Re: plpgsql arrays

> Uh, no, it wouldn't.  Visually:
>
>        L1      -------------------------
>        L2      -----------
>        L3      ---------------------
>
>        R1                     --------
>
> At L2, you'd conclude that you're done matching R1.
>

No, you should conclude that you're done matching L2. You conclude
you're done matching R1 when you reach L4  ( or there exists a j st
Lj.start > R1.end, or equivalently Lj is strictly greater than R1 )

FWIW, this is a very common problem in bioinformatics. I've mostly
implemented this in python and C. The code is available at
encodestatistics.org. Look in encode.py at the overlap method of the
feature_region class, or ( for the C version ) region_overlap in
block_bootstrap.c ( svn is more up to date for the C ).

-Nathan