Use correct trait and update release notes

README.md

@@ -85,11 +85,10 @@ When using peek_many, you can look at all data from oldest to newest.
 ### Unreleased
 
 - *Breaking* Data CRC has been upgraded to 32-bit from 16-bit. Turns out 16-bit has too many collisions.
-  This increases the item header size from 6 to 8.
-- Closed pages that were partially erased now can be detected and repaired. This comes at the cost of always
-  reading both page markers to determine the page state.
-- Item migration that happens in `map` that is interrupted by a shutoff can now be repaired.
-- Made changes to queue to better survive shutoffs
+  This increases the item header size from 6 to 8. The CRC was also moved to the front of the header to
+  aid with shutdown/cancellation issues.
+- When the state is corrupted, many issues can now be repaired with the repair functions in the map and queue modules
+- Made changes to the entire to better survive shutoffs
 
 ### 0.6.2 - 22-12-23
 

src/map.rs

@@ -106,8 +106,6 @@
 
 use core::ops::ControlFlow;
 
-use embedded_storage::nor_flash::MultiwriteNorFlash;
-
 use crate::item::{find_next_free_item_spot, read_items, Item, ItemHeader};
 
 use super::*;
@@ -601,7 +599,7 @@ fn migrate_items<I: StorageItem, S: NorFlash>(
 /// If this function or the function call after this crate returns [Error::Corrupted], then it's unlikely
 /// that the state can be recovered. To at least make everything function again at the cost of losing the data,
 /// erase the flash range.
-pub fn try_repair<I: StorageItem, S: MultiwriteNorFlash>(
+pub fn try_repair<I: StorageItem, S: NorFlash>(
     flash: &mut S,
     flash_range: Range<u32>,
     data_buffer: &mut [u8],