Re: crash while testing imread.m

[Søren Hauberg]

tor, 25 02 2010 kl. 15:26 -0500, skrev John W. Eaton:
> On 25-Feb-2010, Ben Abbott wrote:
> 
> | I reverted my graphicsmagick to 1.3.7. "test imread" now passes.
> 
> Yes, but that is just a temporary solution.  We need to figure out
> what to do about later versions of GraphicsMagick which apparently
> require the init function.
> 
> Can we safely call the init function from a dynamically loaded
> function?  If the dynamically loaded function is cleared, should the
> GraphicsMagick cleanup function be called?  If so, do we have a way to
> handle that now?  If not, I guess we need to create some method of
> registering functions to be called when .oct files are cleared.

Could we create simple wrappers for the image I/O functions and put them
in 'liboctave'? The '__magick_read__', '__magick_write__' and
'__magick_imfinfo__' would then simply call these functions. This would
allow us to initialise GraphicsMagick when 'liboctave' is initialised.

Soren