InnoKB Server ChangeLog
- Fixed: security issues; [Urgent]
- Fixed: the attachment list view may unexpectedly disappear from in the main frame layout;
- Fixed: '_linklist' command not enabled for anonymous to retrieve related info items;
- Fixed: inserting images into HTML content at caret using IE web browser;
- Improvement: enabled '_profile' command for anonymous users to save preferences on server;
- Relocated the query history list at the last band of tool bar;
- Fixed a bug that was possibly causing bad performance during UTF-8 decoding;
- Added a statistics page automatically counting info items for each editors;
- Auto-detection of UTF-8 encoded text files, no matter if BOM exists in file headers;
- Some improvements on text encoding for consistency, see below notes:
If you'd never use any Non-ASCII content with InnoKB, you may ignore this note.
Otherwise, in order for InnoKB Server to better manipulate multilingual text other than US-ASCII,
please make sure that the appropriate language packs are installed properly on your system.
But there're a few notes using ANSI CodePage on different OS platforms;
- Within InnoKB, all text content are encoded in UTF-8, including item titles,
Html content, filenames, labels etc., and all attachments are stored as they originally are
without any coversions, and InnoKB Server will attempt to parse the attachments
and index them for fast search.
In fact, InnoKB Server accepts any files as attachments, that may be not encoded in Unicode
but possibly in ANSI CodePage. If in this case, a translation from ANSI CodePage to Unicode is
required for parsing, being indexed and searchable. So this is the real place where we will care for
Character Set and text Encoding.
Otherwise, you would never care for any Character Set or Encoding with the content you input
via the editor's web interface, you can save, view, search all these text content without a problem
on all the supported OS platforms. This is also true for any attachments that are encoded in
UCS2-LE, UCS2-BE and UTF-8. That's to say, any Unicode based attachments will be fully indexed
and searchable as well. The UCS2 encoded files require BOM in file header, but not the UTF-8 encoded files,
which can be auto-detected.
Anyway, if you'd have to upload any files as attachments that are encoded in ANSI CodePage,
you'd need to set the system locale appropriately or set the 'LC_CTYPE' setting in the 'innokb_config.xml' config file.
The results also depend on the target server OS platform. On Windows and GNU/Linux, InnoKB Server
attempts to translate ANSI CodePage into Unicode by using the system locale or the application sepcific
LC_CTYPE setting. However, on FreeBSD and Solaris, InnoKB Server doesn't translate any non-ACSII text into
Unicode, as results, those non-ASCII text files may be not fully indexable or searchable. Well,
it does not affect other operations, e.g. viewing, downloading, etc.
In short, on Windows, please adjust the system CodePage settings appropriately via 'Control Panel ->
Regional and Language Options -> Advanced -> Language for non-Unicode Programs'; On GNU/Linux, please adjust
the locale settings or set the LC_CTYPE setting in 'innokb_config.xml'. This makes possiblities for
InnoKB Server to parse and index text attachments encoded in the appropriate ANSI CodePage. But on FreeBSD
and Solaris, all user input including file uploads are assumed Unicode encoding (either of UCS2-LE, UCS2-BE
and UTF-8), other encoding based files should be first converted into Unicode before uploading to InnoKB Server,
so that they can be fully indexed and searchable.
Another place that may care for Character Set or Encoding is the Console or log viewer. On Windows and Linux,
it is somewhat easy to correct display for non-ASCII characters by properly setting the system locale.
However, on FreeBSD and Solaris, it currently ignores non-ASCII characters on Console.
- Added: the 'item links' feature which allows you to add related items; As the new commands '_linklist, _link, _unlink' added, the privileges should be adjusted accordingly for this feature to work normally;
- Added: the 'show label path' option which allows to show/hide the full label path; If this option is disabled, only the leaf labels are shown in the items list;
- Added: the up/down arrow icon beside column headers of the items list to indicate if it's sorted in the order of increasing or decreasing;
- Added: the right-arrow icon in the items list to indicate which one is currently opened;
- Added: a few quick links 'newest, updated, labelled, unlabelled, any', which are handy to run the frequent queries;
- Added: the 'label with selection' menu item, which allows to label info items by using the selected text within its content;
- Added: the query history list, which keeps all recent queries in the list, so that you can go back to a history query with ease;
- Added: size limit added for file uploads, 10MB by default and customizeable;
- Added: detect bad http request headers which are too long;
- Improved: when adding new info items, clicking the 'label' box opens the label tree;
- Improved: in the query form, multiple labels can be specified in any queries;
- Improved: in the query form, date created and date modified can be specified in any queries;
- Fixed: a bug in htm.dll, which parsed .html source incorrectly, and failed to retrieve text;
- Fixed: solved the 'pipe broken' problem on Sun Solaris for sparc;
- Fixed: the Enter accepted within HTML source editing;
Big changes: Unlike v1.x, InnoKB 2.0 introduced the 'labels and sub labels' feature for categorizing info items,
and replaced the old tree outline that is cumbersome to manage a large number of info items on the web browser platform,
If you'd want to migrate databases created in v1.x to v2.0, you'd need to create a label tree,
and apply labels to existing info items accordingly.
The main reason of this change is that mixing
categories and info items into a single tree outline view bloats the data transfer and memory usage for the tree structure,
and could heavily slow down the web interface, especially with those databases containing thousands of info items or more.
Note that the v2.0 database file specs are Not fully back compatible with v1.x, so please be sure to create backups
for the v1.x databases before upgrading to v2.0.
- By using labels and sub labels, info items can also be categorized into the tree structure and have more flexibilities.
- InnoKB server 2.0 writes logs specifically for audit purpose, and log filenames are date-based and customizable.
It's more convenient within v2.0 to select and filter info items by running specific queries.
Clicking on a label automatially runs a query with the label and lists resulting info items.
- Ability to search across databases, and labels can be combined into search critera.
- Ability to sort search results by clicking on the list headers, clicking the headers again toggles the sequence between increasing and decreasing.
- Ability to save preferences in user's profile on the server.
- The improved web interfaces for viewers, editors and administrators work more smoothly in the web browser platform.
The indexing process in the administrator interface has been improved. Unlike v1.0, this version will not lock up
the server any more during the indexing process, and the server continues responding to Ajax requests from clients,
and the indexing priority can be adjustable at any time.
New localization mechanism provided, v2.0 uses .INI files for viewer, editor and administrator interfaces, instead of .JS files,
for the convenience of translating the web interfaces to the end-user's native language.
The default localization files in English are included in the server pack, which can be the templates for translating the interfaces.
- The viewer and editor privilege schemes provided, so it'd be easier to grant privileges for viewers and editors.