fix code-review findings in the source-chooser feature

Seven correctness and documentation fixes found in the final review:

- `known_sources()`: check the strtol end pointer and add a round-trip
  guard so a filename that layout_path() could not have produced is
  rejected; without it `foo_8000x.toml` listed as `foo:8000` whose `x`
  would unlink `foo_8000.toml`, a different configured session.
- `chooser_finish()`: destroy the form window before the chooser window;
  a row click or Escape while the form was up left a window whose handlers
  pointed at a dead stack frame (latent use-after-free).
- `choose_source()`: keep the pick_probe timeout id and g_source_remove
  it after the loop; a row picked before the delay elapsed left an armed
  source on the main context pointing at the gone frame.
- `chooser_populate()`: replace a false comment ("not a GtkListBoxRow")
  with the truth — GTK4 auto-wraps, but the wrapped row carries no
  source-index data so the bounds check refuses it.
- `pick_probe()`: collapse the duplicate `known.empty()` / `row==nullptr`
  guards; the second was unreachable and two identical lines for one cause
  makes probe output ambiguous.
- `save_layout()`: state the TOML rule directly in the multicast/interface
  comment rather than cross-referencing the adjacent enabled comment.
- `save_layout_now()`: refuse to write when signals() is empty; a run
  against a dead port receives no CONFIG, so every slot name is empty and
  the save would overwrite a configured session with signals=["",...],
  which the chooser filter refuses, silently removing that source from
  the only list it can be picked from.
- Fix include order in layout_store.cpp (<cstdio> before <cstdlib>).
- Update CLAUDE.md and README.md to document the empty-signals guard and
  fix the probe-output quote to match the emitted text exactly.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

CLAUDE.md

@@ -462,8 +462,8 @@ Two lines of the self-check output are assertions in disguise:
   networking**: write the session file by hand and read the line. With
   `UDPSCOPE_PROBE_NEWHOST=host:port` it drives the form instead — `[sources] form open,
   connect_sensitive=1` then a `picked` line, with no SIGSEGV, which is as much a crash test of a
-  dialog built and destroyed as it is a state check. An empty list must print `picked nothing`
-  rather than dereferencing a null row.
+  dialog built and destroyed as it is a state check. An empty list must print `picked nothing:
+  the list is empty` rather than dereferencing a null row.
 - **The session layout round trip** (`~/.config/udpscope/<host>_<port>.toml`) — run once
   with a drop and a window change, then again with no interaction: the second run's dump
   must read `layout loaded from ...` and `layout: rows=R window_s=W row0=[...]` with the
@@ -1273,4 +1273,10 @@ far the largest file.
   least one signal**, because a probe run leaves a file per throwaway port and all of them hold
   `signals = ["", "", "", ""]`; `known_sources()` splits the filename at the **last** `_`, since
   `sanitize()` turns anything unsafe in a host into one.
+- **`save_layout_now()` refuses to write while `signals()` is empty.** A run against a source
+  whose streamer was down receives no CONFIG, so every slot name stays empty and the save would
+  overwrite a configured session with `signals = ["", ...]`. The chooser then filters that file
+  out — a source whose streamer was merely down during one run would silently vanish from the
+  only list it can be picked from, for ever. Keeping the old file loses nothing: a run that
+  configured nothing had nothing worth persisting.
 - Every source file carries an SPDX MIT header; match it in new files.