On Thu, May 20, 2004 at 02:29:18PM -0700, John-Marc Chandonia wrote:
> 	Why would anybody want to do this?  In my lab, we keep lab
> notebooks in datestamped directories (e.g., /040520/) with symlinks to
> the next and previous weeks (/040513/n/ is the same directory as
> /040520/ or 040527/p/).

Ok, so I'm clear:

swish is indexing /040513/ and comes across "n" (/040513/n) and you have a
FileRules setting to skip that.  But swish marks that /040513/n was
visited, which is actually a symlink for /040520/ -- and it's really
that directory that is marked as seen.

So, later when swish tries to index /040520/ it sees that it was already
indexed and skips it.

Is that correct?

Interesting.

Have to look at that a while -- I wonder if all the filerules checks
should not be before flagging a directory as seen.  I might have put that
check for seen first to avoid checking FileRules on already seen dirs.
But, that may be a bad design choice.

Do have to worry about loops.

-- 
Bill Moseley
moseley@hank.org