Re: [tree-ssa] Removal of gotos from cfg based ir

[Andrew Haley <aph at redhat dot com>]

Andrew MacLeod writes:
 > On Fri, 2003-11-14 at 12:29, Andrew Haley wrote:
 > > Andrew MacLeod writes:
 > >  > On Fri, 2003-11-14 at 11:58, Andrew Haley wrote:
 > >  > > Andrew MacLeod writes:
 > >  > >  > 
 > >  > >  > 
 > >  > >  > How does java determine where an expression might trap to?
 > >  > > 
 > >  > > "Exceptions are precise: when the transfer of control takes place, all
 > >  > > effects of the statements executed and expressions evaluated before
 > >  > > the point from which the exception is thrown must appear to have taken
 > >  > > place. No expressions, statements, or parts thereof that occur after
 > >  > > the point from which the exception is thrown may appear to have been
 > >  > > evaluated. If optimized code has speculatively executed some of the
 > >  > > expressions or statements which follow the point at which the
 > >  > > exception occurs, such code must be prepared to hide this speculative
 > >  > > execution from the user-visible state of the program."
 > >  > > 
 > >  > How does it determine *where* the exception is goig to go?
 > >  > 
 > >  > ie, if there are 4 stmts in a row than might trap, how do you know where
 > >  > the exception might go?
 > > 
 > > They will trap to the handler for the region they're in.
 > > 
 > >  > Are they all contained inside a TRY block?
 > > 
 > > Yes.
 > > 
 > >  > so to make a stmt trap to a different label, you need to be in a
 > >  > different TRY block?
 > > 
 > > Or to have a different exception.  You can have handlers for divide
 > > overflow and null pointer traps, for example, and these can have
 > > different handlers.
 > > 
 > DIfferent handlers, but they are all within the same TRY region, and
 > they simply have different CATCHs, like in c++?

yeah, like

try
{
  a = b/c + foo();
}
catch (DivideOverflow e)
{
   ...
}
catch (NullPointerException e2)
{
  ...
}

 > 
 > So within a TRY region, an exception on any of these stmts is
 > transferred to the same Catch handling region/landing pad/ whatever? All
 > NULL traps will go to the same label, all overflows go to the same label
 > etc?

Yes.

 > SO we can characterize all the stmts within a region the same way?
 > 
 > and can you resume back to where you were, ie the next stmt, or is
 > control premanently in the handler now?

You may not resume, no.

Andrew.