omg, ain't GD so very difficult. I'm looking at Image-Square 0.01_4 testers matrix, what was supposed to be walk in the park is like blood covered battlefield.
I know my knowledge of images is lacking but I was beginning to feel I had done something terribly wrong...
No CoventryCathedral for tests below, simply a red 8 by 8 square to reduce to smaller squares
Isn't the whole point of the tests to check that the module does what it is supposed to in real situations?
Users of the module (me if I am the only one) will be using ti to process large, if not huge, images. If it passes the tests on little tiny images but fails on large ones, doesn't that sort of render the tests meaningless?
My original tests were based on those in Image::Resize in the < href="https://metacpan.org/release/SHERZODR/Image-Resize-0.5/source/t/1.t"1.t</code> file which just checks the dimensions of the generated file. But that module doesn't crop images which is why I wanted to include tests of the actual output.
It appears that copyResampled (and interpolation in general, see further) is unstable between versions and plagued with bugs
I did originally use copy but changed to copyResampled when I decided it would be sensible to add the facility to change the size at the same time...
In reply to Re^4: Testing image output
by Bod
in thread Testing image output
by Bod
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |