You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Greg Stein <gs...@lyra.org> on 2001/04/03 07:01:59 UTC
plugins for hashes? (was: Re: Linux Kernel Summit)
On Tue, Apr 03, 2001 at 05:12:01AM +0000, Tripp Lilley wrote:
> On Tue, 3 Apr 2001, Gavin Thomas Nicol wrote:
>
> > Does anyone see any advantage in using crc64 over
> > tiger/md5? I't be 1/2 the size, and a bit faster,
> > but I'm not sure if there are other advantages.
>
> Being one to advocate plugins, I advocate plugins here :)
More complexity than it is worth. We're already debating their utility. The
marginal utility of a plugin is even less :-)
> While on the topic of hashing, what specifically are the envisioned
> domains of the hashes? I would particularly like the ability to retrieve
> the hash of the head revision of the file, because I can foresee that
> being a very useful tool in scripting. I expect that other hashes would be
> useful (hashes of the diff "list"?), but the head revision comes in handy
> when you're reconciling an offline workspace with the server...
Hmm. I think we'd have the plain text hash, and a second hash for the diff
data. Of course, old revisions don't store the plain text, but when we
restore it, we'll already have its hash (i.e. we computed it when the
revision first arrived; no big reason to toss it).
Cheers,
-g
--
Greg Stein, http://www.lyra.org/
Re: plugins for hashes? (was: Re: Linux Kernel Summit)
Posted by Tripp Lilley <tl...@perspex.com>.
On Tue, 3 Apr 2001, Greg Stein wrote:
> On Tue, Apr 03, 2001 at 05:12:01AM +0000, Tripp Lilley wrote:
> > Being one to advocate plugins, I advocate plugins here :)
>
> More complexity than it is worth. We're already debating their utility. The
> marginal utility of a plugin is even less :-)
Understood. I also re-realized that when I saw the previous post go by
reminding me that you were talking about using the built-in facilities in
Apache to do the MD5 :) So put it on the 3.0+ list and I'll be quiet :)
> Hmm. I think we'd have the plain text hash, and a second hash for the diff
> data. Of course, old revisions don't store the plain text, but when we
> restore it, we'll already have its hash (i.e. we computed it when the
> revision first arrived; no big reason to toss it).
Both sound reasonable and should offer sufficient functionality for the
90% case. Cool. Keeping the already-computed earlier hashes around should
also be helpful for the test harness, I'd think...
--
Joy-Loving * Tripp Lilley * http://stargate.eheart.sg505.net/~tlilley/
------------------------------------------------------------------------------
"Fiber makes you poop." -- From <http://www.pvponline.com/bts_studio.php3>