Imlib 1.9.0 question

[Andreas Tille <tille physik uni-halle de>]

Hello,

if this is not the right place to ask Imlib question please tell
me where to ask.  I didn't found any other place.

1. Tracing the Imlib 1.9.0 sources I detected a change in loading/saving
   image files.  Up to 1.8.2 a hhard function was called.  Now there
   is a pointer to a function which is called to load the image.

   I suppose that the sense behind it is to geve the user the chance
   to define its own loading/saving routines to read/write special
   chunks in PNG or tags in TIFF, etc.

   But there isn't any documentation how to tell Imlib about the
   user defined function.  Could someone explain me how it works?

   It is from great interest for me, because I recently use a modified
   Imlib loading routine to read special chunks/tags and would like
   to go a clean way.

2. At present time it is not possible to use Imlib without X or
   GDK initialized.  I would like to have a chance to use the Imlib
   functionality off line (the way convert of ImageMagick does).
   That's the second reason why I did my own loading routine.
   I would like the way that loading/saving and some image
   manipulation functions could work without initialized display.
   This could cause only problems with XPMs but this could be easily
   avoided.

   The initialisation of the display has to be done before rendering
   the images as pixmaps of course. 

3. Since I used Imlib 1.8.2 I've got SIGSEGVs ofter manipulating the
   rgb_data of an imlib structure.  For instance I used my own
   routine to mirror images.  Switching from 1.8.1 to 1.8.2 lead to
   SIGSEGVs in this routine.  Using Imlibs own routine helped to avoid
   the SIGSEGVs but I have to do certain image manipulation on the RGB
   data which can't be done with Imlib functions.

   Was there any change in the internal Imlib structure???

Kind regards

       Andreas.