Re: [pdf-devel] Problems in the stream implementation

[Juan Pedro Bolivar Puente]

Hi,

Great news, I'm looking forward to start trying to find bugs on your
redesign :) I have been all day working on the fixing of the stream
implementation too, and also required some redesign.

I used the "ring buffer" approach wich does not require any extra error
codes and (IMHO) eases the filter implementation.

I have found many bugs spread around the old implementation in
pdf-stm-filter.c, pdf-stm.c, pdf-stm-f-*.c and pdf-filter.c and tried to
solve most of them.

I have written a cfile backend that can be used to read/write to a C
FILE*. I have rewritten many parts of pdf-filter.c to use this new
backend. Now, both readding and writting can be used by toggling the
--readmode flag and the cache size can be adjusted with a --cache parameter.

I am about to write a bash script to do some automatic tests on the
filters using pdf-filter tool.

I hope my patch is finished and on the mailing list in half an hour. I
don't mind if it does not get into trunk, but anyways some parts may be
of interest. In any case, I have been doing a lot of testing (write and
read operations, variable cache sizes, etc...) and it is almost bug-free
(still have to check seek functions).

Regards,
JP