Indeed! This is like in case of CORBA::Long
MICO::UniCodesetConv::decode (CORBA::Buffer &f, CORBA::ULong len,
CORBA::WChar *t, CORBA::Boolean terminate)
method below the code.
Thanks, the fix is applied now!
Karel
Zoltan Bordas wrote:
> I found a problem when using UTF-8 encoding as the native codeset. When the
> response is being decoded the following code causes an access violation:
>
> CORBA::Long
> MICO::UniCodesetConv::decode (CORBA::Buffer &f, CORBA::ULong len,
> CORBA::Char *t,
> CORBA::Boolean terminate)
> {
> char *cp, *to;
>
> CORBA::ULong flen = _from->codepoint_size();
> if (flen == 3)
> flen = 4;
> flen *= len;
>
> if (len * _from->codepoint_size() > f.length()) {
> return -1;
> }
>
> #ifdef __GNUG__
> char _f[flen+1];
> #else
> CORBA::String_var _f;
> if (_from->codepoint_size() != 1)
> _f = CORBA::string_alloc (flen);
> #endif
>
> <<<< Access violation - _f is uninitialized >>>> if (!f.get (_f, len *
> _from->codepoint_size())) {
> return -1;
>
> The condition needs to be removed from the string initialization otherwise we
> are calling the get method to an uninitialized string.
>
> Zoltan
>
> --- codeset.cc.original 2008-07-25 16:41:44.000000000 -0500
> +++ codeset.cc 2009-01-06 13:11:54.531250000 -0600
> @@ -720,9 +720,7 @@
> #ifdef __GNUG__
> char _f[flen+1];
> #else
> - CORBA::String_var _f;
> - if (_from->codepoint_size() != 1)
> - _f = CORBA::string_alloc (flen);
> + CORBA::String_var _f = CORBA::string_alloc (flen);
> #endif
>
> if (!f.get (_f, len * _from->codepoint_size())) {
>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Mico-devel mailing list
> Mico-devel@mico.org
> http://www.mico.org/mailman/listinfo/mico-devel
--
Karel Gardas kgar...@objectsecurity.com
ObjectSecurity Ltd. http://www.objectsecurity.com
_______________________________________________
Mico-devel mailing list
Mico-devel@mico.org
http://www.mico.org/mailman/listinfo/mico-devel
