[Zope] Re: Upload just file headers?

[Martijn Pieters]

On Wed, Oct 11, 2000 at 03:54:17PM -0400, Paul Winkler wrote:
> For example, take a .gif or .jpeg file and throw away all
> but the first (let's say) 1 kb of the file.
> If you try to open this truncated file with an image editor
> (gimp, photoshop, whatever) you'll have problems. But useful
> information is still stored in that first part of the file,
> as we can see using python and PIL.
> Example, on linux:
> 
> head --bytes 1000 foo.gif > foo_truncated.gif
> python
> ...
> >>> import Image
> >>> test = Image.open('foo_truncated.gif', 'r')
> >>> test.size
> (210, 210)
> 
> So in this case, we can find out that foo.gif is 210 x 210
> pixels even if we only have access to the first 1 kb of the
> 12 kb file.

This doesn't hold true for JPEG's. 

A JPEG encoded image consists of a set of segments, each encoding some
aspect of the image. The image dimensions can be found in the frame header
segment, but this segment may be preceded by tables, comments and
application specific data.

These extra segments could easily take up your 1K chunk. AP press for
example, uses an application specific segment to encode image metadata
such as subject, caption and photographer.

If you have a look at the Image dimension detection code for JPEGs you'll
see that it scans for a frame header (marked by C0, C1, C2 or C3 depending
on the encoding used). Once found, it extracts the width and height. More
information of the JPEG file encoding format can be found at:

  http://www.w3.org/Graphics/JPEG/itu-t81.pdf

For a GIF file, the first 10 bytes should suffice, for a PNG you need the
first 24.

-- 
Martijn Pieters
| Software Engineer  mailto:mj@digicool.com
| Digital Creations  http://www.digicool.com/
| Creators of Zope   http://www.zope.org/
---------------------------------------------