I’ll turn this around. So how did it
work in your environment? If you are allowed to enter Unicode characters in the
New Class Wizard and those names make it to your source code and your compiler
compiled it fine then we did the right thing, no?
It all comes down to whether the compiler
can deal with the encoding of the file. If there is something fancy we need to
do there to prevent users from causing compile errors, then that will get
tricky since the New Class Wizard would need to know the build settings and the
build settings need to say what encodings are allowed. For now, letting the
characters go through seems reasonable.
From:
cdt-dev-bounces@xxxxxxxxxxx [mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of Neeraj U Bhope
Sent: Monday, September 25, 2006
10:24 AM
To: cdt-dev@xxxxxxxxxxx
Subject: [cdt-dev] C/C++
identifier validation and internationalization
The class creation wizard accepts class names in
different languages. e.g. Arabic, German etc. It does not throw up any error
since the CConventions.validateClassName goes through fine. Is this ok? In one
the C++ standard working papers (http://www.open-std.org/jtc1/sc22/wg21/docs/wp/html/oct97/)
there is an annexure for universal character names
(http://www.open-std.org/jtc1/sc22/wg21/docs/wp/html/oct97/extendid.html). So
the standards seems to have defined the support but what is the current level
of support in different compilers and of course in CDT. Do the validations in
CConventions go through "as designed" or do they go through
accidentaly? :)
--
neeraj
