[dev] Re: [cvs] commit: imp/lib Tree.php imp folders.php

[Michael M Slusarz]

Quoting Chuck Hagenbuch <chuck at horde.org>:

| Quoting Michael M Slusarz <slusarz at bigworm.colorado.edu>:
|
| >   Most importatnly, we just go ahead and build the entire tree when
| > initializing
| >   instead of marking as discovered/not discovered.  Since we have to do
| >   'lookaheads' anyway (to determine if a particular element has children or
| >   not) we might as well store this information in the tree at the beginning
| >   since, if using a discovery algorithim, we would have to do these
| >   identical queries later anyway.
|
| Right... but the discovery code was what made IMP even *useable* with setups
| like CMU has, where there are thousands and thousands of IMAP folders
| available. Loading every bit of info from the server, when you really just
| need the top two levels, is *much* quicker.

Yeah - while I was writing my commit message, I was having the same thoughts
actually.  So I went ahead and recoded it, but this time it should be more
efficient about what it stores in the cache (see most recent commit message).

michael

______________________________________________
Michael Slusarz [slusarz at bigworm.colorado.edu]
The University of Colorado at Boulder