Files
zero/internal/tui/hover.go
Anandan d7ac85cbf9 Polish Zero's TUI and harden patch, model, and background workflows (#916)
* Checkpoint UI polish

* Fix structured apply patch handling

* Restore tool previews when resuming sessions

* Improve theme-aware code rendering

* Replace persistent sidebar with run details

* Remove inert transcript interactions

* Cover resumed edit previews

* Smooth active run status feedback

* Preserve semantic colors behind overlays

* Normalize ChatGPT model aliases

* Render staged image attachments in terminal

* Polish transcript and tool feedback

* Resolve rebased diff renderer conflict

* Address UI polish review findings

* Fix smoke test regressions

* Harden patch and transcript edge cases

* Bind structured patches to workspace root

* Use unique structured patch temp files

* Retry structured patch temp allocation

* Reuse rooted patch temp allocator

* Fix Windows CI process checks

* Fix structured patch safety boundaries

Preserve read-before-write credit only when the model already knew the full source, reject ambiguous hunk matches, and retain CRLF line endings. Add regression coverage for unsafe overwrites, positive tracking cases, duplicate matches, and Windows-style files.

* Resolve plan test seam after rebase

Drop the obsolete sticky-plan height helper while retaining the PR status test seams brought forward by the UI changes.

* Address final review correctness gaps

Bound narrow waiting labels, reject malformed plan records, preserve missing final newlines, reject out-of-order EOF hunks, clear stale tracking after failed rollback, and keep deferred tool loading aligned with model visibility.

* Harden patch transactions and review edges

Publish new patch targets without overwriting raced destinations, journal only completed mutations, and create replacement temps beside their targets. Reject pre-plan diagnostics, cover tool-search transcript suppression, and leave sufficient cleanup time in the Windows process fixture.

* Bind patch rollback to committed versions

Capture file identity and content before publication, quarantine rollback targets atomically, and refuse destructive undo when a path was replaced or modified in place. Add deterministic coverage for publication and rollback races.

* Align structured patch failure semantics

Use prevalidated sequential application with explicit partial-success errors instead of unsafe multi-file rollback. Retain atomic per-file publication, no-overwrite creates, unsupported-hardlink fallback, and conservative tracker invalidation.

* Preserve patch context bytes

Track source offsets for structured-patch context lines and reuse the exact matched source text during replacement. This keeps tolerant matching from changing indentation or mixed line endings outside the requested edit.
2026-08-19 22:18:47 +05:30

114 lines
4.5 KiB
Go

package tui
import tea "charm.land/bubbletea/v2"
// hoverKind discriminates which clickable surface a hoverTarget refers to.
type hoverKind int
const (
hoverNone hoverKind = iota
// hoverTranscript: a specialist card or collapse/expand toggle header in the
// main transcript body (or the subchat child view), identified by bodyY.
hoverTranscript
// hoverSidebarAgent: a clickable AGENTS sidebar row, identified by sessionID.
hoverSidebarAgent
// hoverPlanStep: a plan step row in the sidebar's PLAN section, identified by
// step index.
hoverPlanStep
// hoverFileRow: a touched-file row in the sidebar's FILES section, identified
// by path.
hoverFileRow
)
// hoverTarget identifies the single clickable row (if any) currently under the
// mouse cursor with no button pressed. Exactly one of bodyY/sessionID/stepIndex is
// meaningful, discriminated by kind.
//
// The sidebar fields deliberately store a STABLE IDENTITY (sessionID, stepIndex),
// not a raw line offset: the AGENTS/PLAN section sizes can change between when a
// hover is detected and when the sidebar next renders (a swarm member's linger
// window elapsing, a plan step completing) with no mouse motion in between to
// re-resolve it. A cached raw offset would then silently point at whatever
// unrelated row now occupies that slot. hoveredSidebarLineOffset (sidebar.go)
// re-resolves the CURRENT line offset for this identity fresh on every render, so
// a row that's gone simply doesn't highlight rather than a wrong one lighting up.
type hoverTarget struct {
kind hoverKind
bodyY int // hoverTranscript
sessionID string // hoverSidebarAgent
stepIndex int // hoverPlanStep
filePath string // hoverFileRow
}
// mouseHover reports whether msg is a plain cursor-movement event with NO button
// pressed — a hover, not a drag. Requires AllMotion mouse reporting (see View's
// wantsMouseCapture branch); CellMotion only reports motion while a button is held,
// so this predicate would never match under the old mouse mode.
func mouseHover(msg tea.MouseMsg) bool {
if _, ok := msg.(tea.MouseMotionMsg); !ok {
return false
}
return mouseEvent(msg).Button == tea.MouseNone
}
// updateHoverTarget resolves visible transcript targets for a hover motion.
// Context data lives in the run-details overlay; the former sidebar rail is not
// rendered and therefore cannot own a hover target.
func (m model) updateHoverTarget(msg tea.MouseMsg) model {
if line, ok := m.transcriptLineAtMouse(msg); ok {
// A permission option reuses its OWN existing keyboard-cursor highlight
// (see hoverPermissionOption) rather than the m.hover mechanism, so there's
// nothing further to set here.
if line.permOption {
return m.hoverPermissionOption(line.permChoice).withHover(hoverTarget{})
}
// Only a card or a collapse/expand toggle header is "clickable" here; plain
// selectable text (e.g. a user/assistant message) isn't, so hovering over
// it must not light up as if it were.
if line.specialistCard || line.toggle {
return m.withHover(hoverTarget{kind: hoverTranscript, bodyY: line.bodyY})
}
}
return m.withHover(hoverTarget{})
}
// hoverPermissionOption moves the pending permission prompt's cursor to the
// hovered option, reusing its EXISTING keyboard-navigation highlight (the same
// one Tab/Shift+Tab/arrow keys move) instead of new styling — the popup already
// renders whichever option .cursor points at differently. A no-op when no
// permission prompt is pending or the choice isn't found (defensive; every
// permOption line's choice always comes from permissionOptions(request) in the
// first place).
func (m model) hoverPermissionOption(choice permissionDecision) model {
if m.pendingPermission == nil {
return m
}
for index, option := range permissionOptions(m.pendingPermission.request) {
if option.choice == choice {
m.pendingPermission.cursor = index
return m
}
}
return m
}
func (m model) withHover(t hoverTarget) model {
m.hover = t
return m
}
// clearHover drops any hover highlight. Called wherever the underlying content a
// hover target refers to may no longer mean the same thing on the next render — a
// wheel-scroll (the cursor's mapping to a bodyY is only recomputed on the NEXT
// motion event; without clearing, a stale bodyY could coincidentally match a
// DIFFERENT row after the viewport shifts, highlighting the wrong thing) or a
// subchat toggle (bodyY numbering is entirely different between the parent
// transcript and a child session).
func (m model) clearHover() model {
if m.hover.kind == hoverNone {
return m
}
m.hover = hoverTarget{}
return m
}