From: "Shot (Piotr Szotkowski)" <shot@...>
Date: 2008-02-22T02:40:48+09:00
Subject: Re: Object#freeze as a basis for caching of method results?

--B5t8AZ5AOUxuGP3g
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Shot (Piotr Szotkowski):

> Robert Klemme:

>> Now, whether you use the String or demarshal probably mainly depends
>> on memory usage. If the String is short enough that approach is
>> certainly preferable because it incurs less processing overhead
>> (demarshaling).

> Ok, I=E2=80=99m totally lost here. I don=E2=80=99t see any demarshalling =
happening,
> just marshalling (and using that as a key)=E2=80=A6 What am I missing?

Ok, I got it now (stupid me) =E2=80=93 you discuss keing the
cache with Marshal.dump(self) as opposed to keying it
with Marshal.load(Marshal.dump(self)), right?

Are marshalled objects in general much
bigger (memory-wise) than the =E2=80=98originals=E2=80=99?

Also, I usually define dup as Marshal.load(Marshal.dump(self))
=E2=80=93 provided it=E2=80=99s not a bottleneck, does this approach has an=
y vices?

Lastly, assuming I have a working #dup (Marshal-based or otherwise),
wouldn=E2=80=99t dup.freeze be a better/more elegant key than dump+load? Do=
es
it has any vices?

-- Shot
--=20
Also, I think there is a typo in the opening. Where you say
"Ruby is slow", I think what you mean is "Ruby is an extremely
powerful and flexible language". It's an understandable mistake,
all of the keys are so close to each other.     -- GD, ruby-talk

--B5t8AZ5AOUxuGP3g
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)

iD8DBQFHvbFBi/mCfdEo8UoRAtd0AJ9HlSv6adQPUi3J9npT/PSjRMjAxACfU0fj
nSq+H7g5f6NFzbERvEsRKpA=
=LjzJ
-----END PGP SIGNATURE-----

--B5t8AZ5AOUxuGP3g--