fix(desktop): show newest changes in nightly previews

[t3dotgg]

Nightly update previews showed the oldest changes in each GitHub release. Contributor credits could also consume preview slots, so recent changes were often missing.

This change filters release boilerplate, keeps the newest eight changes, and reports the full change and release counts. The sidebar preview now links to each GitHub release and uses a focus-managed popover for pointer and keyboard access.

Before	After

Watch the pointer and keyboard interaction. The 2.5-second excerpt is slowed to 2.5x duration so the focus changes are easier to see.

Verification:

• 86 focused desktop and web tests passed.
• Contracts, desktop, and web typechecks passed.
• Changed-file lint, formatting, and git diff --check passed.
• The affected public nightly release produced the newest eight of ten changes and excluded contributor credits.
• Markdown and HTML section headings are excluded from the reported change count.
• The real app in Chromium passed pointer travel and scrolling, release-link activation, keyboard entry and exit, Escape focus return, download, install cancellation and confirmation, and tooltip fallback checks. The test used a stubbed desktop bridge, so it did not download or install a real update.
• The same update button DOM node kept keyboard focus through idle, checking, nightly available, no-notes, and latest state changes. These flows passed with React Compiler enabled.

This continues #8588. Theo Browne authored its three original commits. This replacement preserves his authorship and adds the review fixes separately.

Made with GPT-5.6 Sol in the Codex harness.

---

Note

Medium Risk
IPC contract changes require totalItems and omittedReleaseCount on update state, which can break older clients or persisted snapshots. Popover focus and React Compiler opt-out add UI regression surface on the nightly update control.

Overview
Nightly desktop update previews now surface the newest changes per release instead of the oldest, and stop treating contributor credits and full-changelog boilerplate as real changes. Normalization returns up to eight items and six release groups, with totalItems per group and omittedReleaseCount when older releases are dropped; those fields flow through IPC and the desktop update reducer.

The sidebar replaces the cramped nightly tooltip with a hover/focus popover (SidebarUpdateReleaseNotes) that links each preview to its GitHub release tag, labels hidden changes within a release, and links omitted releases to the releases history page. Popover behavior keeps the update button as the action target (no double-toggle on click), supports forward Tab into the panel, and restores focus sensibly on Escape.

User docs add a short section describing the nightly preview limits and links.

^{Reviewed by Cursor Bugbot for commit d90cf5f70eb63caae8fc2e2940d1a91f3e8f8d98. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Show newest nightly release changes in sidebar popover with per-release links

• Overhauls release-note extraction to keep the newest eight changes per release in newest-first order, excluding contributor and changelog sections, and returns up to six non-empty release groups plus an omittedReleaseCount
• Replaces the sidebar tooltip with a hover- and focus-aware Popover that renders nightly notes via the new SidebarUpdateReleaseNotes component, while keeping the update button as the install action
• SidebarUpdateReleaseNotes links each displayed release to its exact GitHub tag, reports omitted changes per release via totalItems, and links omitted release groups to the GitHub release history page
• Adds required omittedReleaseCount to DesktopUpdateState and totalItems to DesktopUpdateReleaseNote in the IPC contracts and runtime schema
• Risk: omittedReleaseCount and totalItems are now required fields in ipc.ts; any out-of-tree desktop clients reading the update-state payload must supply or tolerate these fields

^{Macroscope summarized d99162b.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.4 KiB	13.1 KiB	−320 B (−2.3%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.9 KiB	6.9 KiB	−17 B (−0.2%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.2 KiB	−303 B (−4.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	54.7 KiB	−2.3 KiB (−4.1%)	66.4 KiB	✅
Codex	Live turn messages	10	8	−2 (−20.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−18 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.9 KiB	6.9 KiB	+1 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.6 KiB	6.6 KiB	−19 B (−0.3%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: ea71a19 · PR result: d99162b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 109.4 KiB
• Claude decoded thread snapshot: 110.1 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR changes production desktop update state and IPC contracts while replacing the nightly tooltip with a focus-managed, hover-enabled release-notes popover and new external-link workflow. The interaction changes introduce meaningful focus and state-transition risk that should receive human review.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d99162b. Configure here.}

[t3dotgg]

Nightly update previews showed the oldest changes in each GitHub release. Contributor credits could also consume preview slots, so recent changes were often missing.

This change filters release boilerplate, keeps the newest eight changes, and reports the full change and release counts. The sidebar preview now links to each GitHub release and uses a focus-managed popover for pointer and keyboard access.

Before	After

Watch the pointer and keyboard interaction. The 2.5-second excerpt is slowed to 2.5x duration so the focus changes are easier to see.

Verification:

• 86 focused desktop and web tests passed.
• Contracts, desktop, and web typechecks passed.
• Changed-file lint, formatting, and git diff --check passed.
• The affected public nightly release produced the newest eight of ten changes and excluded contributor credits.
• Markdown and HTML section headings are excluded from the reported change count.
• The real app in Chromium passed pointer travel and scrolling, release-link activation, keyboard entry and exit, Escape focus return, download, install cancellation and confirmation, and tooltip fallback checks. The test used a stubbed desktop bridge, so it did not download or install a real update.
• The same update button DOM node kept keyboard focus through idle, checking, nightly available, no-notes, and latest state changes. These flows passed with React Compiler enabled.

This continues #8588. Theo Browne authored its three original commits. This replacement preserves his authorship and adds the review fixes separately.

Made with GPT-5.6 Sol in the Codex harness.

---

Note

Medium Risk
IPC contract changes require totalItems and omittedReleaseCount on update state, which can break older clients or persisted snapshots. Popover focus and React Compiler opt-out add UI regression surface on the nightly update control.

Overview
Nightly desktop update previews now surface the newest changes per release instead of the oldest, and stop treating contributor credits and full-changelog boilerplate as real changes. Normalization returns up to eight items and six release groups, with totalItems per group and omittedReleaseCount when older releases are dropped; those fields flow through IPC and the desktop update reducer.

The sidebar replaces the cramped nightly tooltip with a hover/focus popover (SidebarUpdateReleaseNotes) that links each preview to its GitHub release tag, labels hidden changes within a release, and links omitted releases to the releases history page. Popover behavior keeps the update button as the action target (no double-toggle on click), supports forward Tab into the panel, and restores focus sensibly on Escape.

User docs add a short section describing the nightly preview limits and links.

^{Reviewed by Cursor Bugbot for commit d90cf5f70eb63caae8fc2e2940d1a91f3e8f8d98. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Show newest nightly release changes in sidebar popover with per-release links

• Overhauls release-note extraction to keep the newest eight changes per release in newest-first order, excluding contributor and changelog sections, and returns up to six non-empty release groups plus an omittedReleaseCount
• Replaces the sidebar tooltip with a hover- and focus-aware Popover that renders nightly notes via the new SidebarUpdateReleaseNotes component, while keeping the update button as the install action
• SidebarUpdateReleaseNotes links each displayed release to its exact GitHub tag, reports omitted changes per release via totalItems, and links omitted release groups to the GitHub release history page
• Adds required omittedReleaseCount to DesktopUpdateState and totalItems to DesktopUpdateReleaseNote in the IPC contracts and runtime schema
• Risk: omittedReleaseCount and totalItems are now required fields in ipc.ts; any out-of-tree desktop clients reading the update-state payload must supply or tolerate these fields

^{Macroscope summarized d99162b.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.4 KiB	13.1 KiB	−320 B (−2.3%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.9 KiB	6.9 KiB	−17 B (−0.2%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.2 KiB	−303 B (−4.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	54.7 KiB	−2.3 KiB (−4.1%)	66.4 KiB	✅
Codex	Live turn messages	10	8	−2 (−20.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−18 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.9 KiB	6.9 KiB	+1 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.6 KiB	6.6 KiB	−19 B (−0.3%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: ea71a19 · PR result: d99162b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 109.4 KiB
• Claude decoded thread snapshot: 110.1 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR changes production desktop update state and IPC contracts while replacing the nightly tooltip with a focus-managed, hover-enabled release-notes popover and new external-link workflow. The interaction changes introduce meaningful focus and state-transition risk that should receive human review.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d99162b. Configure here.}

[t3dotgg]

Nightly update previews showed the oldest changes in each GitHub release. Contributor credits could also consume preview slots, so recent changes were often missing.

This change filters release boilerplate, keeps the newest eight changes, and reports the full change and release counts. The sidebar preview now links to each GitHub release and uses a focus-managed popover for pointer and keyboard access.

Before	After

Watch the pointer and keyboard interaction. The 2.5-second excerpt is slowed to 2.5x duration so the focus changes are easier to see.

Verification:

• 86 focused desktop and web tests passed.
• Contracts, desktop, and web typechecks passed.
• Changed-file lint, formatting, and git diff --check passed.
• The affected public nightly release produced the newest eight of ten changes and excluded contributor credits.
• Markdown and HTML section headings are excluded from the reported change count.
• The real app in Chromium passed pointer travel and scrolling, release-link activation, keyboard entry and exit, Escape focus return, download, install cancellation and confirmation, and tooltip fallback checks. The test used a stubbed desktop bridge, so it did not download or install a real update.
• The same update button DOM node kept keyboard focus through idle, checking, nightly available, no-notes, and latest state changes. These flows passed with React Compiler enabled.

This continues #8588. Theo Browne authored its three original commits. This replacement preserves his authorship and adds the review fixes separately.

Made with GPT-5.6 Sol in the Codex harness.

---

Note

Medium Risk
IPC contract changes require totalItems and omittedReleaseCount on update state, which can break older clients or persisted snapshots. Popover focus and React Compiler opt-out add UI regression surface on the nightly update control.

Overview
Nightly desktop update previews now surface the newest changes per release instead of the oldest, and stop treating contributor credits and full-changelog boilerplate as real changes. Normalization returns up to eight items and six release groups, with totalItems per group and omittedReleaseCount when older releases are dropped; those fields flow through IPC and the desktop update reducer.

The sidebar replaces the cramped nightly tooltip with a hover/focus popover (SidebarUpdateReleaseNotes) that links each preview to its GitHub release tag, labels hidden changes within a release, and links omitted releases to the releases history page. Popover behavior keeps the update button as the action target (no double-toggle on click), supports forward Tab into the panel, and restores focus sensibly on Escape.

User docs add a short section describing the nightly preview limits and links.

^{Reviewed by Cursor Bugbot for commit d90cf5f70eb63caae8fc2e2940d1a91f3e8f8d98. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Show newest nightly release changes in sidebar popover with per-release links

• Overhauls release-note extraction to keep the newest eight changes per release in newest-first order, excluding contributor and changelog sections, and returns up to six non-empty release groups plus an omittedReleaseCount
• Replaces the sidebar tooltip with a hover- and focus-aware Popover that renders nightly notes via the new SidebarUpdateReleaseNotes component, while keeping the update button as the install action
• SidebarUpdateReleaseNotes links each displayed release to its exact GitHub tag, reports omitted changes per release via totalItems, and links omitted release groups to the GitHub release history page
• Adds required omittedReleaseCount to DesktopUpdateState and totalItems to DesktopUpdateReleaseNote in the IPC contracts and runtime schema
• Risk: omittedReleaseCount and totalItems are now required fields in ipc.ts; any out-of-tree desktop clients reading the update-state payload must supply or tolerate these fields

^{Macroscope summarized d99162b.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.4 KiB	13.1 KiB	−320 B (−2.3%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.9 KiB	6.9 KiB	−17 B (−0.2%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.2 KiB	−303 B (−4.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	54.7 KiB	−2.3 KiB (−4.1%)	66.4 KiB	✅
Codex	Live turn messages	10	8	−2 (−20.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−18 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.9 KiB	6.9 KiB	+1 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.6 KiB	6.6 KiB	−19 B (−0.3%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: ea71a19 · PR result: d99162b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 109.4 KiB
• Claude decoded thread snapshot: 110.1 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR changes production desktop update state and IPC contracts while replacing the nightly tooltip with a focus-managed, hover-enabled release-notes popover and new external-link workflow. The interaction changes introduce meaningful focus and state-transition risk that should receive human review.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d99162b. Configure here.}

[t3dotgg]

Nightly update previews showed the oldest changes in each GitHub release. Contributor credits could also consume preview slots, so recent changes were often missing.

This change filters release boilerplate, keeps the newest eight changes, and reports the full change and release counts. The sidebar preview now links to each GitHub release and uses a focus-managed popover for pointer and keyboard access.

Before	After

Watch the pointer and keyboard interaction. The 2.5-second excerpt is slowed to 2.5x duration so the focus changes are easier to see.

Verification:

• 86 focused desktop and web tests passed.
• Contracts, desktop, and web typechecks passed.
• Changed-file lint, formatting, and git diff --check passed.
• The affected public nightly release produced the newest eight of ten changes and excluded contributor credits.
• Markdown and HTML section headings are excluded from the reported change count.
• The real app in Chromium passed pointer travel and scrolling, release-link activation, keyboard entry and exit, Escape focus return, download, install cancellation and confirmation, and tooltip fallback checks. The test used a stubbed desktop bridge, so it did not download or install a real update.
• The same update button DOM node kept keyboard focus through idle, checking, nightly available, no-notes, and latest state changes. These flows passed with React Compiler enabled.

This continues #8588. Theo Browne authored its three original commits. This replacement preserves his authorship and adds the review fixes separately.

Made with GPT-5.6 Sol in the Codex harness.

---

Note

Medium Risk
IPC contract changes require totalItems and omittedReleaseCount on update state, which can break older clients or persisted snapshots. Popover focus and React Compiler opt-out add UI regression surface on the nightly update control.

Overview
Nightly desktop update previews now surface the newest changes per release instead of the oldest, and stop treating contributor credits and full-changelog boilerplate as real changes. Normalization returns up to eight items and six release groups, with totalItems per group and omittedReleaseCount when older releases are dropped; those fields flow through IPC and the desktop update reducer.

The sidebar replaces the cramped nightly tooltip with a hover/focus popover (SidebarUpdateReleaseNotes) that links each preview to its GitHub release tag, labels hidden changes within a release, and links omitted releases to the releases history page. Popover behavior keeps the update button as the action target (no double-toggle on click), supports forward Tab into the panel, and restores focus sensibly on Escape.

User docs add a short section describing the nightly preview limits and links.

^{Reviewed by Cursor Bugbot for commit d90cf5f70eb63caae8fc2e2940d1a91f3e8f8d98. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Show newest nightly release changes in sidebar popover with per-release links

• Overhauls release-note extraction to keep the newest eight changes per release in newest-first order, excluding contributor and changelog sections, and returns up to six non-empty release groups plus an omittedReleaseCount
• Replaces the sidebar tooltip with a hover- and focus-aware Popover that renders nightly notes via the new SidebarUpdateReleaseNotes component, while keeping the update button as the install action
• SidebarUpdateReleaseNotes links each displayed release to its exact GitHub tag, reports omitted changes per release via totalItems, and links omitted release groups to the GitHub release history page
• Adds required omittedReleaseCount to DesktopUpdateState and totalItems to DesktopUpdateReleaseNote in the IPC contracts and runtime schema
• Risk: omittedReleaseCount and totalItems are now required fields in ipc.ts; any out-of-tree desktop clients reading the update-state payload must supply or tolerate these fields

^{Macroscope summarized d99162b.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.4 KiB	13.1 KiB	−320 B (−2.3%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.9 KiB	6.9 KiB	−17 B (−0.2%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.2 KiB	−303 B (−4.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	54.7 KiB	−2.3 KiB (−4.1%)	66.4 KiB	✅
Codex	Live turn messages	10	8	−2 (−20.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−18 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.9 KiB	6.9 KiB	+1 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.6 KiB	6.6 KiB	−19 B (−0.3%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: ea71a19 · PR result: d99162b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 109.4 KiB
• Claude decoded thread snapshot: 110.1 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR changes production desktop update state and IPC contracts while replacing the nightly tooltip with a focus-managed, hover-enabled release-notes popover and new external-link workflow. The interaction changes introduce meaningful focus and state-transition risk that should receive human review.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d99162b. Configure here.}

[t3dotgg]

Nightly update previews showed the oldest changes in each GitHub release. Contributor credits could also consume preview slots, so recent changes were often missing.

This change filters release boilerplate, keeps the newest eight changes, and reports the full change and release counts. The sidebar preview now links to each GitHub release and uses a focus-managed popover for pointer and keyboard access.

Before	After

Watch the pointer and keyboard interaction. The 2.5-second excerpt is slowed to 2.5x duration so the focus changes are easier to see.

Verification:

• 86 focused desktop and web tests passed.
• Contracts, desktop, and web typechecks passed.
• Changed-file lint, formatting, and git diff --check passed.
• The affected public nightly release produced the newest eight of ten changes and excluded contributor credits.
• Markdown and HTML section headings are excluded from the reported change count.
• The real app in Chromium passed pointer travel and scrolling, release-link activation, keyboard entry and exit, Escape focus return, download, install cancellation and confirmation, and tooltip fallback checks. The test used a stubbed desktop bridge, so it did not download or install a real update.
• The same update button DOM node kept keyboard focus through idle, checking, nightly available, no-notes, and latest state changes. These flows passed with React Compiler enabled.

This continues #8588. Theo Browne authored its three original commits. This replacement preserves his authorship and adds the review fixes separately.

Made with GPT-5.6 Sol in the Codex harness.

---

Note

Medium Risk
IPC contract changes require totalItems and omittedReleaseCount on update state, which can break older clients or persisted snapshots. Popover focus and React Compiler opt-out add UI regression surface on the nightly update control.

Overview
Nightly desktop update previews now surface the newest changes per release instead of the oldest, and stop treating contributor credits and full-changelog boilerplate as real changes. Normalization returns up to eight items and six release groups, with totalItems per group and omittedReleaseCount when older releases are dropped; those fields flow through IPC and the desktop update reducer.

The sidebar replaces the cramped nightly tooltip with a hover/focus popover (SidebarUpdateReleaseNotes) that links each preview to its GitHub release tag, labels hidden changes within a release, and links omitted releases to the releases history page. Popover behavior keeps the update button as the action target (no double-toggle on click), supports forward Tab into the panel, and restores focus sensibly on Escape.

User docs add a short section describing the nightly preview limits and links.

^{Reviewed by Cursor Bugbot for commit d90cf5f70eb63caae8fc2e2940d1a91f3e8f8d98. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Show newest nightly release changes in sidebar popover with per-release links

• Overhauls release-note extraction to keep the newest eight changes per release in newest-first order, excluding contributor and changelog sections, and returns up to six non-empty release groups plus an omittedReleaseCount
• Replaces the sidebar tooltip with a hover- and focus-aware Popover that renders nightly notes via the new SidebarUpdateReleaseNotes component, while keeping the update button as the install action
• SidebarUpdateReleaseNotes links each displayed release to its exact GitHub tag, reports omitted changes per release via totalItems, and links omitted release groups to the GitHub release history page
• Adds required omittedReleaseCount to DesktopUpdateState and totalItems to DesktopUpdateReleaseNote in the IPC contracts and runtime schema
• Risk: omittedReleaseCount and totalItems are now required fields in ipc.ts; any out-of-tree desktop clients reading the update-state payload must supply or tolerate these fields

^{Macroscope summarized d99162b.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.4 KiB	13.1 KiB	−320 B (−2.3%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.9 KiB	6.9 KiB	−17 B (−0.2%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.2 KiB	−303 B (−4.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	54.7 KiB	−2.3 KiB (−4.1%)	66.4 KiB	✅
Codex	Live turn messages	10	8	−2 (−20.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−18 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.9 KiB	6.9 KiB	+1 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.6 KiB	6.6 KiB	−19 B (−0.3%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: ea71a19 · PR result: d99162b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 109.4 KiB
• Claude decoded thread snapshot: 110.1 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR changes production desktop update state and IPC contracts while replacing the nightly tooltip with a focus-managed, hover-enabled release-notes popover and new external-link workflow. The interaction changes introduce meaningful focus and state-transition risk that should receive human review.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d99162b. Configure here.}

[t3dotgg]

Nightly update previews showed the oldest changes in each GitHub release. Contributor credits could also consume preview slots, so recent changes were often missing.

This change filters release boilerplate, keeps the newest eight changes, and reports the full change and release counts. The sidebar preview now links to each GitHub release and uses a focus-managed popover for pointer and keyboard access.

Before	After

Watch the pointer and keyboard interaction. The 2.5-second excerpt is slowed to 2.5x duration so the focus changes are easier to see.

Verification:

• 86 focused desktop and web tests passed.
• Contracts, desktop, and web typechecks passed.
• Changed-file lint, formatting, and git diff --check passed.
• The affected public nightly release produced the newest eight of ten changes and excluded contributor credits.
• Markdown and HTML section headings are excluded from the reported change count.
• The real app in Chromium passed pointer travel and scrolling, release-link activation, keyboard entry and exit, Escape focus return, download, install cancellation and confirmation, and tooltip fallback checks. The test used a stubbed desktop bridge, so it did not download or install a real update.
• The same update button DOM node kept keyboard focus through idle, checking, nightly available, no-notes, and latest state changes. These flows passed with React Compiler enabled.

This continues #8588. Theo Browne authored its three original commits. This replacement preserves his authorship and adds the review fixes separately.

Made with GPT-5.6 Sol in the Codex harness.

---

Note

Medium Risk
IPC contract changes require totalItems and omittedReleaseCount on update state, which can break older clients or persisted snapshots. Popover focus and React Compiler opt-out add UI regression surface on the nightly update control.

Overview
Nightly desktop update previews now surface the newest changes per release instead of the oldest, and stop treating contributor credits and full-changelog boilerplate as real changes. Normalization returns up to eight items and six release groups, with totalItems per group and omittedReleaseCount when older releases are dropped; those fields flow through IPC and the desktop update reducer.

The sidebar replaces the cramped nightly tooltip with a hover/focus popover (SidebarUpdateReleaseNotes) that links each preview to its GitHub release tag, labels hidden changes within a release, and links omitted releases to the releases history page. Popover behavior keeps the update button as the action target (no double-toggle on click), supports forward Tab into the panel, and restores focus sensibly on Escape.

User docs add a short section describing the nightly preview limits and links.

^{Reviewed by Cursor Bugbot for commit d90cf5f70eb63caae8fc2e2940d1a91f3e8f8d98. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Show newest nightly release changes in sidebar popover with per-release links

• Overhauls release-note extraction to keep the newest eight changes per release in newest-first order, excluding contributor and changelog sections, and returns up to six non-empty release groups plus an omittedReleaseCount
• Replaces the sidebar tooltip with a hover- and focus-aware Popover that renders nightly notes via the new SidebarUpdateReleaseNotes component, while keeping the update button as the install action
• SidebarUpdateReleaseNotes links each displayed release to its exact GitHub tag, reports omitted changes per release via totalItems, and links omitted release groups to the GitHub release history page
• Adds required omittedReleaseCount to DesktopUpdateState and totalItems to DesktopUpdateReleaseNote in the IPC contracts and runtime schema
• Risk: omittedReleaseCount and totalItems are now required fields in ipc.ts; any out-of-tree desktop clients reading the update-state payload must supply or tolerate these fields

^{Macroscope summarized d99162b.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.4 KiB	13.1 KiB	−320 B (−2.3%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.9 KiB	6.9 KiB	−17 B (−0.2%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.2 KiB	−303 B (−4.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	54.7 KiB	−2.3 KiB (−4.1%)	66.4 KiB	✅
Codex	Live turn messages	10	8	−2 (−20.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−18 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.9 KiB	6.9 KiB	+1 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.6 KiB	6.6 KiB	−19 B (−0.3%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: ea71a19 · PR result: d99162b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 109.4 KiB
• Claude decoded thread snapshot: 110.1 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR changes production desktop update state and IPC contracts while replacing the nightly tooltip with a focus-managed, hover-enabled release-notes popover and new external-link workflow. The interaction changes introduce meaningful focus and state-transition risk that should receive human review.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d99162b. Configure here.}

[t3dotgg]

Nightly update previews showed the oldest changes in each GitHub release. Contributor credits could also consume preview slots, so recent changes were often missing.

This change filters release boilerplate, keeps the newest eight changes, and reports the full change and release counts. The sidebar preview now links to each GitHub release and uses a focus-managed popover for pointer and keyboard access.

Before	After

Watch the pointer and keyboard interaction. The 2.5-second excerpt is slowed to 2.5x duration so the focus changes are easier to see.

Verification:

• 86 focused desktop and web tests passed.
• Contracts, desktop, and web typechecks passed.
• Changed-file lint, formatting, and git diff --check passed.
• The affected public nightly release produced the newest eight of ten changes and excluded contributor credits.
• Markdown and HTML section headings are excluded from the reported change count.
• The real app in Chromium passed pointer travel and scrolling, release-link activation, keyboard entry and exit, Escape focus return, download, install cancellation and confirmation, and tooltip fallback checks. The test used a stubbed desktop bridge, so it did not download or install a real update.
• The same update button DOM node kept keyboard focus through idle, checking, nightly available, no-notes, and latest state changes. These flows passed with React Compiler enabled.

This continues #8588. Theo Browne authored its three original commits. This replacement preserves his authorship and adds the review fixes separately.

Made with GPT-5.6 Sol in the Codex harness.

---

Note

Medium Risk
IPC contract changes require totalItems and omittedReleaseCount on update state, which can break older clients or persisted snapshots. Popover focus and React Compiler opt-out add UI regression surface on the nightly update control.

Overview
Nightly desktop update previews now surface the newest changes per release instead of the oldest, and stop treating contributor credits and full-changelog boilerplate as real changes. Normalization returns up to eight items and six release groups, with totalItems per group and omittedReleaseCount when older releases are dropped; those fields flow through IPC and the desktop update reducer.

The sidebar replaces the cramped nightly tooltip with a hover/focus popover (SidebarUpdateReleaseNotes) that links each preview to its GitHub release tag, labels hidden changes within a release, and links omitted releases to the releases history page. Popover behavior keeps the update button as the action target (no double-toggle on click), supports forward Tab into the panel, and restores focus sensibly on Escape.

User docs add a short section describing the nightly preview limits and links.

^{Reviewed by Cursor Bugbot for commit d90cf5f70eb63caae8fc2e2940d1a91f3e8f8d98. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Show newest nightly release changes in sidebar popover with per-release links

• Overhauls release-note extraction to keep the newest eight changes per release in newest-first order, excluding contributor and changelog sections, and returns up to six non-empty release groups plus an omittedReleaseCount
• Replaces the sidebar tooltip with a hover- and focus-aware Popover that renders nightly notes via the new SidebarUpdateReleaseNotes component, while keeping the update button as the install action
• SidebarUpdateReleaseNotes links each displayed release to its exact GitHub tag, reports omitted changes per release via totalItems, and links omitted release groups to the GitHub release history page
• Adds required omittedReleaseCount to DesktopUpdateState and totalItems to DesktopUpdateReleaseNote in the IPC contracts and runtime schema
• Risk: omittedReleaseCount and totalItems are now required fields in ipc.ts; any out-of-tree desktop clients reading the update-state payload must supply or tolerate these fields

^{Macroscope summarized d99162b.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.4 KiB	13.1 KiB	−320 B (−2.3%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.9 KiB	6.9 KiB	−17 B (−0.2%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.2 KiB	−303 B (−4.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	54.7 KiB	−2.3 KiB (−4.1%)	66.4 KiB	✅
Codex	Live turn messages	10	8	−2 (−20.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−18 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.9 KiB	6.9 KiB	+1 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.6 KiB	6.6 KiB	−19 B (−0.3%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: ea71a19 · PR result: d99162b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 109.4 KiB
• Claude decoded thread snapshot: 110.1 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR changes production desktop update state and IPC contracts while replacing the nightly tooltip with a focus-managed, hover-enabled release-notes popover and new external-link workflow. The interaction changes introduce meaningful focus and state-transition risk that should receive human review.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d99162b. Configure here.}

[t3dotgg]

Nightly update previews showed the oldest changes in each GitHub release. Contributor credits could also consume preview slots, so recent changes were often missing.

This change filters release boilerplate, keeps the newest eight changes, and reports the full change and release counts. The sidebar preview now links to each GitHub release and uses a focus-managed popover for pointer and keyboard access.

Before	After

Watch the pointer and keyboard interaction. The 2.5-second excerpt is slowed to 2.5x duration so the focus changes are easier to see.

Verification:

• 86 focused desktop and web tests passed.
• Contracts, desktop, and web typechecks passed.
• Changed-file lint, formatting, and git diff --check passed.
• The affected public nightly release produced the newest eight of ten changes and excluded contributor credits.
• Markdown and HTML section headings are excluded from the reported change count.
• The real app in Chromium passed pointer travel and scrolling, release-link activation, keyboard entry and exit, Escape focus return, download, install cancellation and confirmation, and tooltip fallback checks. The test used a stubbed desktop bridge, so it did not download or install a real update.
• The same update button DOM node kept keyboard focus through idle, checking, nightly available, no-notes, and latest state changes. These flows passed with React Compiler enabled.

This continues #8588. Theo Browne authored its three original commits. This replacement preserves his authorship and adds the review fixes separately.

Made with GPT-5.6 Sol in the Codex harness.

---

Note

Medium Risk
IPC contract changes require totalItems and omittedReleaseCount on update state, which can break older clients or persisted snapshots. Popover focus and React Compiler opt-out add UI regression surface on the nightly update control.

Overview
Nightly desktop update previews now surface the newest changes per release instead of the oldest, and stop treating contributor credits and full-changelog boilerplate as real changes. Normalization returns up to eight items and six release groups, with totalItems per group and omittedReleaseCount when older releases are dropped; those fields flow through IPC and the desktop update reducer.

The sidebar replaces the cramped nightly tooltip with a hover/focus popover (SidebarUpdateReleaseNotes) that links each preview to its GitHub release tag, labels hidden changes within a release, and links omitted releases to the releases history page. Popover behavior keeps the update button as the action target (no double-toggle on click), supports forward Tab into the panel, and restores focus sensibly on Escape.

User docs add a short section describing the nightly preview limits and links.

^{Reviewed by Cursor Bugbot for commit d90cf5f70eb63caae8fc2e2940d1a91f3e8f8d98. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Show newest nightly release changes in sidebar popover with per-release links

• Overhauls release-note extraction to keep the newest eight changes per release in newest-first order, excluding contributor and changelog sections, and returns up to six non-empty release groups plus an omittedReleaseCount
• Replaces the sidebar tooltip with a hover- and focus-aware Popover that renders nightly notes via the new SidebarUpdateReleaseNotes component, while keeping the update button as the install action
• SidebarUpdateReleaseNotes links each displayed release to its exact GitHub tag, reports omitted changes per release via totalItems, and links omitted release groups to the GitHub release history page
• Adds required omittedReleaseCount to DesktopUpdateState and totalItems to DesktopUpdateReleaseNote in the IPC contracts and runtime schema
• Risk: omittedReleaseCount and totalItems are now required fields in ipc.ts; any out-of-tree desktop clients reading the update-state payload must supply or tolerate these fields

^{Macroscope summarized d99162b.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.4 KiB	13.1 KiB	−320 B (−2.3%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.9 KiB	6.9 KiB	−17 B (−0.2%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.2 KiB	−303 B (−4.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	57.0 KiB	54.7 KiB	−2.3 KiB (−4.1%)	66.4 KiB	✅
Codex	Live turn messages	10	8	−2 (−20.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−18 B (−0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.9 KiB	6.9 KiB	+1 B (+0.0%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.6 KiB	6.6 KiB	−19 B (−0.3%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.8 KiB	57.8 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	10	10	0 (0.0%)	21	✅

Baseline: ea71a19 · PR result: d99162b · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 109.4 KiB
• Claude decoded thread snapshot: 110.1 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[macroscopeapp]

Approvability

Verdict: Not approved

Macroscope's review found this PR not approvable — The PR changes production desktop update state and IPC contracts while replacing the nightly tooltip with a focus-managed, hover-enabled release-notes popover and new external-link workflow. The interaction changes introduce meaningful focus and state-transition risk that should receive human review.

^{You can add or adjust custom eligibility rules. Learn more.}

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit d99162b. Configure here.}