Hi again William, I spent some time this week playing with the HTTP filesystem and ended up with a quite simplified implementation:
bzr branch lp:~aleksander-m/gnupdf/fsys-api Specifically here: http://bazaar.launchpad.net/~aleksander-m/gnupdf/fsys-api/view/head:/src/base/pdf-fsys-http.c It would be great if you could take a look at it and comment it. I already tested the file_open() part, where the HTTP-HEAD is done to know if "Accept-Ranges: bytes" is supported. Also implemented the file_read() part with a range request via HTTP-GET, but didn't have time to test it yet. Note that I modified the utils/pdf-filereader program a bit to easier test all this. Again, not saying that this is the way we should go forward with the HTTP filesystem, just suggesting an easy implementation. Oh, and still, Disk filesystem should be our main priority right now, not the HTTP one :-/. I plan to fully focus on it and merge to that branch I'm working on some patches which didn't get merged yet [1]. Also, I think Jose is preparing a patch for gnulib to include url creation capabilities ported from wget, which we will also use in both the Disk and HTTP filesystems. Cheers! [1] http://lists.gnu.org/archive/html/pdf-devel/2011-04/msg00010.html -- Aleksander
signature.asc
Description: This is a digitally signed message part
