>Hi,
>Be aware that the fix that Masa provided breaks the rootcint.
>Jacek.
Thank you for pointing this out. As I sent follow-up message already,
my first fix was not good. There is a choice between
&& (G__func_now>=0 || G__def_struct_member)
and
&& (G__func_now>=0)
but both has small problem. The other choice
&& (G__func_now>=0 || 0==G__def_struct_member)
does not work.
My recommendation is to define G__OLDIMPLEMENTATION1083 macro to
take following. (CINT_parse.c line 470)
&& (G__func_now>=0)
At least this implementation had been working for sometime before.
Thank you
Masaharu Goto