On 05/01/06, Dharmaweb.org dharmaweb@gmail.com wrote:
What I don't understand is where do you specified the path to the actual file in the image directory?
I don't know if it *is* stored anywhere, for performance reasons or whatever, but it's worth pointing out that the divisions in the upload directory are not arbitrary, and rely exclusively on the filenames. IIRC, they're the first and second characters of an MD5 hash of the actual name. So it's perfectly possible to not store them anywhere, since they can be recalculated whenever you know the name of the file you're looking for.
The purpose, I think, is to avoid servers containing large numbers of uploads having one directory with a huge number of images in it. (Indeed, the images can even be transparently split across multiple servers this way, by mounting the sub-directories over the network.)
There was talk of using file *content* hashes and all sorts a while back, but I'm not sure anything got implemented. -- Rowan Collins BSc [IMSoP]