| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
focused input, it doesn't look good on init
|
|
- Fix reset-search action so it resets list page number to 1
|
|
database.
The object must be cleaned up before storing in DB, because some code
(out of rcube_message and rcube_imap classes) is changing objects data
e.g. parts mimetype or body. Now it's become a problem because we're
saving the object in cache at shutdown (after all modifications).
- Make stored message object smaller by removing some redundant data.
|
|
|
|
Indexes are stored in a separate table, so there's no need to store all messages in a folder
Added threads data caching
Flags are stored separately, so flag change doesn't cause DELETE+INSERT, just UPDATE
- Partial QRESYNC support
- Improved FETCH response handling
- Improvements in response tokenization method
|
|
|
|
|
|
|
|
|
|
|
|
- Added spellchecker exceptions dictionary (shared or per-user)
- Added possibility to ignore words containing caps, numbers, symbols (spellcheck_ignore_* options)
|
|
to trigger session gc handlers
|
|
|
|
more appropriate message to the user
|
|
print error to the log
|
|
|
|
'user_create' hook arguments (#1488024)
|
|
|
|
|
|
|
|
clearing search (#1488066)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
- pl_PL update
- Fix label: databaserror -> dberror
|
|
|
|
|
|
|
|
SESSION['timezone'] when saving prefs
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Explorer
|
|
changelog
|
|
instead of 'enctype'
|
