CodeSnip Old Code (SVN)

Changed code that restores files so that the restored files are created with the specified encoding.

Modified call to TStringStream constructor in TCopyInfoMgr.GenerateDoc so that string stream does not free the passed encoding. This avoids a potential bug. It hasn't manifested itself because TCopyInfoMgr.GenerateDoc is at present only called with standard format encodings which are never freed in TStringStream

Fixed bug in overloaded versions of StreamToString and StringToStream that take and Encoding parameter where TStringStream was freeing the encoding.

Heavily refactored code that reads and parses backup file. Split out loading and parsing of backup file into helper classes - there is one class for each file format version.

Prevented possible bug in TDataStreamReader when attempt is made to read 0 or -ve length strings.

* Renamed all methods of TDataStreamReader that return ANSI strings with "Ansi" in name and made them return RawByteString instead of Windows1252String.
* Added new Read methods to TDataStreamReader that are named as old ANSI read methods that return Unicode strings.
* TDataStreamReader uses a TEncoding object to get Unicode strings from the stream. User can pass this object in constructor.
* Made minimal changes to all units that use TDataStreamReader to use the renamed ANSI string reader methods. Functionality not changed.

Fixed bug that was adding duplicate entries to code page value map - default code page was duplicating one of the other code pages.

Added a new static class, TEncodingHelper, to the unit.

Updated database download web service and updating code to work with v5 of the codesnip-updt web service on delphidabbler.com. Web service no longer sends an MD5 checksum as part of the download. Instead it includes a checksum in the HTTP headers, which is automatically handled by the UTHHTPEx unit.

Fixed bug where status bar manager was causing an AV by being called after managed status bar has been destroyed.