FAQ-0492 How can I test support for interesting Unicode characters?


FAQ-0492 How can I test support for interesting Unicode characters?

» Symbian OS » Developer Knowledgebase » FAQ-0492

Classification:	C++	Category:	Debugging
Created:	04/27/2000	Modified:	06/07/2001
Number:	FAQ-0492
Platform:	Symbian OS v6.0, Symbian OS v6.1

Question:
Are there any simple ways to test that my application is correctly supporting Unicode characters - I have to do some conversion between Unicode and 8-bit characters and I want to ensure that I'm testing the cases where nontrivial Unicode characters are encountered.

Answer:
The most convenient way to test support for Unicode characters with values greater than 256 is to use the Euro symbol.

The Unicode encoding for the Euro is 0x20AC (decimal 8364), and it is supported in the standard EPOC fonts.
The CodePage 1252 encoding for the Euro is 0x80 (decimal 128) and it is supported in some Windows TrueType fonts, and in the EPOC encoding used with the narrow builds (ER5 and before).

Both EPOC and Windows file systems can support file names involving the Euro, and the Euro can be entered from the keyboard by means of a control sequence:

Unicode EPOC: Press and hold the CTRL key, type the digits 8 3 6 4, then release the CTRL key
Windows NT: Press and hold the ALT key, type the digits 0 1 2 8 on the numeric keypad, then release the ALT key
When converting from Unicode into UTF8, the Euro should be translated into the 3-byte sequence 0xE2 0x82 0xAC - note that the final 0xAC is a coincidence and not simply the low byte of the Unicode character!

As an example of how the Euro symbol improves the testing, it should be noted that the descriptor functions

TDes8::Copy(const TDesC16& aDes)
TDes16::Copy(const TDesC8& aDes)

do not do proper character conversion, and will corrupt the Unicode Euro value into 0xAC, which is the Unicode and CP1252 encoding for ¬ ( NOT SIGN ).