CodeSnip Old Code (SVN)

Fixed TSnippetExtraHelper.BuildActiveText overload that processes REML to correctly handle code that produces code where some text is embedded within block level tags while other text is not. All text not in block level tags is placed within paragraph tags.

Changed TSnippetsEditorDlg.BuildExtraActiveText method to no longer pre-process REML code string to compress spaces. No longer needed now that tagged text lexer processes spaces correctly itself.

Modified TTaggedTextLexer to emit only first of any sequence of white space characters. Any intervening tags are ignored for this purpose so that "x <em> y </em> x" is parsed as "x <em>y </em>x". A side effect is that a text token with an empty string can be returned.

Changed TREMLReader.Parse to ignore empty text tokens received from lexer: no active text element is generated for empty string.

Fixed problem with REML parsing where nested block level tags in REML, such as <p>...<p>...</p>...</p> were being permitted.
Required new TActiveTextActionElemKinds set type in UActiveText unit.

NOTE: This fix may cause exceptions when loading any old REML code when block level tags are nested. All main database is OK, but some user defined snippets could be affected and their extra information will be lost.

CodeSnip version 4.0 alpha 3

Change log updated to record changes in release 4.0 alpha 3.
Removed notes from v4-notes.txt that are now in change log.

Updated ready for Release 4.0 alpha 3
* Set version numbers to 3.98.3 build 199.

Recorded details of changes since r1882.

New file format documentation file that describes the format of selection files (used when selections (search result sets) are saved to disk.

Changed File | Save Snippet shortcut to Shift+Ctrl+N from Ctrl+S which has been assigned to Database | Save User Database.
Updated File and Database menu help topics re shortcut key changes.