(`[hwnd] "${sessionId}" resolved OS handle ${bh} — birth-time capture (off-screen at creation)`); return Number(bh);@@ -5547,6 +5565,7 @@ async function resolveSessionHwndByBounds(session, sessionId) { let resolved = null, how = ''; if (!_geomBackoff && matches.size === 1) { resolved = [...matches][0]; how = markerW != null ? 'unique-width wiggle' : 'window bounds'; } else if (!_geomBackoff && _profWins.size === 1) { resolved = [..._profWins][0]; how = 'sole window of profile process (wiggle-free)'; }+ if (resolved && hwndOwnedByOtherSession(resolved, sessionId)) { console.log(`[hwnd] "${sessionId}" geometry picked ${resolved} but another session owns it — refusing`); resolved = null; } if (resolved) { session._hwnd = resolved; persistSessionHwnd(sessionId, resolved);@@ -5576,7 +5595,7 @@ async function resolveSessionHwndByBounds(session, sessionId) { let got = null; for (let i = 0; i < 8 && !got; i++) { const rr = await enumChromeWidgetWindows();- const m = rr.find(r => Math.abs(r.T - Math.round(-33000)) <= 300 && [1, 1.25, 1.5, 1.75, 2, 0.8].some(sc => Math.abs(r.L - Math.round(probeLeft * sc)) <= 80));+ const m = rr.find(r => !hwndOwnedByOtherSession(r.h, sessionId) && Math.abs(r.T - Math.round(-33000)) <= 300 && [1, 1.25, 1.5, 1.75, 2].some(sc => Math.abs(r.L - Math.round(probeLeft * sc)) <= 40)); if (m) got = m.h; else await new Promise(r => setTimeout(r, 120)); } // restore on-screen (z-bottom park handles final placement); use original size if we had it@@ -7504,6 +7523,10 @@ async function createPageInNewWindow(browser, profileName) { const birthLeft = -32000 - ((_birthSeq++ % 300) * 40); try { const before = new Set(await browser.pages());+ // v2.0.34: snapshot existing OS window handles BEFORE creating, so the SET-DIFF below identifies+ // the new window unambiguously — coordinate matching alone collided concurrent births.+ let _beforeHwnds = new Set();+ if (process.platform === 'win32') { try { _beforeHwnds = new Set((await enumChromeWidgetWindows()).map(r => Number(r.h))); } catch (e) {} } cdp = await browser.target().createCDPSession(); // v1.9.144 (John: "you keep opening pup windows on top of the one i'm working in") — birth // the window OFF-SCREEN so the moment between creation and the park is invisible. 1.9.135@@ -7529,21 +7552,25 @@ async function createPageInNewWindow(browser, profileName) { // window of this profile at ~birthLeft" is unambiguous. Stash it on the page for the resolver. if (fresh && process.platform === 'win32' && profileName) { try {- // v2.0.31: match by the window's UNIQUE off-screen BIRTH POSITION, not by pid membership.- // pup births every new window at top:-32000 and a unique birthLeft — no other app parks- // there — so that coordinate IS the identity. The old pid filter (pupBrowserPidSet) came up- // empty for the shared adom-you Chrome process (its real pid is often untracked), so birth- // capture silently failed and the window fell through to the geometry heuristic that cannot- // disambiguate a crowded profile — the raw-Chrome-icon bug John caught.+ // v2.0.34: SET-DIFF birth capture. The window that just APPEARED in the birth zone (off-screen,+ // and NOT present before createTarget, and not already owned by another session) is ours by+ // construction. If concurrency put several new windows there at once, disambiguate by MY unique+ // birthLeft with a TIGHT tolerance (births are 40px apart, so >18px would cross-wire — the+ // fusion+nb-shared-hwnd bug). No coordinate-only match, no pid filter. let picked = null;- for (let i = 0; i < 12 && !picked; i++) {+ for (let i = 0; i < 15 && !picked; i++) { const rows = await enumChromeWidgetWindows();- const off = rows.filter(r => r.T < -10000); // pup's birth zone (top:-32000)- for (const sc of [1, 1.25, 1.5, 1.75, 2, 0.8, 1.4]) { const m = off.find(r => Math.abs(r.L - Math.round(birthLeft * sc)) <= 60); if (m) { picked = m.h; break; } }- if (!picked && off.length === 1) picked = off[0].h;- if (!picked) await new Promise(r => setTimeout(r, 150));+ const newOff = rows.filter(r => r.T < -10000 && !_beforeHwnds.has(Number(r.h)) && !hwndOwnedByOtherSession(r.h, null));+ if (newOff.length === 1) picked = newOff[0].h;+ else if (newOff.length > 1) {+ let best = null, bestErr = 1e9;+ for (const r of newOff) for (const sc of [1, 1.25, 1.5, 1.75, 2]) { const e = Math.abs(r.L - Math.round(birthLeft * sc)); if (e < bestErr) { bestErr = e; best = r.h; } }+ if (best && bestErr <= 18) picked = best; // tight: unique birthLeft, no adjacent collision+ }+ if (!picked) await new Promise(r => setTimeout(r, 120)); }- if (picked) { fresh.__pupHwnd = picked; console.log(`[hwnd] birth-resolved new window on "${profileName}" -> ${picked} (off-screen L≈${birthLeft})`); }+ if (picked) { fresh.__pupHwnd = picked; console.log(`[hwnd] birth-resolved new window on "${profileName}" -> ${picked} (off-screen L≈${birthLeft}, set-diff)`); }+ else console.log(`[hwnd] birth capture on "${profileName}" left AMBIGUOUS (concurrent births) — will resolve later`); } catch (e) {} } try {