ARBIL 2.3.x RELEASE NOTES
----------------------------------------
$Date: 2011-09-27 14:50:05 +0200 (Tue, 27 Sep 2011) $ 
$Revision: 26544 $ (this means that the descriptions below apply to Arbil 2.3.x with x >= revision)

Contacts:
- Peter Withers (peter.withers@mpi.nl)
- Twan Goosen (twan.goosen@mpi.nl)
----------------------------------------

Below is an overview of changes in Arbil since the previous minor version (Arbil 2.2). Where 
applicable, related ticket numbers from TRAC are included at the end of the description.

** ADDITIONS **

The following functionality has been added to Arbil:

- SEARCH: The 'column' field in the search window is now a drop-down that will contain a list of
previously entered column names, as well as a number of pre-defined common values. Users can add any
column name to this list as well as remove arbitrary items. (#574, #633)

- TABLES: Individual columns in table views can be optionally set to a fixed width. This can be used
on arbitrary open tables (and will be kept across sessions) as well as in saved column views. (#678)

- CONTROLLED VOCABULARIES: Editing controlled vocabulary fields (e.g. 'Genre') has been extended
with 'type-ahead' (autocomplete) functionality, both for single value and list vocabularies. (#584)

- CONTROLLED VOCABULARIES: In the table view and long field editor, an icon is shown indicating the
type of vocabulary (open/closed, single value/list) on the field.

- EDIT ALL METADATA: A new context menu option 'Edit/view all metadata' is available for metadata
files in the tree view (except for Corpus nodes). It opens a window containing tables for the 
selected node all of its subnodes in an hierarchical view. (#603)

- TYPECHECKER CONFIGURATION: A custom typechecker filetypes configuration can be used by putting a
filetypes.txt file in the Arbil working directory. It will then automatically be used instead of the
internal default configuration (#905).

- WINDOWS: A new option 'Close all windows' in the Windows menu on the menu bar closes all floating
windows on the application workspace at once. (#570)


** CHANGES **

These parts of the Arbil functionality have changed:

- SEARCH: The 'search' options is now also available in the context menu on the local corpus node in
the tree view. This means that one can now easily search through all data in the local corpus at 
once. (#68)

- METADATA STRUCTURE: Metadata nodes can now be moved between metada files using drag/drop (#658)
- TABLE MENU: 'Save changes' has been added to the table context menu if it there are changes to
be saved. (#596)

- TABLE ORDERING: Tables can be sorted according to field order, field name or field value. In the
latter case, the table is now sorted by the value stored on disk and not the actual value to prevent
the rows from jumping after an edit that causes a change in the alphabetical order. (#948)

- LONG FIELD EDITOR: The long field editor (LFE) now has 'previous' and 'next' buttons that allow the
user to quickly navigate the fields of a given node in a single LFE window. (#622)

- LONG FIELD EDITOR: The LFE can now be opened on a selected node in the tree view. It will open on
the first field; the user can then use the previous and next buttons to navigate the fields.

- CMDI SCHEMA: Several schema changes in CMDI (Clarin Metadata Infrastructure) are now reflected in
Arbil so that it generates valid metadata instance files.

- RESOURCE LINK: Resources with no specified resource link no longer get the 'file not found' icon
type (a red diagonal line) but rather a non-specified mark (a red question mark). (#676)

- NODE SORTING: Nodes that contain a numeral value in their displayed name (e.g. Node1, Node2)
are now sorted by the value of this numeral if they are otherwise equal. So Node10 follows Node2
even though character-by-character the latter comes first. This applies to the tree views as well
as tables. (#557)

** FIXES **

Among other, mostly minor, bugfixes, the following issues have been resolved:

- MOVING SESSIONS: When moving multiple sessions between corpus nodes by drag and drop,
a bug would cause sessions to 'disappear' (i.e. be removed from the source corpus but not
linked to the destination) on some platforms. This has been resolved. (#925)

- FAVOURITES: Favourites of a subnode type (e.g. an Actor) could not be dragged onto their matching
parent node (e.g. Actors), but instead had to be dragged onto the root node of the file (Session).
This has been resolved. (#942)

- TREE MENU: Subnodes that are not allowed to be added to a given node according to te schema appear
as disabled in the 'Add' submenu of the context menu on nodes in the tree. (#111)

- WINDOWS: The state of floating windows (tables, search results, subnode views) as well as docked
components is stored across sessions including size and position. (#33)

- CHARACTERS IN FILENAME: Including files with non-ASCII characters in their filename used to cause
issues on some platforms. This has been fixed. (#782)

- QUIT ON MacOS: Arbil now fully supports native methods for exiting the applications on MacOS. This
means that using the 'Quit' option on the application menu or pression CMD-Q will ask for 
confirmation if there are unsaved changes and store the application state before actually quiting.
(#600, MacOS only)

- CMDI: CONTROLLED VOCABULARIES: If the CV provides descriptions, they are used as display value,
while the actual value (e.g. code) is written to the XML. (#997)

- CMDI RESOURCES: When resource are added to CMDI metadata files, this is now also visible in the
tree view. The resources can be added and deleted. Multiple references to a single resource result
in a single resource proxy being referenced multiple times (rather than copies of the same 
resource). (#979)

- IMDI: FOLLOW-UP VOCABULARIES: When selecting a Genre on Content, the proper Subgenre options are 
now shown. (#6)

- IMDI: MULTIPLE VALUES: Multiple values are now possible for Location.Region and all Description
fields.

** KNOWN ISSUES **

Among others, the following issues have been described and will be addressed in future version of
Arbil:

- CMDI: DATA CATEGORIES: Field descriptions should be retrieved from ISOcat but currently are not.
(#595)

- IMPORT RESOURCES: Resources that are protected and require authentication can currently not be
imported into the local corpus since the authentication mechanism is broken. (#679)

- SHOW IMDI FORMATTED: When showing an IMDI metadata file 'IMDI formatted', the formatting does
not look as expected. (#899)

- PERFORMANCE: After importing (certain) corpus nodes, CPU load gets and stays high. (#929)

- STABILITY: During search, memory usage may exceed the limits of the Java virtual machine which
could cause the application to terminate unexpectedly.


