CodeSnip Old Code (SVN)

Added Classes to uses cause to permit inlining of TObjectList methods.

Added conditionally defined $WARN directive to switch off warning about explicit string cast from string to ANSI string potentially loosing data. The referenced cast is OK.

* Changed UEncrypt unit's Encode routine to define a character map as an array of AnsiChar instead of Char.
* Commented changes to original unit.

Made implicit casts from AnsiString to UnicodeString to explicit casts in order to document the intention.

* Changed character set testing code in several units to use char set helper functions added to UUnicodeHelper.pas instead of testing directly against literal or constant sets. This prevents Delphi 2010 warnings about WideChar being reduced to byte char in set expressions.
* Changed use of UpCase built-in to use ToUpperCase function added to UUnicodeHelper.pas.
* Added various set handling routines along with ToUpperCase function. These mainly call TCharacter methods on Unicode Delphis and have custom implementations on other Delphis.

Modified TAppInfo.GenerateKey to be Unicode compatible: now uses new ASCIIBytesOf routine (added to UUnicodeHelper unit) to ensure program key contains only ASCII characters.

* TUserDBBackup.Backup now uses TBytes version of TChecksum.Calculate to get checksum of a file's contents.
* Changed TUserDBBackup.Restore to work with new TDataStreamString string type used by classes in UDataStreamIO unit.

* Changed TFileUpdater to work with new TDataStreamString string type used by classes in UDataStreamIO unit.
* TFileUpdater now uses TCheckSum.Compare rather than TChecksum.Calculate when checking checksums.

* Added support for Latin1 code page: string definition, code page constant, custom encoding class and Latin1BytesOf helper function.
* Defined TBytes for non-Unicode Delphis.
* Provided BytesOf function for non-Unicode Delphis.

Made TBytes version of TChecksum.Calculate compilable in non-Unicode Delphis as well as Unicode versions.