-Notmuch 0.8 (2011-09-05)
+Notmuch 0.9 (2011-10-01)
+========================
+
+New, general features
+---------------------
+
+Correct handling of interruptions during "notmuch new"
+
+ "notmuch new" now operates as a series of small, self-consistent
+ transactions, so it can correctly resume after an interruption or
+ crash. Previously, interruption could lose existing tags, fail to
+ detect messages on resume, or leave the database in a state
+ temporarily or permanently inconsistent with the mail store.
+
+Library changes
+---------------
+
+New functions
+
+ notmuch_database_begin_atomic and notmuch_database_end_atomic allow
+ multiple database operations to be performed atomically.
+
+ notmuch_database_find_message_by_filename does exactly what it says.
+
+Python bindings changes
+-----------------------
+
+ - Re-encode python unicode objects to utf-8 before passing back to
+ libnotmuch.
+ - Support Database().begin_atomic()/end_atomic()
+ - Support Database().find_message_by_filename()
+ NB! This needs a db opened in READ-WRITE mode currently, or it will crash
+ the python process. The is a limitation (=bug) of the underlying libnotmuch.
+ - Fixes where we would not throw NotmuchErrors when we should (Justus Winter)
+
+Ruby bindings changes
+---------------------
+
+ Wrap new library functions notmuch_database_{begin,end}_atomic,
+ and add new exception Notmuch::UnbalancedAtomicError. Rename destroy
+ to destroy! according to Ruby naming conventions.
+
+
+Notmuch 0.8 (2011-09-10)
========================
Improved handling of message/rfc822 parts
Added documentation, typo fixes, and improved support for rdoc.
+Unicode, iterator, PEP8 changes for python bindings
+
+ - PEP8 (code formatting) changes for python files.
+ - Remove Tags.__len__ ; see 0.6 release notes for motivation.
+ - Decode headers as UTF8, encode (unicode) database paths as UTF8.
+
Notmuch 0.7 (2011-08-01)
========================
projects / notmuch / blobdiff