>Stephen,
>I agree somehow with your point. I have experienced the same
>frustation. The point is that (by definition) I cannot check
>in TTree::Draw because the function is not even called by CINT
>in this case. What happens typically (intuitive) is that you do not type
>the } (as suggested by the CINT error info and as indicated
>by Masa's reply), but rather type CRTL/C. At this point, CINT gets
>confused and looses the control.
In a sense I agree. My instruction is not too kind in this context.
I can improve command prompt behavior.
>Masa,
>This is a quite serious point in my mind as it is becoming one of
>the most serious complaints we have from many users.
>We must discuss the relationship between the CRTL/C event
>and the CINT recovery/cleanup procedure in this case.
As I tried, cint completely keeps its' control in this senario. The only
thing you have to do is type '}' which is directly shown to the user.
CRTL/C event behaves differently in UNIX and Win32 and both cases it is
in control of ROOT interrupt handler. I do not think this is a problem of
CTRL/C behavior.
Masaharu Goto