fix(cli): flip the expansion default and explain when Ctrl+O/Ctrl+T finds no block in view by Astro-Han · Pull Request #1140 · apache/maka

Astro-Han · 2026-07-16T16:21:50Z

Summary

Closes#1134. Follow-up to #1097 / #1130.

Expanding a block taller than the terminal pushes the block's own head above the live viewport into terminal scrollback. Scrollback is append-only, so collapsing that block in place is physically impossible without the scrollback-clearing full redraw the #1097 contract forbids — the second Ctrl+O/Ctrl+T found no in-viewport target, returned false, and silently left both the block and the expansion default stuck.

Per the decision on #1134, this accepts the physical limitation instead of "fixing" it (option 3), and removes the two real defects around it:

toggleAllToolExpansion / toggleAllThinkingExpansion now always flip the expansion default when any candidate entry exists and entry positions are known, so a stuck default no longer leaks into future entries.
When every candidate sits above the viewport, the toggle appends an info notice — the keypress is no longer silent — saying the content above stays readable in scrollback and which state new output starts in.
Toggles stay fully inert while entry positions are unknown after a wholesale replacement (unchanged fix(cli): anchor Ctrl+O/Ctrl+T expansion to the live viewport so toggles never clear scrollback #1130 contract), and entries above the viewport stay byte-identical across a toggle (unchanged fix(cli): expanding an off-screen block triggers a full redraw that clears pre-Maka scrollback #1097 contract).

Not done, deliberately: clamping expanded rendering (regresses full-content-into-scrollback), letting toggles touch head-scrolled blocks (reintroduces the #1097 wipe), or an explicit scrollback-clearing redraw command (possible orthogonal follow-up).

Verification

packages/cli: npm test 465/465 (3 new/updated unit tests, 1 new integration test), root npm run lint clean.
New unit tests lock: head-scrolled second toggle flips the default back, freezes the entry, keeps scrollback lines byte-identical, and appends the notice; a session with no candidate still reports nothing to toggle; the post-replacement inert window is preserved.
New integration test drives the real pi-tui renderer through runMakaPiTui with a 24-row terminal: an 80-line thinking block is expanded past the viewport, the second Ctrl+T renders the notice, and the whole output stream contains no ESC[3J.
Manually drove the built TUI in a real tmux TTY (100×24): expanded an 80-line thinking block (head scrolled into tmux scrollback), pressed Ctrl+T again → notice rendered, expansion frozen, all 80 rows still present in tmux scrollback with no screen clear; Ctrl+O on the in-viewport tool card still round-trips expand/collapse.

 reason-row-78
reason-row-79
● Bash $ npm test (41 lines)
done answering
Note: No thinking in view to toggle — thinking above stays as rendered in scrollback. New thinking
starts collapsed.

…inds no block in view A block expanded past the terminal pushes its own head into scrollback, which the #1097 contract forbids rewriting, so the second Ctrl+O/Ctrl+T found no target, returned false, and silently left both the block and the expansion default stuck. The toggles now always flip the default when any candidate entry exists and positions are known; when every candidate sits above the viewport they append an info notice saying the content stays readable in scrollback and which state new output starts in. Toggles stay fully inert while entry positions are unknown after a wholesale replacement, and entries above the viewport stay byte-identical as before. Closes#1134

…primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode

…bove the viewport (#4011) (#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the #1097 contract forbids rewriting, so they stay expanded with no way back (#1134). #1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (#1134 option 2, deferred by #1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

…bove the viewport (apache#4011) (apache#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (apache#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

Astro-Han · 2026-07-16T16:21:50Z

Summary

Closes#1134. Follow-up to #1097 / #1130.

Expanding a block taller than the terminal pushes the block's own head above the live viewport into terminal scrollback. Scrollback is append-only, so collapsing that block in place is physically impossible without the scrollback-clearing full redraw the #1097 contract forbids — the second Ctrl+O/Ctrl+T found no in-viewport target, returned false, and silently left both the block and the expansion default stuck.

Per the decision on #1134, this accepts the physical limitation instead of "fixing" it (option 3), and removes the two real defects around it:

toggleAllToolExpansion / toggleAllThinkingExpansion now always flip the expansion default when any candidate entry exists and entry positions are known, so a stuck default no longer leaks into future entries.
When every candidate sits above the viewport, the toggle appends an info notice — the keypress is no longer silent — saying the content above stays readable in scrollback and which state new output starts in.
Toggles stay fully inert while entry positions are unknown after a wholesale replacement (unchanged fix(cli): anchor Ctrl+O/Ctrl+T expansion to the live viewport so toggles never clear scrollback #1130 contract), and entries above the viewport stay byte-identical across a toggle (unchanged fix(cli): expanding an off-screen block triggers a full redraw that clears pre-Maka scrollback #1097 contract).

Not done, deliberately: clamping expanded rendering (regresses full-content-into-scrollback), letting toggles touch head-scrolled blocks (reintroduces the #1097 wipe), or an explicit scrollback-clearing redraw command (possible orthogonal follow-up).

Verification

packages/cli: npm test 465/465 (3 new/updated unit tests, 1 new integration test), root npm run lint clean.
New unit tests lock: head-scrolled second toggle flips the default back, freezes the entry, keeps scrollback lines byte-identical, and appends the notice; a session with no candidate still reports nothing to toggle; the post-replacement inert window is preserved.
New integration test drives the real pi-tui renderer through runMakaPiTui with a 24-row terminal: an 80-line thinking block is expanded past the viewport, the second Ctrl+T renders the notice, and the whole output stream contains no ESC[3J.
Manually drove the built TUI in a real tmux TTY (100×24): expanded an 80-line thinking block (head scrolled into tmux scrollback), pressed Ctrl+T again → notice rendered, expansion frozen, all 80 rows still present in tmux scrollback with no screen clear; Ctrl+O on the in-viewport tool card still round-trips expand/collapse.

 reason-row-78
reason-row-79
● Bash $ npm test (41 lines)
done answering
Note: No thinking in view to toggle — thinking above stays as rendered in scrollback. New thinking
starts collapsed.

…inds no block in view A block expanded past the terminal pushes its own head into scrollback, which the #1097 contract forbids rewriting, so the second Ctrl+O/Ctrl+T found no target, returned false, and silently left both the block and the expansion default stuck. The toggles now always flip the default when any candidate entry exists and positions are known; when every candidate sits above the viewport they append an info notice saying the content stays readable in scrollback and which state new output starts in. Toggles stay fully inert while entry positions are unknown after a wholesale replacement, and entries above the viewport stay byte-identical as before. Closes#1134

…primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode

…bove the viewport (#4011) (#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the #1097 contract forbids rewriting, so they stay expanded with no way back (#1134). #1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (#1134 option 2, deferred by #1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

…bove the viewport (apache#4011) (apache#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (apache#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

Astro-Han · 2026-07-16T16:21:50Z

Summary

Closes#1134. Follow-up to #1097 / #1130.

Expanding a block taller than the terminal pushes the block's own head above the live viewport into terminal scrollback. Scrollback is append-only, so collapsing that block in place is physically impossible without the scrollback-clearing full redraw the #1097 contract forbids — the second Ctrl+O/Ctrl+T found no in-viewport target, returned false, and silently left both the block and the expansion default stuck.

Per the decision on #1134, this accepts the physical limitation instead of "fixing" it (option 3), and removes the two real defects around it:

toggleAllToolExpansion / toggleAllThinkingExpansion now always flip the expansion default when any candidate entry exists and entry positions are known, so a stuck default no longer leaks into future entries.
When every candidate sits above the viewport, the toggle appends an info notice — the keypress is no longer silent — saying the content above stays readable in scrollback and which state new output starts in.
Toggles stay fully inert while entry positions are unknown after a wholesale replacement (unchanged fix(cli): anchor Ctrl+O/Ctrl+T expansion to the live viewport so toggles never clear scrollback #1130 contract), and entries above the viewport stay byte-identical across a toggle (unchanged fix(cli): expanding an off-screen block triggers a full redraw that clears pre-Maka scrollback #1097 contract).

Not done, deliberately: clamping expanded rendering (regresses full-content-into-scrollback), letting toggles touch head-scrolled blocks (reintroduces the #1097 wipe), or an explicit scrollback-clearing redraw command (possible orthogonal follow-up).

Verification

packages/cli: npm test 465/465 (3 new/updated unit tests, 1 new integration test), root npm run lint clean.
New unit tests lock: head-scrolled second toggle flips the default back, freezes the entry, keeps scrollback lines byte-identical, and appends the notice; a session with no candidate still reports nothing to toggle; the post-replacement inert window is preserved.
New integration test drives the real pi-tui renderer through runMakaPiTui with a 24-row terminal: an 80-line thinking block is expanded past the viewport, the second Ctrl+T renders the notice, and the whole output stream contains no ESC[3J.
Manually drove the built TUI in a real tmux TTY (100×24): expanded an 80-line thinking block (head scrolled into tmux scrollback), pressed Ctrl+T again → notice rendered, expansion frozen, all 80 rows still present in tmux scrollback with no screen clear; Ctrl+O on the in-viewport tool card still round-trips expand/collapse.

 reason-row-78
reason-row-79
● Bash $ npm test (41 lines)
done answering
Note: No thinking in view to toggle — thinking above stays as rendered in scrollback. New thinking
starts collapsed.

…inds no block in view A block expanded past the terminal pushes its own head into scrollback, which the #1097 contract forbids rewriting, so the second Ctrl+O/Ctrl+T found no target, returned false, and silently left both the block and the expansion default stuck. The toggles now always flip the default when any candidate entry exists and positions are known; when every candidate sits above the viewport they append an info notice saying the content stays readable in scrollback and which state new output starts in. Toggles stay fully inert while entry positions are unknown after a wholesale replacement, and entries above the viewport stay byte-identical as before. Closes#1134

…primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode

…bove the viewport (#4011) (#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the #1097 contract forbids rewriting, so they stay expanded with no way back (#1134). #1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (#1134 option 2, deferred by #1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

…bove the viewport (apache#4011) (apache#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (apache#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

Astro-Han · 2026-07-16T16:21:50Z

Summary

Closes#1134. Follow-up to #1097 / #1130.

Expanding a block taller than the terminal pushes the block's own head above the live viewport into terminal scrollback. Scrollback is append-only, so collapsing that block in place is physically impossible without the scrollback-clearing full redraw the #1097 contract forbids — the second Ctrl+O/Ctrl+T found no in-viewport target, returned false, and silently left both the block and the expansion default stuck.

Per the decision on #1134, this accepts the physical limitation instead of "fixing" it (option 3), and removes the two real defects around it:

toggleAllToolExpansion / toggleAllThinkingExpansion now always flip the expansion default when any candidate entry exists and entry positions are known, so a stuck default no longer leaks into future entries.
When every candidate sits above the viewport, the toggle appends an info notice — the keypress is no longer silent — saying the content above stays readable in scrollback and which state new output starts in.
Toggles stay fully inert while entry positions are unknown after a wholesale replacement (unchanged fix(cli): anchor Ctrl+O/Ctrl+T expansion to the live viewport so toggles never clear scrollback #1130 contract), and entries above the viewport stay byte-identical across a toggle (unchanged fix(cli): expanding an off-screen block triggers a full redraw that clears pre-Maka scrollback #1097 contract).

Not done, deliberately: clamping expanded rendering (regresses full-content-into-scrollback), letting toggles touch head-scrolled blocks (reintroduces the #1097 wipe), or an explicit scrollback-clearing redraw command (possible orthogonal follow-up).

Verification

packages/cli: npm test 465/465 (3 new/updated unit tests, 1 new integration test), root npm run lint clean.
New unit tests lock: head-scrolled second toggle flips the default back, freezes the entry, keeps scrollback lines byte-identical, and appends the notice; a session with no candidate still reports nothing to toggle; the post-replacement inert window is preserved.
New integration test drives the real pi-tui renderer through runMakaPiTui with a 24-row terminal: an 80-line thinking block is expanded past the viewport, the second Ctrl+T renders the notice, and the whole output stream contains no ESC[3J.
Manually drove the built TUI in a real tmux TTY (100×24): expanded an 80-line thinking block (head scrolled into tmux scrollback), pressed Ctrl+T again → notice rendered, expansion frozen, all 80 rows still present in tmux scrollback with no screen clear; Ctrl+O on the in-viewport tool card still round-trips expand/collapse.

 reason-row-78
reason-row-79
● Bash $ npm test (41 lines)
done answering
Note: No thinking in view to toggle — thinking above stays as rendered in scrollback. New thinking
starts collapsed.

…inds no block in view A block expanded past the terminal pushes its own head into scrollback, which the #1097 contract forbids rewriting, so the second Ctrl+O/Ctrl+T found no target, returned false, and silently left both the block and the expansion default stuck. The toggles now always flip the default when any candidate entry exists and positions are known; when every candidate sits above the viewport they append an info notice saying the content stays readable in scrollback and which state new output starts in. Toggles stay fully inert while entry positions are unknown after a wholesale replacement, and entries above the viewport stay byte-identical as before. Closes#1134

…primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode

…bove the viewport (#4011) (#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the #1097 contract forbids rewriting, so they stay expanded with no way back (#1134). #1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (#1134 option 2, deferred by #1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

…bove the viewport (apache#4011) (apache#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (apache#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

Astro-Han · 2026-07-16T16:21:50Z

Summary

Closes#1134. Follow-up to #1097 / #1130.

Expanding a block taller than the terminal pushes the block's own head above the live viewport into terminal scrollback. Scrollback is append-only, so collapsing that block in place is physically impossible without the scrollback-clearing full redraw the #1097 contract forbids — the second Ctrl+O/Ctrl+T found no in-viewport target, returned false, and silently left both the block and the expansion default stuck.

Per the decision on #1134, this accepts the physical limitation instead of "fixing" it (option 3), and removes the two real defects around it:

toggleAllToolExpansion / toggleAllThinkingExpansion now always flip the expansion default when any candidate entry exists and entry positions are known, so a stuck default no longer leaks into future entries.
When every candidate sits above the viewport, the toggle appends an info notice — the keypress is no longer silent — saying the content above stays readable in scrollback and which state new output starts in.
Toggles stay fully inert while entry positions are unknown after a wholesale replacement (unchanged fix(cli): anchor Ctrl+O/Ctrl+T expansion to the live viewport so toggles never clear scrollback #1130 contract), and entries above the viewport stay byte-identical across a toggle (unchanged fix(cli): expanding an off-screen block triggers a full redraw that clears pre-Maka scrollback #1097 contract).

Not done, deliberately: clamping expanded rendering (regresses full-content-into-scrollback), letting toggles touch head-scrolled blocks (reintroduces the #1097 wipe), or an explicit scrollback-clearing redraw command (possible orthogonal follow-up).

Verification

packages/cli: npm test 465/465 (3 new/updated unit tests, 1 new integration test), root npm run lint clean.
New unit tests lock: head-scrolled second toggle flips the default back, freezes the entry, keeps scrollback lines byte-identical, and appends the notice; a session with no candidate still reports nothing to toggle; the post-replacement inert window is preserved.
New integration test drives the real pi-tui renderer through runMakaPiTui with a 24-row terminal: an 80-line thinking block is expanded past the viewport, the second Ctrl+T renders the notice, and the whole output stream contains no ESC[3J.
Manually drove the built TUI in a real tmux TTY (100×24): expanded an 80-line thinking block (head scrolled into tmux scrollback), pressed Ctrl+T again → notice rendered, expansion frozen, all 80 rows still present in tmux scrollback with no screen clear; Ctrl+O on the in-viewport tool card still round-trips expand/collapse.

 reason-row-78
reason-row-79
● Bash $ npm test (41 lines)
done answering
Note: No thinking in view to toggle — thinking above stays as rendered in scrollback. New thinking
starts collapsed.

…inds no block in view A block expanded past the terminal pushes its own head into scrollback, which the #1097 contract forbids rewriting, so the second Ctrl+O/Ctrl+T found no target, returned false, and silently left both the block and the expansion default stuck. The toggles now always flip the default when any candidate entry exists and positions are known; when every candidate sits above the viewport they append an info notice saying the content stays readable in scrollback and which state new output starts in. Toggles stay fully inert while entry positions are unknown after a wholesale replacement, and entries above the viewport stay byte-identical as before. Closes#1134

…primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode

…bove the viewport (#4011) (#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the #1097 contract forbids rewriting, so they stay expanded with no way back (#1134). #1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (#1134 option 2, deferred by #1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

…bove the viewport (apache#4011) (apache#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (apache#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

Astro-Han · 2026-07-16T16:21:50Z

Summary

Closes#1134. Follow-up to #1097 / #1130.

Expanding a block taller than the terminal pushes the block's own head above the live viewport into terminal scrollback. Scrollback is append-only, so collapsing that block in place is physically impossible without the scrollback-clearing full redraw the #1097 contract forbids — the second Ctrl+O/Ctrl+T found no in-viewport target, returned false, and silently left both the block and the expansion default stuck.

Per the decision on #1134, this accepts the physical limitation instead of "fixing" it (option 3), and removes the two real defects around it:

toggleAllToolExpansion / toggleAllThinkingExpansion now always flip the expansion default when any candidate entry exists and entry positions are known, so a stuck default no longer leaks into future entries.
When every candidate sits above the viewport, the toggle appends an info notice — the keypress is no longer silent — saying the content above stays readable in scrollback and which state new output starts in.
Toggles stay fully inert while entry positions are unknown after a wholesale replacement (unchanged fix(cli): anchor Ctrl+O/Ctrl+T expansion to the live viewport so toggles never clear scrollback #1130 contract), and entries above the viewport stay byte-identical across a toggle (unchanged fix(cli): expanding an off-screen block triggers a full redraw that clears pre-Maka scrollback #1097 contract).

Not done, deliberately: clamping expanded rendering (regresses full-content-into-scrollback), letting toggles touch head-scrolled blocks (reintroduces the #1097 wipe), or an explicit scrollback-clearing redraw command (possible orthogonal follow-up).

Verification

packages/cli: npm test 465/465 (3 new/updated unit tests, 1 new integration test), root npm run lint clean.
New unit tests lock: head-scrolled second toggle flips the default back, freezes the entry, keeps scrollback lines byte-identical, and appends the notice; a session with no candidate still reports nothing to toggle; the post-replacement inert window is preserved.
New integration test drives the real pi-tui renderer through runMakaPiTui with a 24-row terminal: an 80-line thinking block is expanded past the viewport, the second Ctrl+T renders the notice, and the whole output stream contains no ESC[3J.
Manually drove the built TUI in a real tmux TTY (100×24): expanded an 80-line thinking block (head scrolled into tmux scrollback), pressed Ctrl+T again → notice rendered, expansion frozen, all 80 rows still present in tmux scrollback with no screen clear; Ctrl+O on the in-viewport tool card still round-trips expand/collapse.

 reason-row-78
reason-row-79
● Bash $ npm test (41 lines)
done answering
Note: No thinking in view to toggle — thinking above stays as rendered in scrollback. New thinking
starts collapsed.

…inds no block in view A block expanded past the terminal pushes its own head into scrollback, which the #1097 contract forbids rewriting, so the second Ctrl+O/Ctrl+T found no target, returned false, and silently left both the block and the expansion default stuck. The toggles now always flip the default when any candidate entry exists and positions are known; when every candidate sits above the viewport they append an info notice saying the content stays readable in scrollback and which state new output starts in. Toggles stay fully inert while entry positions are unknown after a wholesale replacement, and entries above the viewport stay byte-identical as before. Closes#1134

…primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode

…bove the viewport (#4011) (#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the #1097 contract forbids rewriting, so they stay expanded with no way back (#1134). #1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (#1134 option 2, deferred by #1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

…bove the viewport (apache#4011) (apache#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (apache#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

Astro-Han · 2026-07-16T16:21:50Z

Summary

Closes#1134. Follow-up to #1097 / #1130.

Expanding a block taller than the terminal pushes the block's own head above the live viewport into terminal scrollback. Scrollback is append-only, so collapsing that block in place is physically impossible without the scrollback-clearing full redraw the #1097 contract forbids — the second Ctrl+O/Ctrl+T found no in-viewport target, returned false, and silently left both the block and the expansion default stuck.

Per the decision on #1134, this accepts the physical limitation instead of "fixing" it (option 3), and removes the two real defects around it:

toggleAllToolExpansion / toggleAllThinkingExpansion now always flip the expansion default when any candidate entry exists and entry positions are known, so a stuck default no longer leaks into future entries.
When every candidate sits above the viewport, the toggle appends an info notice — the keypress is no longer silent — saying the content above stays readable in scrollback and which state new output starts in.
Toggles stay fully inert while entry positions are unknown after a wholesale replacement (unchanged fix(cli): anchor Ctrl+O/Ctrl+T expansion to the live viewport so toggles never clear scrollback #1130 contract), and entries above the viewport stay byte-identical across a toggle (unchanged fix(cli): expanding an off-screen block triggers a full redraw that clears pre-Maka scrollback #1097 contract).

Not done, deliberately: clamping expanded rendering (regresses full-content-into-scrollback), letting toggles touch head-scrolled blocks (reintroduces the #1097 wipe), or an explicit scrollback-clearing redraw command (possible orthogonal follow-up).

Verification

packages/cli: npm test 465/465 (3 new/updated unit tests, 1 new integration test), root npm run lint clean.
New unit tests lock: head-scrolled second toggle flips the default back, freezes the entry, keeps scrollback lines byte-identical, and appends the notice; a session with no candidate still reports nothing to toggle; the post-replacement inert window is preserved.
New integration test drives the real pi-tui renderer through runMakaPiTui with a 24-row terminal: an 80-line thinking block is expanded past the viewport, the second Ctrl+T renders the notice, and the whole output stream contains no ESC[3J.
Manually drove the built TUI in a real tmux TTY (100×24): expanded an 80-line thinking block (head scrolled into tmux scrollback), pressed Ctrl+T again → notice rendered, expansion frozen, all 80 rows still present in tmux scrollback with no screen clear; Ctrl+O on the in-viewport tool card still round-trips expand/collapse.

 reason-row-78
reason-row-79
● Bash $ npm test (41 lines)
done answering
Note: No thinking in view to toggle — thinking above stays as rendered in scrollback. New thinking
starts collapsed.

…inds no block in view A block expanded past the terminal pushes its own head into scrollback, which the #1097 contract forbids rewriting, so the second Ctrl+O/Ctrl+T found no target, returned false, and silently left both the block and the expansion default stuck. The toggles now always flip the default when any candidate entry exists and positions are known; when every candidate sits above the viewport they append an info notice saying the content stays readable in scrollback and which state new output starts in. Toggles stay fully inert while entry positions are unknown after a wholesale replacement, and entries above the viewport stay byte-identical as before. Closes#1134

…primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode

…bove the viewport (#4011) (#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the #1097 contract forbids rewriting, so they stay expanded with no way back (#1134). #1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (#1134 option 2, deferred by #1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

…bove the viewport (apache#4011) (apache#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (apache#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

Astro-Han · 2026-07-16T16:21:50Z

Summary

Closes#1134. Follow-up to #1097 / #1130.

Expanding a block taller than the terminal pushes the block's own head above the live viewport into terminal scrollback. Scrollback is append-only, so collapsing that block in place is physically impossible without the scrollback-clearing full redraw the #1097 contract forbids — the second Ctrl+O/Ctrl+T found no in-viewport target, returned false, and silently left both the block and the expansion default stuck.

Per the decision on #1134, this accepts the physical limitation instead of "fixing" it (option 3), and removes the two real defects around it:

toggleAllToolExpansion / toggleAllThinkingExpansion now always flip the expansion default when any candidate entry exists and entry positions are known, so a stuck default no longer leaks into future entries.
When every candidate sits above the viewport, the toggle appends an info notice — the keypress is no longer silent — saying the content above stays readable in scrollback and which state new output starts in.
Toggles stay fully inert while entry positions are unknown after a wholesale replacement (unchanged fix(cli): anchor Ctrl+O/Ctrl+T expansion to the live viewport so toggles never clear scrollback #1130 contract), and entries above the viewport stay byte-identical across a toggle (unchanged fix(cli): expanding an off-screen block triggers a full redraw that clears pre-Maka scrollback #1097 contract).

Not done, deliberately: clamping expanded rendering (regresses full-content-into-scrollback), letting toggles touch head-scrolled blocks (reintroduces the #1097 wipe), or an explicit scrollback-clearing redraw command (possible orthogonal follow-up).

Verification

packages/cli: npm test 465/465 (3 new/updated unit tests, 1 new integration test), root npm run lint clean.
New unit tests lock: head-scrolled second toggle flips the default back, freezes the entry, keeps scrollback lines byte-identical, and appends the notice; a session with no candidate still reports nothing to toggle; the post-replacement inert window is preserved.
New integration test drives the real pi-tui renderer through runMakaPiTui with a 24-row terminal: an 80-line thinking block is expanded past the viewport, the second Ctrl+T renders the notice, and the whole output stream contains no ESC[3J.
Manually drove the built TUI in a real tmux TTY (100×24): expanded an 80-line thinking block (head scrolled into tmux scrollback), pressed Ctrl+T again → notice rendered, expansion frozen, all 80 rows still present in tmux scrollback with no screen clear; Ctrl+O on the in-viewport tool card still round-trips expand/collapse.

 reason-row-78
reason-row-79
● Bash $ npm test (41 lines)
done answering
Note: No thinking in view to toggle — thinking above stays as rendered in scrollback. New thinking
starts collapsed.

…inds no block in view A block expanded past the terminal pushes its own head into scrollback, which the #1097 contract forbids rewriting, so the second Ctrl+O/Ctrl+T found no target, returned false, and silently left both the block and the expansion default stuck. The toggles now always flip the default when any candidate entry exists and positions are known; when every candidate sits above the viewport they append an info notice saying the content stays readable in scrollback and which state new output starts in. Toggles stay fully inert while entry positions are unknown after a wholesale replacement, and entries above the viewport stay byte-identical as before. Closes#1134

…primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode

…bove the viewport (#4011) (#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the #1097 contract forbids rewriting, so they stay expanded with no way back (#1134). #1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (#1134 option 2, deferred by #1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

…bove the viewport (apache#4011) (apache#4025) * feat(cli): name stranded expansion blocks and add confirmed-collapse primitives (apache#4011) A Ctrl+O/Ctrl+T collapse can strand expanded blocks above the live viewport: their heads sit in terminal scrollback, which the apache#1097 contract forbids rewriting, so they stay expanded with no way back (apache#1134). apache#1140 made the all-stranded case explained but left the far more common partial case silent — some cards collapse, some stay stuck, and the keypress reads as broken. The toggle now appends a notice whenever a collapse strands blocks, naming the count and offering the escape hatch: a second press within EXPANSION_COLLAPSE_CONFIRM_WINDOW_MS collapses them too via one knowingly-accepted, scrollback-clearing full redraw (apache#1134 option 2, deferred by apache#1140 as an orthogonal follow-up). The confirm offer exists for collapses only — collapsed blocks above the viewport are compact and harmless, and arming on expand would make a quick expand-then- collapse pair read the second press as "expand everything". New exports for the runner wiring: hasExpandedEntriesAboveViewport (arming predicate) and applyExpansionDefaultToAll (applies the current default to every entry including above-viewport ones, without flipping the default the way a plain toggle would). The window constant lives beside the notice copy so the offer text and the runner's confirm window share one authority. Generated-by: OpenCode * feat(cli): second Ctrl+O/Ctrl+T press pays one full redraw to collapse stranded blocks (apache#4011) Wire the confirm gesture into the key handler: a collapse toggle that leaves expanded blocks above the viewport arms a 2s window; pressing the same key again inside it applies the collapsed default to every entry and forces one scrollback-clearing full redraw through pi-tui's existing public requestRender(true), which re-anchors the viewport at the tail. Transcript content is fully re-rendered into fresh scrollback, so nothing from the session is lost; only pre-session shell scrollback is cleared, and only after the explicit second press the notice announced. The integration test drives the real renderer on a 24-row terminal: an 80-line thinking block is expanded past the viewport, the first collapse press renders the offer with no ESC[3J in the stream, and the confirmed second press emits exactly the deliberate clear while collapsing the block back to its compact row. Generated-by: OpenCode * fix(cli): re-offer expired expansion confirmation

Astro-Han added 2 commits July 17, 2026 00:21

docs(cli): state why unscrolled unknown geometry keeps the toggles live
77e2f55

Astro-Han merged commit e96f8b7 into mainJul 16, 2026
3 checks passed

Astro-Han deleted the fix/cli-headscrolled-toggle-feedback branch July 24, 2026 09:05

This was referenced Aug 27, 2026
proposal(cli): confirmed full redraw so Ctrl+O/Ctrl+T can collapse blocks stuck above the viewport #4011
Closed
feat(cli): confirmed second Ctrl+O/Ctrl+T collapses blocks stranded above the viewport (#4011) #4025
Merged

Astro-Han added 2 commits July 17, 2026 00:21

docs(cli): state why unscrolled unknown geometry keeps the toggles live
77e2f55

Astro-Han merged commit e96f8b7 into mainJul 16, 2026
3 checks passed

Astro-Han deleted the fix/cli-headscrolled-toggle-feedback branch July 24, 2026 09:05

This was referenced Aug 27, 2026
proposal(cli): confirmed full redraw so Ctrl+O/Ctrl+T can collapse blocks stuck above the viewport #4011
Closed
feat(cli): confirmed second Ctrl+O/Ctrl+T collapses blocks stranded above the viewport (#4011) #4025
Merged

Astro-Han added 2 commits July 17, 2026 00:21

docs(cli): state why unscrolled unknown geometry keeps the toggles live
77e2f55

Astro-Han merged commit e96f8b7 into mainJul 16, 2026
3 checks passed

Astro-Han deleted the fix/cli-headscrolled-toggle-feedback branch July 24, 2026 09:05

This was referenced Aug 27, 2026
proposal(cli): confirmed full redraw so Ctrl+O/Ctrl+T can collapse blocks stuck above the viewport #4011
Closed
feat(cli): confirmed second Ctrl+O/Ctrl+T collapses blocks stranded above the viewport (#4011) #4025
Merged

Astro-Han added 2 commits July 17, 2026 00:21

docs(cli): state why unscrolled unknown geometry keeps the toggles live
77e2f55

Astro-Han merged commit e96f8b7 into mainJul 16, 2026
3 checks passed

Astro-Han deleted the fix/cli-headscrolled-toggle-feedback branch July 24, 2026 09:05

This was referenced Aug 27, 2026
proposal(cli): confirmed full redraw so Ctrl+O/Ctrl+T can collapse blocks stuck above the viewport #4011
Closed
feat(cli): confirmed second Ctrl+O/Ctrl+T collapses blocks stranded above the viewport (#4011) #4025
Merged

Astro-Han added 2 commits July 17, 2026 00:21

docs(cli): state why unscrolled unknown geometry keeps the toggles live
77e2f55

Astro-Han merged commit e96f8b7 into mainJul 16, 2026
3 checks passed

Astro-Han deleted the fix/cli-headscrolled-toggle-feedback branch July 24, 2026 09:05

This was referenced Aug 27, 2026
proposal(cli): confirmed full redraw so Ctrl+O/Ctrl+T can collapse blocks stuck above the viewport #4011
Closed
feat(cli): confirmed second Ctrl+O/Ctrl+T collapses blocks stranded above the viewport (#4011) #4025
Merged

Astro-Han added 2 commits July 17, 2026 00:21

docs(cli): state why unscrolled unknown geometry keeps the toggles live
77e2f55

Astro-Han merged commit e96f8b7 into mainJul 16, 2026
3 checks passed

Astro-Han deleted the fix/cli-headscrolled-toggle-feedback branch July 24, 2026 09:05

This was referenced Aug 27, 2026
proposal(cli): confirmed full redraw so Ctrl+O/Ctrl+T can collapse blocks stuck above the viewport #4011
Closed
feat(cli): confirmed second Ctrl+O/Ctrl+T collapses blocks stranded above the viewport (#4011) #4025
Merged

Astro-Han added 2 commits July 17, 2026 00:21

docs(cli): state why unscrolled unknown geometry keeps the toggles live
77e2f55

Astro-Han merged commit e96f8b7 into mainJul 16, 2026
3 checks passed

Astro-Han deleted the fix/cli-headscrolled-toggle-feedback branch July 24, 2026 09:05

This was referenced Aug 27, 2026
proposal(cli): confirmed full redraw so Ctrl+O/Ctrl+T can collapse blocks stuck above the viewport #4011
Closed
feat(cli): confirmed second Ctrl+O/Ctrl+T collapses blocks stranded above the viewport (#4011) #4025
Merged

Astro-Han added 2 commits July 17, 2026 00:21

docs(cli): state why unscrolled unknown geometry keeps the toggles live
77e2f55

Astro-Han merged commit e96f8b7 into mainJul 16, 2026
3 checks passed

Astro-Han deleted the fix/cli-headscrolled-toggle-feedback branch July 24, 2026 09:05

This was referenced Aug 27, 2026
proposal(cli): confirmed full redraw so Ctrl+O/Ctrl+T can collapse blocks stuck above the viewport #4011
Closed
feat(cli): confirmed second Ctrl+O/Ctrl+T collapses blocks stranded above the viewport (#4011) #4025
Merged

Uh oh!

Conversation

Astro-Han commented Jul 16, 2026

Summary

Verification

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

Astro-Han commented Jul 16, 2026

Summary

Verification

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

Astro-Han commented Jul 16, 2026

Summary

Verification

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

Astro-Han commented Jul 16, 2026

Summary

Verification

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

Astro-Han commented Jul 16, 2026

Summary

Verification

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

Astro-Han commented Jul 16, 2026

Summary

Verification