On Wed, 16 Jul 2008 17:27:44 +0100, Tom Corey
... except for that loud contingent of users who are absolutely
certain that Unicode Will Break Everything And Doom Civilization.
The proposed change of meaning of "string" means that it will allow
Unicode characters to be entered and displayed in every TEdit, TComboBox,
TStringGrid, etc, etc - which is indeed a major advance after a simple
How does this modification in a Delphi front-end application magically
persuade our customer's non-Delphi, flat-record, 1char=1byte, ANSI-only,
widely-distributed database to accept Unicode data?
a) A Chinese/Russian/... user will be able to enter "char>255" into a
TEdit and see the characters displayed correctly (unlike at present) but
that data will appear "mangled" on subsequent read-back because the data
will not (cannot) be stored correctly in the database.
or b) Each of the several thousand data files worldwide is extended to
handle Unicode data
or c) Each and every Edit1, ComboBox2, StringGrid2143, ... is modified to
filter out all non-ANSI data entry.
The first - "What You See Is NOT What You Get Back" - is completely
unacceptable. (Note - this would also be the effect were the filtering
done in the database interface layer)
The second and third options are completely uneconomic - unless CG are
going to issue some magic class helpers for /all/ data entry controls or
add a "Don't accept Unicode entry" property to the Application and/or
FTR I don't recall ever saying this change would "Doom Civilization", but
it certainly won't be good news for the CG CFO if he's expecting any
upgrade revenue from our customer.
Information Management Systems