CodeSnip Old Code (SVN)

Replaced printable characters specified in form #xxx with actual character literal. This to avoid cases where Unicode code point for a character is different to Ansi / Latin-1 code point.

* Rationalised TDataStreamWriter code, removing redundant methods.
* Changed Unicode only TDataStreamWriter methods that write strings to specifically use UnicodeString type.
* Revised and completed commenting of TDataStreamWriter.

* Extended TLatin1Encoding class with a class property the provides a singleton instance.
* Added Latin1Encoding encoding function that provide alternative access to TLatin1Encoding singleton.
* Removed unused IsLatin1Char function.
* Reimplemented ANSI version of ASCIIBytesOf function to replace non-ASCII chars with '?' character.
* Documented the UUnicodeHelper.pas unit.

* Changed TUserDBBackup.Backup to convert string for which checksum is required to Latin1 bytes using Latin1Bytes from UUnicodeHelper.pas rather than BytesOf.
* Removed Ansi compiler only implementation of BytesOf from public section of UUnicodeHelper.pas and renamed implementation as AnsiStringBytesOf for internal use only.

* Replaced TDataStreamReader string type with Latin1String.
* Deleted TDataStreamReader from UDataStreamIO.pas.

Added some new general TODOs.

* Removed TODO comments that simply documented changes made to compile with Delphi 2010.
* Removed any commented out code.

Delete redundant private IsWhiteSpace routine from UUtils unit. Version in UUnicodeHelper unit now used in preference.

Added routines to UUnicodeHelper unit to safely convert a string to a Latin-1 encoded AnsiString and to check if a character is Latin-1.

Modified code that reads and writes encrypted data to ensure the strings to be encrypted and decrypted are safely converted to the Latin1 code page.