[cfe-dev] new design of RecursiveASTVisitor

[Jordy Rose]

On Sat, 5 Jun 2010 23:07:40 -0700, Zhanyong Wan (λx.x x) <wan at google.com>
wrote:
> On Sat, Jun 5, 2010 at 4:20 PM, Jordy Rose <jediknil at belkadan.com>
wrote:
>>
>> Having just written a small tree-walker myself, I want to ask what
>> returning "false" means. There's lots of ways you might want to "stop"
in
>> the middle of a traversal:
>>
>> 1. Stop traversing altogether. This is what the original proposal says,
>> if
>> I'm reading it correctly.
> 
> Yes, that's what I meant.  Think of it as throwing an exception.

<snip>

>> Do you (or does anyone) think there are enough uses of #3, #4, or #5 to
>> warrant implementing something like this? A nice way to do it could be
a
>> simple push()/pop() system -- when you say abort, it pop()s back to the
>> last saved node. Alternately, you could return a special value that
says
>> how far back to pop() (such as a particular Stmt*), with special
>> constants
>> for "Continue" and "Abort".
> 
> I suspect we are getting a bit ahead of ourselves here.  I like to
> make the design as general as possible too - as long as we don't
> introduce unnecessary complexity.  I think the use cases you described
> are rare, and I'm fine with not supporting them directly.  The user
> can always override enough methods to make the AST visitor do whatever
> he wants to do.  Or he may just create a new AST visitor class if that
> makes more sense.

Heh, I suppose so. I started out writing with the intent to make sure
"stop" was well-defined (and eventually, clearly documented), but then got
carried away with all the possible uses. Of course it's easy to modify
behavior by subclasing RecursiveASTVisitor.

Thanks,
Jordy