Re: cvs: php4 /ext/standard image.c

On Wed, 16 May 2001, Thies C. Arntzen wrote: >On Tue, May 15, 2001 at 11:55:54PM -0700, Rasmus Lerdorf wrote: >> > are you sure this is a wise thing to do? this code has been >> > more-or-less taken from the ijg-jpeg and has been in there as >> > long as this (getimagesize) function exists. could you send >> > me a failing sample-image so i could see myself? >> >> It made my images work. Before this change none of the images that came >> from my digital camera worked. And when I originally wrote/cloned the >> code they certainly worked. > > the problem is that now we start parsing the jpeg-file with a > 3-byte offset and not from the beginning. this was changed in > order to allow GetImageSize work on remote images (no fseek() > possible). > > i'm just not sure if this change(s alltogeter) won't break on > other valid jpeg-images. > > the clean approach would be to buffer the 3 bytes that we > need to identify the image as a jpeg and feed them back into > the next_marker functions somehow. From jdmarker.c: <---clip---> /* Skip any non-FF bytes. * This may look a bit inefficient, but it will not occur in a valid file. * We sync after each discarded byte so that a suspending data source * can discard the byte from its buffer. */ while (c != 0xFF) { cinfo->marker->discarded_bytes++; INPUT_SYNC(cinfo); INPUT_BYTE(cinfo, c, return FALSE); } <---clip--->