docs(handoff): record statusline first-review resolution

Update the volatile ledger with the rebased feature head, real TUI PTY
smoke evidence, post-rebase gates, review comment, and recovery state.

Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

docs/active-work.md

@@ -55,28 +55,42 @@ If it does not, stop and repair the remote/fetch configuration.
 
 - Portable branch: `githubsucks/statusline-segments`
 - Feature head:
-  `c3d125ac96b7bf0a8eba60530f8a2adb083cd7a6`
+  `66765b3cbc0044bfd33fc6505b8cd93968315823`
 - Pull request: #125,
   **feat(statusline): composable modeline segments at protocol v18** —
   <https://github.com/levineuwirth/pmacs/pull/125>
-- State: Revision 3's Q#SL1-Q#SL11 and Acceptance 1-27 are implemented,
-  committed, pushed, and awaiting review. The branch is **not merged**.
+- State: first review addressed; Revision 3's Q#SL1-Q#SL11 and
+  Acceptance 1-27 remain implemented and awaiting review. The four
+  feature commits were rebased without conflict onto canonical `main`
+  `bb17ec955e083a56aa937b596906fce84a00533a`. The branch is **not
+  merged**. Review resolution:
+  <https://github.com/levineuwirth/pmacs/pull/125#issuecomment-5036372634>.
 - Scope delivered: strict composable per-window `pmacs.statusline`
   providers; borrow-released evaluation and per-context failure latches;
   legacy-preserving TUI composition; a pure built-in LSP provider;
   dynamic modeline-face inventory; protocol-v18 authoritative
   `StatuslineSegments`; snapshot/version symmetry; and atomic GPU
   validation, face resolution, shaping, clipping, and caching.
-- Full verification, sequential with no flaky rerun: `cargo fmt --check`;
-  workspace/all-target Clippy with `-D warnings`; 1,610 default library
-  tests (3 ignored); 1,784 CRDT library tests (3 ignored); 7 default and
-  8 CRDT statusline acceptance tests; 110 M4 acceptance tests (3
-  ignored, `basedpyright` filtered); 108 required GPU tests; one
-  workspace invocation with 2,704 passing tests across 78 suites (19
-  ignored, `basedpyright` filtered); and `git diff --check`.
+- Real TUI smoke: hermetic XDG roots under `/tmp`, live tmux PTYs at
+  100x24 and verified 46x24, faced left/right custom runs with CJK,
+  combining text, and an injected ESC. The ESC rendered as a space;
+  ordering, clipping, protected `L1:C1 All`, and the separate echo row
+  remained legible without overlap. Both sessions exited 0; focused TUI
+  statusline/mode-line tests passed 6 + 1.
+- Final verification after the rebase, sequential: `cargo fmt --check`;
+  workspace/all-target Clippy with `-D warnings`; 1,617 default library
+  tests (3 ignored); 1,791 CRDT library tests (3 ignored); 7 default and
+  8 CRDT statusline acceptance tests; 114 M4 acceptance tests (3
+  ignored, `basedpyright` filtered); 108 required GPU tests; and
+  `git diff --check`. The first workspace sweep hit only
+  `m6_5_repl_spawns_zsh`; it reproduced on current main. A hermetic
+  wrapper invoking real `/usr/bin/zsh -f` passed isolated, and the
+  resulting full workspace rerun passed 2,715 tests across 78 suites
+  (19 ignored, `basedpyright` filtered). No feature code changed.
 - Recovery state at handoff: `/home/jeans/Repos/active/pmacs-statusline`
   is clean, its local branch and `githubsucks/statusline-segments` both
-  point at the feature head above, and PR #125 targets canonical `main`.
+  point at the feature head above, PR #125 targets canonical `main`, and
+  the PR reports 22 changed files (+5,572/-302) and mergeable.
 
 Recovery worktree on a machine that does not already have the local
 branch: