pureboot: review-pass fixes to the host tool and device runner

pureboot.py: reject an empty image file with a clear error instead of
an IndexError deep in the vector-surgery planner; tighten the erase
docstring (order is irrelevant there — every target byte is the same
value, unlike a real flash where page 0 must go last).

pureboot_device.c: the GPIO bridge's bit_cycles used plain truncating
division where the firmware computes its own bit period with
round-to-nearest (uart.hpp: (Clock.hz + Baud.bd/2)/Baud.bd) — one
cycle off per bit on both tinies, harmless in practice but needless
drift against a firmware built to a different constant. Matched
exactly. Also clear the queued-bytes/decode-in-progress bridge state
on the test-only reset signal, so a future reset-mid-transfer scenario
can't feed a freshly reset chip bytes queued for its previous life.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

pureboot/pureboot.py

@@ -193,6 +193,8 @@ def load_image(path):
     """Raw binary, or Intel HEX by extension (.hex/.ihx/.ihex)."""
     data = open(path, "rb").read()
     if not path.lower().endswith((".hex", ".ihx", ".ihex")):
+        if not data:
+            raise Error(f"{path}: empty image")
         return data
     memory = {}
     for number, line in enumerate(data.decode("ascii", "replace").splitlines(), 1):
@@ -284,8 +286,9 @@ def covered(pages, skip_blank):
 
 
 def op_erase_flash(loader):
-    """0xff over the whole application area, page 0 first — an interrupted
-    erase leaves the entry word blank and the chip still boots the loader."""
+    """0xff over the whole application area. Order does not matter here —
+    every target byte is the same value — and a blank word 0 still falls
+    through to the loader, so an interruption is harmless."""
     blank = bytes([0xFF] * loader.info.page)
     for address in range(0, loader.info.base, loader.info.page):
         loader.write_page(address, blank)