- Ground up rewrite for node 0.8. Code is now much more maintainable, readable,
- New architecture: reed -> APIs -> redis connector/filesystem helper/file processor.
- No more blocking code in the library (not necessarily dependencies).
- Articles no longer need to have dashes in the filename.
- Reed will now watch files that end in ".markdown" as well as ".md".
- Reed now properly detects file additions, updates, and removals that happened while it was not running.
reed.refreshmethods deprecated. They can still be called but will only emit a warning. They will be removed in the next version.
reed.removeAllmethod is now atomic (Redis MULTI).
- Method blocking is now much more efficient by using a queue instead of what amounted to fancy spinlock.
removeevents for reed pages.
- Fixed bug where pages API would not open unless the blog portion wasn't open.
- Fixed issue where pages would throw exceptions when creating Redis keys.
lastModifiedvalues will now be exposed as Date objects. They are still stored as Unix timestamp strings in Redis.
- The redis client will now only shut down if both pages and blog portions are closed. So, if both are open, close needs to be called on both to stop reed.
- Removed all page-related events because they did not fire in the same manner as the blog portion. Pages are for a different purpose than blog anyway.
- Added unit tests for Pages API, using Vows.
- Added changelog file to the project to keep track of history.
- Fixed a bug in would cause the library to crash if there was no metadata defined in articles.
- Added the ability to configure reed so that it can connect to redis running on different hosts/ports and use authentication. This is just a passthrough to the redis module's methods, so you can send in other options as well.
reed.allto get all posts in the system, ordered by date.
- Changed metadata properties to be camelCase instead of all lowercase.
- Clarified in readme that reed will not have comments functionality any time soon.