Debugger vs -E

Hello.  I solicit your thoughts on how this combination of runtime flags 
might be made to cooperate:

	-dE

the goal being that perl code evaled in the debugger might include such 
modern gems as 'say', etc.

The processing of the -E flag currently takes place in toke.c after the 
processing of the -d flag, but hoisting it up is not sufficient, because 
perl5db.pl is invoked via 'require' and the "use feature ':5.10'" that is 
generated by -E has a lexical scope.  My ideas of how to solve this so far 
are unpalatable (hardcode a version of 'require' that doesn't have the 
lexical hiding, modify perl5db.pl to accept some sort of argument, do 
horrible things with $evalarg, etc).  Can anyone see a nice solution?

-- 
Peter Scott
http://www.perlmedic.com/
http://www.perldebugged.com/
http://www.informit.com/store/product.aspx?isbn=0137001274

• Debugger vs -E by Peter Scott
• Re: Debugger vs -E by Ben Morrow
• Re: Debugger vs -E by Rafael Garcia-Suarez
• Re: Debugger vs -E by H.Merijn Brand