fix(terminal): trim the padding and shared indent out of a copied selection

xterm hands back whole screen rows and trims only the cells that were
never written to, so the real spaces a full-screen TUI paints across the
unused part of a row count as content and reach the clipboard. Measured
against Claude Code in a 282-column pane, single lines arrived carrying
138 trailing spaces, and every line carried the two-space transcript
indent as well. Windows Terminal, iTerm2 and GNOME Terminal all trim that
for you, decideAutoCopy already calls a wall of spaces "never what the
gesture meant", and _selectTouchSelectionLine already treats those cells
as padding — the mouse and keyboard paths never had the same rule.

CodemanCopySelection.clean lives in constants.js beside decideAutoCopy,
its pure sibling. It drops the trailing run from each line, and removes
the leading run only where every selected row shares one. A selection of
a single row keeps its run, because one row shares nothing with anything
and stripping it would silently reindent one line of `git log` body text
or one line out of `less`. A drag that began inside a row keeps its
partial first line untouched and out of the measurement, which otherwise
pins the shared run to zero and leaves every following row indented.

Every pass over a line is a scan rather than a regex. `/[ \t]+(\r?)$/` is
quadratic on a line whose spaces are followed by a non-space character,
which is what right-aligned or centred TUI content looks like: measured
over 50 000 rows with a 280-column run it took 2.9s, against 1.3ms for
the scan, and a 2 000-column run took 16s. The scan is also the faster of
the two on an ordinary padded row.

cleanedTerminalSelection in terminal-ui.js is the half that needs the
live terminal. It returns a COLUMN selection untouched: Alt+drag makes
one, and a rectangle's rows lining up is the point of the gesture, so
both halves of the clean would destroy it. xterm exposes the mode nowhere
public, so the check reads terminal._core._selectionService, the way this
file already reads terminal._core for cell dimensions, and cleans
normally if a future xterm renames the field. A test pins that assumption
against the library rather than against a stub repeating the literal.

The Ctrl+C chord decides on the cleaned selection, not the raw one. A
drag across the blank part of a row selects real padding spaces, so the
raw text is truthy, and testing it would spend that press on a copy of
nothing and make the user press again to interrupt. A padding-only
selection is now dropped and the press falls through to the PTY, while
Ctrl+Shift+C still never falls through. copyTerminalSelection gates on
trim() for the same reason, since a multi-row drag across padding cleans
to line breaks alone and a bare newline pasted into a chat composer
submits it.

All four of the main terminal's copy paths go through it: the Ctrl+C
chord, right-click, the phone selection button and Auto Copy. The
browser's own Edit menu copy, a disabled copy shortcut and the subagent
windows still copy raw rows, as they did before, and the invariants doc
now says so rather than claiming every copy is cleaned. Auto Copy
resolves its own toggle before it reads the selection, since it is off by
default and a selection can run to the 50 000-row scrollback ceiling.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

src/web/public/constants.js

@@ -806,6 +806,94 @@ function decideAutoCopy({ enabled, text, lastCopied, pending } = {}) {
   return 'copy';
 }
 
+// The text a copy should put on the clipboard, given xterm's raw selection.
+// Pure: the caller reads the selection and decides the mode, this transforms.
+//
+// xterm hands back whole screen ROWS, and its own trim only drops cells that
+// were never written to. A full-screen TUI writes real spaces across the part
+// of a row it is not using, so that padding counts as content and rides along
+// to the clipboard: measured against Claude Code in a 282-column pane, single
+// lines arrived carrying 138 trailing spaces. Native terminals trim it on copy
+// (Windows Terminal, iTerm2 and GNOME Terminal all do), decideAutoCopy above
+// already calls a wall of spaces "never what the gesture meant", and
+// _selectTouchSelectionLine already treats those cells as padding. This is that
+// same rule for the mouse and keyboard paths, which never had it.
+//
+// The leading run is the other half, and it applies ACROSS ROWS ONLY. A TUI
+// that indents its whole transcript repeats the indent on every row, so a
+// multi-row selection arrives with the chrome baked into each line; only the
+// run every selected row shares is removed, which is a no-op for shell output
+// and keeps the relative indentation of anything nested inside. A selection of
+// ONE row shares nothing with anything, so its leading spaces are content and
+// stay put — otherwise a single line of `git log` body text, or one line out of
+// `less`, would silently lose its indentation.
+//
+// ⚠ The trailing trim takes spaces AND tabs while the leading run counts spaces
+// only, so one tab-led row disables the strip for its whole block. Terminals
+// expand tabs into cells, so a tab should never reach either rule; the
+// asymmetry is deliberate caution rather than an oversight.
+//
+// ⚠ A WRAPPED logical line keeps its continuation indent. xterm appends a
+// wrapped row to the previous entry instead of starting a new line, so rows
+// 2..n of one wrapped line sit mid-string where no line rule can see them. That
+// is inherent to cleaning xterm's output rather than a gap to fix here.
+function cleanCopiedSelection(text, { startedMidRow = false } = {}) {
+  if (typeof text !== 'string' || !text) return '';
+  // Split on \n and leave any \r in place: xterm joins rows with \r\n on
+  // Windows, and the clipboard should keep the endings xterm chose.
+  // Scanned rather than matched, throughout. A selection can run to the 50 000-row
+  // scrollback ceiling, and `/[ \t]+(\r?)$/` is QUADRATIC on a line whose spaces
+  // are followed by any non-space character, which is what right-aligned or
+  // centred TUI content looks like: the engine retries the run from every
+  // whitespace position and backtracks over it. Measured over 50 000 rows with a
+  // 280-column run, that regex took 2.9s against 1.3ms for the scan below, and a
+  // 2 000-column run took 16s. It is also the faster of the two on an ordinary
+  // padded row. A length is returned rather than a trimmed string so a
+  // \r-terminated line costs no substring either.
+  const trimEnd = (line) => {
+    let end = line.length;
+    if (end > 0 && line[end - 1] === '\r') end--;
+    let cut = end;
+    while (cut > 0 && (line[cut - 1] === ' ' || line[cut - 1] === '\t')) cut--;
+    return cut === end ? line : line.slice(0, cut) + line.slice(end);
+  };
+  const lines = text.split('\n').map(trimEnd);
+  const bareLen = (line) => (line.endsWith('\r') ? line.length - 1 : line.length);
+  const leadingRun = (line) => {
+    let n = 0;
+    while (n < line.length && line[n] === ' ') n++;
+    return n;
+  };
+
+  // ⚠ Counted over EVERY line, including a partial first line the loop below
+  // skips. A mid-row drag across two rows therefore measures one row and strips
+  // it. That is deliberate: both rows of a wrapped paragraph wear the TUI's
+  // margin, and the drag only hid the first one's. The cost is that a two-row
+  // mid-row drag over genuinely indented content loses that indent.
+  let contentRows = 0;
+  for (const line of lines) if (bareLen(line)) contentRows++;
+  if (contentRows < 2) return lines.join('\n');
+
+  // startedMidRow keeps the first line out of the measurement. A drag that
+  // begins inside a row gives a first line with no leading run at all, which
+  // would otherwise pin the shared run to zero and leave every row after it
+  // still wearing the indent. That partial line is never stripped either.
+  let shared = Infinity;
+  for (let i = startedMidRow ? 1 : 0; i < lines.length; i++) {
+    // A row left blank by the trailing trim says nothing about the indent, and
+    // counting it as zero would disable the strip for the whole block.
+    if (!bareLen(lines[i])) continue;
+    shared = Math.min(shared, leadingRun(lines[i]));
+    if (shared === 0) break;
+  }
+  if (!shared || shared === Infinity) return lines.join('\n');
+  // The clamp is what keeps a blank row's lone \r intact when the shared run is
+  // wider than that row is long.
+  return lines
+    .map((line, i) => (startedMidRow && i === 0 ? line : line.slice(Math.min(shared, leadingRun(line)))))
+    .join('\n');
+}
+
 if (typeof window !== 'undefined') {
   window.WEBGL_FALLBACK = WEBGL_FALLBACK;
   window.evaluateWebGLLongTaskTrip = evaluateWebGLLongTaskTrip;
@@ -854,6 +942,9 @@ if (typeof window !== 'undefined') {
     decide: decideAutoCopy,
     MAX_CHARS: AUTO_COPY_MAX_CHARS,
   };
+  window.CodemanCopySelection = {
+    clean: cleanCopiedSelection,
+  };
   window.CodemanTerminalFont = {
     DEFAULT_STACK: TERMINAL_FONT_DEFAULT_STACK,
     resolve: resolveTerminalFontFamily,