fix(13e6): partition + SPI corruption + bootstrap stay-awake

Three problems surfaced during the first 13.3" end-to-end run:

1) LittleFS IntegerDivideByZero on 200 → write /img.bin. Cause: the
   ~3.5 MB SPIFFS in default_16MB.csv can't fit three 960 KB setup
   screens + a 960 KB cached image (~3.84 MB). Switching to a custom
   partitions_13e6.csv with 24 MB LittleFS on the 32 MB flash.

2) Yellow wash across the panel on long SPI bursts. Cause: SPI DMA
   from a PSRAM-backed scratch buffer hits a cache-coherency window —
   the CPU's writes hadn't reached PSRAM yet when DMA read it. Push
   each half in 8 KB chunks through an internal-SRAM (DMA-coherent)
   scratch, and drop the bus clock to 4 MHz to match the 7.3"
   production speed.

3) Bootstrap window (no image yet) was deep-sleeping for 15 s between
   polls — each cycle a ~5 s ROM-boot + Wi-Fi reconnect, so the user
   waited ~20 s × N retries between scanning the setup QR and seeing
   their first photo land. Now normal_operation_impl returns early
   during bootstrap and main.cpp's normal_operation loops with a
   2 s delay, keeping Wi-Fi up. Once the first image arrives, the
   normal scheduled deep sleep takes over.

   Also fixes a related bug Matt called out: a transient TLS hiccup
   during bootstrap was hitting the 5xx fallback path and painting a
   full yellow fill over the green setup QR, leaving the user with
   no claim path. Criterion is now "does /img.bin exist?" (panel has
   something worth showing with a border) rather than "is currentImgId
   set?", so a fresh device with no cached image preserves the setup
   screen through transient network errors.

Diagnostic prints in the panel driver + [op] start/code lines in
normal_operation_impl that proved invaluable during bringup; leaving
them in for now. Tests updated for the new bootstrap semantics
(deep sleep no longer arms on bootstrap-cycle 204/404/5xx); 43/43
native tests pass, 7.3" production build stays byte-identical.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

test/test_normal_operation/test_main.cpp

@@ -141,7 +141,9 @@ void test_fw04_204_no_prior_image_does_not_redraw() {
         "204 must not redraw the QR — panel already holds it from provisioning");
     TEST_ASSERT_EQUAL(0, g_epd_init_count);
     TEST_ASSERT_EQUAL(0, g_epd_draw_image_count);
-    TEST_ASSERT_TRUE(g_deep_sleep_started);
+    // Bootstrap: no deep sleep. Caller (main.cpp normal_operation loop) keeps
+    // WiFi up and retries on a short BOOTSTRAP_RETRY_INTERVAL_MS timer.
+    TEST_ASSERT_FALSE(g_deep_sleep_started);
 }
 
 // FW-04b: 204 after a real image was previously displayed — panel holds the
@@ -157,14 +159,15 @@ void test_fw04b_204_with_prior_image_does_not_redraw() {
     TEST_ASSERT_EQUAL(0, g_epd_fill_count);
 }
 
-// FW-05: 404 — same logic as 204; panel keeps whatever's there.
+// FW-05: 404 — same logic as 204; panel keeps whatever's there. Without a
+// prior image, the bootstrap path also skips deep sleep (caller retries).
 void test_fw05_404_does_not_redraw() {
     g_http_get_code = 404;
     normal_operation_impl(String("mac"), http, String("url"), prefs);
     TEST_ASSERT_EQUAL(0, g_show_setup_qr_count);
     TEST_ASSERT_EQUAL(0, g_epd_init_count);
     TEST_ASSERT_EQUAL(0, g_epd_draw_image_count);
-    TEST_ASSERT_TRUE(g_deep_sleep_started);
+    TEST_ASSERT_FALSE(g_deep_sleep_started);
 }
 
 // FW-06a: 5xx error WITH a cached image → preserve last image and overlay a
@@ -185,18 +188,24 @@ void test_fw06a_error_with_cache_draws_border_not_fill() {
     TEST_ASSERT_EQUAL(1, prefs.getInt(NVS_KEY_ERR_BORDER, -1));
 }
 
-// FW-06b: 5xx error with NO cached image → fall back to full yellow fill so
-// the user still sees a sync-fail signal on a fresh device.
-void test_fw06b_error_without_cache_falls_back_to_fill() {
+// FW-06b: 5xx error during the bootstrap window (no cached image) MUST NOT
+// paint anything — the panel is holding the green setup-claim QR, which the
+// user still needs to scan. A transient TLS hiccup or server blip wiping it
+// to yellow leaves the recipient stranded with no path to claim the frame.
+// err_border MUST NOT be set either, since there's nothing to "recover" from
+// on the next healthy response.
+void test_fw06b_error_without_cache_preserves_setup_screen() {
     g_http_get_code = 500;
     // LittleFS has no IMAGE_PATH entry
 
     normal_operation_impl(String("mac"), http, String("url"), prefs);
 
     TEST_ASSERT_EQUAL(0, g_epd_draw_border_count);
-    TEST_ASSERT_EQUAL(1, g_epd_fill_count);
-    TEST_ASSERT_EQUAL(COLOR_YELLOW, g_epd_fill_last_color);
-    TEST_ASSERT_EQUAL(1, prefs.getInt(NVS_KEY_ERR_BORDER, -1));
+    TEST_ASSERT_EQUAL_MESSAGE(0, g_epd_fill_count,
+        "5xx during bootstrap must NOT paint yellow over the setup QR");
+    // err_border NOT set — default sentinel (-1) means key wasn't written.
+    TEST_ASSERT_EQUAL(-1, prefs.getInt(NVS_KEY_ERR_BORDER, -1));
+    TEST_ASSERT_FALSE(g_deep_sleep_started);
 }
 
 // FW-06c: 304 with err_border flag set (sync recovered after a previous
@@ -389,17 +398,20 @@ void test_fw_deep_sleep_arms_ext0_button_wakeup() {
 }
 
 // FW-FIRST-IMG-A: device has never received an image (img_id = -1) AND the
-// poll didn't deliver one (e.g. 204 because no images approved yet). Sleep
-// must be the 15s bootstrap interval, NOT whatever the server's schedule
-// says. Without this, a fresh device on a noon-daily schedule sits dark
-// for up to 24 h before the first photo lands.
-void test_fw_first_image_bootstrap_polls_at_15s_when_no_image_yet() {
+// poll didn't deliver one (e.g. 204 because no images approved yet). The
+// bootstrap path must SKIP deep sleep entirely — the caller (main.cpp's
+// normal_operation loop) keeps WiFi alive and re-invokes the function on
+// a short BOOTSTRAP_RETRY_INTERVAL_MS timer so the user doesn't watch a
+// ~5 s deep-sleep + wifi-reconnect on every "no image yet" cycle.
+// Without this contract, a fresh device on a noon-daily schedule would
+// deep-sleep for the server's 6-hour interval and sit dark all day.
+void test_fw_first_image_bootstrap_skips_deep_sleep() {
     g_http_get_code = 204;
     // Server tries to set a 6-hour interval for the user's noon-daily schedule.
     g_http_response_headers["X-Interval-Ms"] = String((unsigned long)(6ULL * 60 * 60 * 1000)).c_str();
     // No img_id in NVS → device has never seen an image.
     normal_operation_impl(String("mac"), http, String("url"), prefs);
-    TEST_ASSERT_EQUAL_UINT64(FIRST_IMAGE_POLL_INTERVAL_MS * 1000ULL, g_sleep_us);
+    TEST_ASSERT_FALSE(g_deep_sleep_started);
 }
 
 // FW-FIRST-IMG-B: once we've persisted an image (200 path wrote img_id), the
@@ -542,7 +554,7 @@ int main(int argc, char** argv) {
     RUN_TEST(test_fw04b_204_with_prior_image_does_not_redraw);
     RUN_TEST(test_fw05_404_does_not_redraw);
     RUN_TEST(test_fw06a_error_with_cache_draws_border_not_fill);
-    RUN_TEST(test_fw06b_error_without_cache_falls_back_to_fill);
+    RUN_TEST(test_fw06b_error_without_cache_preserves_setup_screen);
     RUN_TEST(test_fw06c_304_after_error_repaints_clean);
     RUN_TEST(test_fw06d_304_steady_state_does_not_fill_yellow);
     RUN_TEST(test_fw06e_200_after_error_clears_flag);
@@ -560,7 +572,7 @@ int main(int argc, char** argv) {
     RUN_TEST(test_fw_no_flag_means_no_header);
     RUN_TEST(test_fw_X_Claimed_response_clears_flag);
     RUN_TEST(test_fw_no_X_Claimed_response_keeps_flag);
-    RUN_TEST(test_fw_first_image_bootstrap_polls_at_15s_when_no_image_yet);
+    RUN_TEST(test_fw_first_image_bootstrap_skips_deep_sleep);
     RUN_TEST(test_fw_first_image_bootstrap_clears_after_200);
     RUN_TEST(test_fw_first_image_just_arrived_uses_server_interval);
     RUN_TEST(test_fw_deep_sleep_arms_ext0_button_wakeup);