export-runner: classify a completed-but-empty backtest as a data hole, not a failure

The first full production run aborted on Copper 2017-04: a month with no data
*inside* the symbol's range. cTrader does not emit the clean "No historical data"
message for an in-range hole; it runs the backtest to its `}` summary but with a
degenerate empty result and no EXPORT_SUCCESS. run_one misread that as a transient
failure and hard-aborted after retries.

Reaching the `}` summary means the backtest completed cleanly, so the absence of an
export marker is an empty month (rc 3, skipped), not a transient failure. A genuine
transient failure never reaches `}`, which is what distinguishes the two.

docs/export-runner-quirks.md

@@ -24,7 +24,15 @@ outcome, instead of trusting a single stdout sentinel:
 
 - `DATA_FOUND` / `EXPORT_SUCCESS` → **success** (rc 0)
 - `No historical data for the specified period` → **clean-empty** (rc 3, skipped)
-- anything else — timeout, `ERROR:`, a result block with no verdict, a dead stream → **transient failure** (rc 2, retried then aborted)
+- the backtest reaches its `}` summary but exported nothing → **clean-empty** (rc 3, skipped).
+  This catches a cTrader-side **data hole** (a month with no data *inside* a symbol's
+  range, e.g. Copper 2017-04), which emits a degenerate empty result block
+  (`{ "Equity":, … }`) instead of the clean "No historical data" message — surfaced by
+  the first real production run.
+- bot `ERROR:` (an OnStop exception, e.g. a write failure) → **failure** (rc 2)
+- no `}` terminator at all — timeout, crash, dead stream → **transient failure** (rc 2,
+  retried then aborted). A genuine transient failure never reaches the clean `}` summary,
+  which is what separates it from a data hole.
 
 A produced ZIP is validated (`testzip` + non-empty entry) **before** the `mv`, so a
 truncated artifact from a cut-off run is caught instead of being promoted to a final file.