[SIP-devel] [I'm considering another backend library for SIP]

[Zhang C. <ch...@ma...>]

Hi, all

I'm sorry that I'm inactive for these days, because I'm quite busy with
my homework. I spent some time on SIP's source code yesterday in order
to fix the bug that Ricardo reported. Let me restate it:

The calling to mogrify() from Scilab will cause Scilab to crash. I
checked out mogrify_int.c, but found nothing unusual. So I suspect that
this is a bug for Window ImageMagick, but I don't have time to prove it.

A workaound currently is to rewrite mogrify() in Scilab script. If I
have time, I'll give a try to it. But I come up with an idea that
ImageMagick maybe not the best backend library for SIP. ImageMagick is
huge in size, 25mega bytes, which makes SIP cumbersome when installed on
Windows. What's more, I think what SIP really nead is a lightweight
image I/O library which can handle with the variety of image formats.
The image manipulation should be mainly implemented by SIP instead of
strong dependent on ImageMagick. Thus we can change SIP's functions as
we need, and the basic I/O image library can be packaged with SIP which
makes SIP's distribution easier. However, I know switching means much
coding and testing.

I've found such a library, Free Image Library, free, open source, simple
to use and deliver. I plan to test its ability as SIP's backend. I want
to hear your comments and opinions on the issue.

I hope SIP will become the official image processing toolbox for Scilab
in the future. I'll keep working for it.

Zhang Cheng