2026.10.22 p2 - upgrade my yasb bar to: render icons (whose apps I am working on) on the timeline !

Timeline activity semantics, durable icons, renderer consolidation, and terminal glyph repair Stream window: 0:00:00 – 1:49:28 24-h (UTC+8): 02:56:24 – 04:45:52 ⏰ Timestamps  │   ├─ 🖼️ 0:00:00 YASB’s diagonally stacked workspace icons provided the visual reference. ⟦↗ till 0:10:21⟧   ├─ 🧩 0:02:18 Renderer analysis favored a unified preview layer while preserving the isolated main search UI. ⟦↗ till 0:17:17⟧  │   ├─ 🔍 0:08:06 Foreground-only timeline requirements revealed a producer-level data-contract gap. ⟦↗ till 0:15:36⟧  │　 ├─ ⚠️ 0:08:06 Microblock app counts could still include UI events lacking foreground metadata.  │　 └─ 🏗️ 0:08:06 The scope expanded toward a shared lower-layer foreground-activity projection. ⟦↗ till 0:15:36⟧  │   ├─ 📐 0:08:06 Workspace stack offsets and indices were inspected for the planned icon layout. ⟦↗ till 0:10:21⟧   ├─ 🔤 0:08:06 Terminal glyph diagnosis compared installed fonts, configuration, and renderer behavior. ⟦↗ till 0:35:24⟧   ├─ 📊 0:20:56 Codex session quota and reset telemetry fields were inspected. ⟦↗ till 0:25:55⟧   ├─ ✅ 0:31:40 The existing Tauri implementation was inspected and its Rust tests passed.   ├─ 🧠 0:32:24 Screenpipe focus semantics were checked before deriving engagement signals.  │   ├─ ⚙️ 0:36:55 Middle-stage architecture joined layered activity semantics, renderer consolidation, and versioned icon assets. ⟦↗ till 0:50:12⟧  │　 ├─ 📶 0:36:55 A layered activity model distinguished raw focus from derived engagement and passive presence. ⟦↗ till 0:40:16⟧  │　 │　 ├─ 🧾 0:36:55 Operating-system focus remained an observed fact, not proof of attention.  │　 │　 ├─ ⌨️ 0:36:55 Typing, scrolling, clicks, and clipboard actions were planned as bounded interaction bursts.  │　 │　 └─ 🪟 0:37:40 Engagement, foreground presence, and passive viewing became separate derived signals. ⟦↗ till 0:40:16⟧  │　 ├─ 🧱 0:37:40 The auxiliary UI was consolidated into one canvas using one fewer renderer. ⟦↗ till 0:46:26⟧  │　 └─ 🗂️ 0:49:52 A versioned manifest was proposed for content-addressed application icons. ⟦↗ till 0:50:12⟧  │   ├─ ✨ 0:50:12 A Windows Terminal fallback chain was applied and verified to restore missing glyphs. ⟦↗ till 0:51:26⟧   ├─ 🧪 0:53:22 A repeated baseline-versus-branch resource benchmark was prepared. ⟦↗ till 0:54:08⟧   ├─ 💾 0:58:04 Icon availability was bound to the timeline data that references it. ⟦↗ till 0:58:55⟧   ├─ 📦 1:02:34 Cold checkpoints were planned to bundle referenced icons with deterministic restoration.   ├─ 🛠️ 1:04:03 The terminal-font script was slated to preserve the tested fallback rule. ⟦↗ till 1:05:34⟧   ├─ 🌐 1:08:58 An asset repository was proposed to fill icons when applications first appear.   ├─ 🎨 1:08:58 Stack depth, activity significance, passive presence, and original alpha shaped icon rendering. ⟦↗ till 1:10:16⟧   ├─ 📝 1:10:16 Terminal font diagnostics were committed and pushed.   ├─ 🏎️ 1:11:11 Both release variants were rebuilt for comparative runtime sampling.   ├─ 🚫 1:14:20 A cold-start sample was rejected after lingering WebView2 helpers contaminated it.   ├─ 🛡️ 1:15:07 Icon resolution became non-blocking, concurrency-limited, retryable, and fallback-safe.   ├─ 🐛 1:15:52 A PowerShell reserved-variable collision was isolated to the sampling wrapper.  │   ├─ 🔀 1:16:39 Independent terminal maintenance closed alongside the late validation work. ⟦↗ till 1:20:42⟧  │　 └─ 🚀 1:16:39 Terminal fallback verified and pushed. ⟦↗ till 1:20:42⟧  │   ├─ 🪪 1:23:11 Durable Windows app identity. ⟦↗ till 1:33:43⟧  │　 ├─ 🌍 1:23:11 Ordinary pages stayed under browser identity while reliably identified PWAs could stand alone. ⟦↗ till 1:33:43⟧  │　 ├─ 🏷️ 1:23:11 Publisher, product, and original filename metadata defined stable Win32 identity. ⟦↗ till 1:29:22⟧  │　 └─ 🔗 1:23:11 Time-bounded evidence-backed aliases could upgrade identity without rewriting history.  │   ├─ ✅ 1:33:43 The first baseline resource round produced the expected process topology.   ├─ 🔎 1:38:56 Existing extraction code was checked for paths, AUMIDs, and shortcut identities.   ├─ 📈 1:39:41 Three switching-load rounds sampled CPU and GPU in matched windows. ⟦↗ till 1:39:51⟧   ├─ 🎯 1:41:07 Foreground and control-focus references refined identity evidence without flooding the graph. ⟦↗ till 1:42:51⟧   ├─ ⚖️ 1:44:47 zgui showed lower private memory but higher dedicated GPU use, without justifying migration.  │   └─ 🧱 1:46:59 The final timeline contract normalized evidence once into ranked application buckets.     　 ├─ 📏 1:46:59 Identity confidence was separated from retained metadata detail.     　 └─ 🪣 1:46:59 Stable ranked app buckets. #TimelineArchitecture #AppIdentity #YASB #Tauri #WindowsTerminal