>root [0] a=3
>Warning: Undeclared data member a FILE:/tmp/aaaaapEma LINE:1
>(int)3
>*** Interpreter error recovered ***
>
>seems to be ok, (int)3 seems to indicate that the interpreter have still
>taken a for an int by default. But if one tries to see what is this var
>
>root [1] int a=3
>Error: a already declared as different type FILE:/tmp/baaaapEma LINE:1
>*** Interpreter error recovered ***
>
>So, is a defined or not ?
This was a bug. I fixed this in cint5.13.77.
a should have been declared as an automatic int variable.
Now, please be careful that CINT handles automatic variable in the way
different from C/C++ standard. CINT distinguishes following 2 declarations.
a=3;
int a=3;
So, at the second line,cint still complains a has been already declared.
>There is perhaps a problem in some cases when trying to use time.h
>As suggested by Masa a few months ago, I ran mkincld.c in cint and
>transferred the resulting header files in $ROOTSYS/cint/include.
>This cleaned up some problems but there remains one.
>After doing a #include <time.h> in the interpreter, the following macro
>doesn't work :
>
>{
> time_t tt, gts1;
> double sampleRate;
> tt = time(NULL);
> gts1 = mktime(gmtime(&tt));
> sampleRate = 2.e+4;
>}
>
>namely, it says :
>
>root [31] .x FrTest.C
>Error: sampleRate already declared as different type
>FILE:/big_data/virgo/users/buskulic/virgo_dev/vega/v0.1/test/./FrTest.C
>LINE:6
>*** Interpreter error recovered ***
I have difficulty debugging this problem because it only happens in ROOT
session. CINT alone works fine. Sorry but I can not make progress now.
Masaharu Goto