Re: [pdf-devel] DCT implemenet

[jemarch]

    > I'm writing DCT filter.I've some issues to discuss with you.
    > 
    > 1. For the base layer, should the input and output  be file or memory? Or
    > both?
    > 
    > 2. How to get the parameters for encode, e.g. the image size? If in the 
hash
    > table filter_params, should I add the key name of the parameters in
    > pdf-stm.h?
    > 

   Hi Yank, for both questions I recommend you to read some already implemented
   filter like, pdf-f-ahex. I'll give you some hints anyways.
   The input and output are two buffers, whether on disk or memory should not 
be of
   your concern.
   You can read more here:
   http://gnupdf.org/Lib:Architecture/Base_Layer/Stream_Module#Buffers

   Regarding the parameters you only ask for the key, as it's supposed to be 
already
   added prior the filter's call.

You can see an example on the use of pdf_hash_t to hold the filter's
parameters in the JBIG2 decoder (src/base/pdf-stm-f-jbig2.[ch]).