On Tue, Jan 18, 2000 at 10:15:31AM -0600, Jeff Hayward wrote:
> On Fri, 14 Jan 2000, Russell Nelson wrote:
>
> One way to do that would be for Dan to change the
> Maildir specification so that a Maildir may have multiple "cur"
> directories. Then, keep a CDB containing a subset of the message
> headers.
>
> Why multiple "cur" directories? I'm guessing that you're trying to avoid
> rebuilding a large CDB when any cachable item changes. Why not simply use
> multiple CDB's in a single directory instead? Select a CDB by hashing the
> file names.
CDB is hashed itself. Using multiple CDB's to share one load is useless.
The multiple "cur" directory idea helps performance on average filesystems.
> I'm also presuming that the CDB will be indexed by something like the
> message file name. How efficient are things like string searches going to
> be in that case? My dream states include things like results of previous
> searches being cached (I have several large folders that I search on the
> same subset of strings frequently). How would you do that with a CDB?
Well the CDB (in my idea, at least) will be indexed to the unchanging part
of a message filename (without new/ or cur/ in front), and contain the headers
that mutt normally reads from the file itself while opening. [Yes, I am
targeting mutt specifically, don't flame me ;)]
For searches thru headers, the cdb can be used. For body-text-searches my
solution won't help much.
Greetz, Peter.
--
Peter van Dijk - student/sysadmin/ircoper/madly in love/pretending coder
|
| 'C makes it easy to shoot yourself in the foot;
| C++ makes it harder, but when you do it blows your whole leg off.'
| Bjarne Stroustrup, Inventor of C++