|
Revision History |
|
Versions Released in 2002 |
|
|
|
The following is a list of
versions released in 2002.
- Major Update! The Builder has a new graphic user interface. The
application is now a MDI application (multiple windows). Each of the 'tabs' in the
previous version of the GOLD Parser Builder are now separate windows.
- 'Auto Whitespace' parameter was added. In the previous version of the GOLD Parser, the
whitespace terminal was always created when omitted in the grammar. Unfortunately, not all
grammars make use of whitespace. The new parameter is set to 'True' by default, but can be
changed to 'False'. When 'False', the system will not automatically create a whitespace
terminal unless it is manually defined.
- Added the ability to export tables to formatted text. When the tables are saved to the
text file, the information is formatted same way as it is in the GOLD Parser Builder
windows.
- Added the ability to export the parse tree from the Test Grammar Window. When a string
is parsed in the Test Grammar Window, the system creates a graphical parse tree. When the
parse tree is saved, it is formatted using standard text.
- Updated some of the messages displayed by the system.
- The 'Check for Updates' menu opens a different webpage. Rather than open the Download
Page, a page for the current version of GOLD Parser Builder is opened.
- Enhanced printing. When you select 'Print' from the 'File' menu, the system will now
allow you to select the printer. The margin of the printout was increased from 0 to 0.75
inches.
- Control characters will now be displayed with friendly names. When the DFA state diagram
was displayed, control characters would appear as small squares. Now each control
character will be represented by its character name.
- Updated how terminal names are displayed. When a terminal name was displayed, the system
would always put some characters between single quotes. Now the system is far more
intelligent when single quotes should be used.
- Fixed a bug in the predefined character sets The No-Break Space (character #160) was a
member of the {Printable Extended} character set. This character was supposed to be in the
normal {Printable} character set.
- Fixed some tab-order problems.
- Fixed an 'Untitled' filename bug When a new grammar was started, it was tentatively
called '(Untitled)'. Unfortunately, when Save was selected, the system created a file with
this name rather than prompting the user.
- Fixed an save bug when a CGT was opened If a CGT file is opened, the system displays a
text message in the grammar editor describing the nature of the CGT data. If the user
selected 'Save' at this point, the original CGT was overwritten by this message -
destroying it.
- Fixed the syntax when exporting C# enumerated constants. The default values were wrong.
- Fixed a few spelling errors ... and probably added a few new ones too!
|
- Fixed a temporary file bug. When the Builder opens a website, it creates a temporary
file containing the redirect information for the browser. Unfortunately, the system was
failing to delete this file after it was used.
- Fixed a drag-and-drop bug. If an icon, which was not a valid file, was dragged and
dropped on the main window, the program would crash.
- Fixed a minor bug in the ActiveX Parser Engine. The parser was failing to increment the
line counter when line-comments were read.
- Fixed a bug in the ActiveX Parser Engine. When a syntax error occurs, the parser loads
the expected tokens into the Tokens() property. The system failed to load the End Symbol
(if present).
|
- Added the 'TrimReductions' property to the Engine. When enabled, the Parser Engine will
remove unnessary reductions from the parse tree. As a result, the tree will be both
smaller and less memory intensive.
- Added the 'Log Read Tokens' item to the Test Window menu. When selected, the Test Window
will display all the tokens read from the source; otherwise, only Reductions will be
visible.
- The GOLD Parser help is now online.
- Added Java and C# to the Enumerated Constants Window.
- Major update! The time needed for the Engine to perform 'garbage collection' was
greatly improved. When a large number of reductions were created, the system
would take a long time during garbage collection - nearly hanging the system. The system
was taking O(nē) operations to clear the parse tree rather than O(n) operations.
- Increased the default size of the Test Window.
- Rearranged the menus in the Test Window. The menus are now organized by File, Behavior
and Display.
- The Test Window will no longer display very large parse trees. The TreeView control that
is used to display the tree cannot handle excessive amounts of data.
- Fixed a bug in the DFA table construction. The Builder was creating case-sensitive DFA
tables for case-insensitive grammars and vice versa. The problem occurred when the Builder
added support for the Unicode characters from #160 to #255. This bug only occurred in
version 1.0 Beta 19.
|
2001 Revisions
|