[TYPO3-dev] ImageMagick - you'll never get it right!

[Ernesto Baschny [cron IT]]

Elias Humbolt schrieb am 21.04.2006 17:36:

>> As an example, take a look of what Gallery 1
>> (http://gallery.menalto.com) did and how the new Gallery 2 totally
>> reworked the image processing, including full "out-of-the-box" support
>> for every IM/GM version there is, where no user-setting is needed.

> Yes, that's what I am up to! If you examine the code you will be stunned at 
> how many places somebody tried to deal with the IM issue. A million places 
> where no such code should be! This needs to be cleared out. And this can be 
> cleared out even without risking backwards compatibility. It can even be 
> automated without risking backwards compatibility!
>
> This is no problem, because the source of the problem is so damn simple. It is 
> just the way the T3 team dealt with it which is unnecessarily complicated.
> 
> I don't blame anybody here, because at the point when the problem occurred, I 
> imagine everything just started as a quick fix and it was especially 
> difficult to figure out what caused the problem in the first place.

I couldn't agree more.

> It's just, after so many versions it is time to get rid of this first try and 
> get something better in place.

Let's hope you're right! :) I can already see people using extensions
that directly access those weird "im_version", "im_invert_mask",
"im5_but_dont_invert" settings directly to generate some images.

> I am willing to do the work necessary. I would even go so far to test my code 
> with all extensions that use any Imagemagick functions to assure backwards 
> compatibility.

Well, there we go! That would simply be amazing, and I think lots of
people that (like me) find this issue very interesting and worth some
major changes would appreciate your efford and be helpful in reviewing
and/or testing. We just need to have some initial CODE (patch) so that
we can see what can be done. And I'm pretty sure no core developer would
have anything against this experiment and if some results can be seen,
why not have that included in maybe even in 4.1?

Cheers,
Ernesto