I'll chime in, as I'm the one who initially raised the idea : )
I'll start with a use-case: my initial motivation for having is_deeply
handle coderefs came up while building certain unit tests for a
rewrite of DBD::Mock. Several of the worker functions return complex
data structures -- which may contain coderefs -- and I'd like to
simply hand the whole lot off to is_deeply and have the work done for
me.
My inital strategy for implementing this was a two-tiered approach.
First, compare the references; if they're the same, return true, go no
futher. If they differ, however, say if anonymous subs were thrown
into the mix, then use B::Deparse to decompile both subs, comparing
their textual representations.
My initial quick-glance at B::Deparse's documentation mentions
something about perl optimising certain constants away, which could
well throw a spanner into the works. Storable uses B::Deparse when
serialising coderefs, though, so I'm certain there's a way around
this.
Collin Winter
On 6/25/05, Michael G Schwern <[EMAIL PROTECTED]> wrote:
> Currently, throwing is_deeply() a code ref causes it to barf.
>
> perl -MTest::More -wle 'print is_deeply sub {}, sub {}'
> WHOA! No type in _deep_check
> This should never happen! Please contact the author immediately!
> # Looks like your test died before it could output anything.
>
>
> is_deeply() doesn't know what to do with code refs. What should it do?
>
> What it *shouldn't* do is what Test.pm does, namely execute the code ref and
> compare the values returned. It would just compare the refernces.
>
> This much is obvious, this should pass:
>
> is_deeply \&foo, \&foo;
>
> The same code ref is being compared. But here's the tricky part.
>
> is_deeply sub { 42 }, sub { 42 };
>
> Does that pass? ie. Are we just comparing the reference value or should
> is_deeply() decompile the subroutine and see if they contain the same code?
> (Technical considerations aside, for the moment).
>
> Given that is_deeply() considers these to be equivalent:
>
> is_deeply [ 42 ], [ 42 ];
>
> I think comparing the decompiled source code is consistent, but is it
> useful? Anyone have use cases for this?
>
>
> --
> Michael G Schwern [EMAIL PROTECTED] http://www.pobox.com/~schwern
> Don't try the paranormal until you know what's normal.
> -- "Lords and Ladies" by Terry Prachett
>
