Ruby » ruby-trunk

Question: any chance of getting this [or something like it] included before feature freezing all the way?
Thanks all. Ruby rox :)

thanks for looking into that :)
-=R

Any progress on this? any way I can help?
Thanks!
-=r

Hi,

At Sun, 21 Sep 2008 03:02:44 +0900,
Roger Pack wrote in [ruby-core:18762]:
> I wasn't sure how to patch 
> RubyVM::InstructionSequence.disassemble
> to work with proc objects AND methods, so added a new method
> RubyVM::InstructionSequence.disassemble_proc
> which works for proc objects, in the accompanying patch.
> Either way would work.

It's pretty easy.


Index: iseq.c
===================================================================
--- iseq.c	(revision 23974)
+++ iseq.c	(working copy)
@@ -1007,4 +1007,13 @@ iseq_s_disasm(VALUE klass, VALUE body)
 	}
     }
+    else {
+	rb_proc_t *proc;
+	VALUE iseqval;
+	GetProcPtr(body, proc);
+	iseqval = proc->block.iseq->self;
+	if (RUBY_VM_NORMAL_ISEQ_P(iseqval)) {
+	    ret = rb_iseq_disasm(iseqval);
+	}
+    }

     return ret;

-- 
Nobu Nakada

Another possiblility would be to add an instruction-sequence method (iseq) into Proc. Off of the instruction sequence, one could use the methods like disasm or to_a.

The actual additional code (put in the right place) for this seems pretty small:

VALUE
proc_iseq(VALUE self)
{
    rb_proc_t *proc;
    rb_iseq_t *iseq;
    VALUE rb_iseq;
    GetProcPtr(self, proc);
    iseq = proc->block.iseq;
    if (!iseq) return Qnil;
    rb_iseq = iseq_alloc_shared(rb_cISeq);
    RDATA(rb_iseq)->data = iseq;
    return rb_iseq;
}

 rb_define_method(rb_cProc, "iseq", proc_iseq, 0);

Warning: the above may have flaws in it.

Would it be possible for somebody to commit this? Do we need more approval?
Thanks.
-r

Hi,

In message "Re: [ruby-core:25962] [Feature #578] add method to disassemble Proc objects"
    on Tue, 6 Oct 2009 20:55:57 +0900, Roger Pack <redmine@ruby-lang.org> writes:

|Would it be possible for somebody to commit this? Do we need more approval?

Yea, a patch from Nobu in [ruby-core:24169] needs permission from
Koichi.

							matz.

Roger Pack wrote::
>> Yea, a patch from Nobu in [ruby-core:24169] needs permission from
>> Koichi.
> 
> Ok.  Koichi do you have any feedback on the patch?

no.  go ahead.

-- 
// SASADA Koichi at atdot dot net

The fact that we can get backtraces from the VM easily almost makes me wonder if the following API is possible:

Thread#backtrace_with_bindings
=>
[[file, line, method, [[:arg1_name, current_value], [:arg2_name, current_value]]

or
[[file, line, method, binding]] # so you can lookup argument values if so desired

mostly inspired by http://barelyenough.org/blog/2005/04/ruby-backtraces/ -- not sure if it's actually necessary/high priority since you can get somewhat higher quality backtraces using ruby-debug already (http://github.com/rdp/backtracer)


Thoughts?
-r