Re: duplicate body remover

[Li Guanglei <guanglei at cn dot ibm dot com>]

Frank Ch. Eigler wrote:
> Li Guanglei <guanglei@cn.ibm.com> writes:
>
> > [...]  today when I tried to shrink the size of the generate .c file
> > from stap,
>
> The biggest current source of repetition is the entry- and
> registration-related functions, not the final probe handlers or
> auxiliary functions.  I'm close to checking in a major improvement on
> this front.
looking forward to your new codes :)

> > [...]
> > +    if(typeid(*statements[i])==typeid(null_statement))
>
> (We try to avoid run-time typing checks like that - it indicates poor
> OO design.)
>
> While your code does shrink pass-2 output, as David said, it is almost
> certainly unnecessary given the pass-3-level duplicate elimination for
> probe handlers.

but that's only partial optimization. If probe_* are duplicated, 
shouldn't we optimize further to eliminate duplicated enter_probe_* 
and register_* too? Seems that your new coming codes will optimize for 
this, right?

> - FChE