feat: Support TMX relative path for Source texture

[DeathPhoenix22]

Hi,

Issue Link: #1302

This change ensures that we can use the path specified on the tmx map file instead of removing the source file path at resource load time.

[AlmasB]

Hi, could you add a test for this. Can we copy an existing texture into another test directory, so that the tmx references a relative path, that should make it fail (i.e. return a dummy texture)?

[codecov-commenter]

Codecov Report

Merging #1303 (4a6d32c) into dev (76cdf21) will decrease coverage by 0.04%.
Report is 6 commits behind head on dev.
The diff coverage is 100.00%.

❗ Current head 4a6d32c differs from pull request most recent head f8853b4. Consider uploading reports for the commit f8853b4 to get more accurate results

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

@@             Coverage Diff              @@
##                dev    #1303      +/-   ##
============================================
- Coverage     59.60%   59.57%   -0.04%     
+ Complexity     2962     2956       -6     
============================================
  Files           308      308              
  Lines         15199    15199              
  Branches       1558     1558              
============================================
- Hits           9060     9055       -5     
- Misses         5610     5620      +10     
+ Partials        529      524       -5     

Files	Coverage Δ
...om/almasb/fxgl/entity/level/tiled/TilesetLoader.kt	77.72% <100.00%> (ø)

... and 4 files with indirect coverage changes

📣 Codecov offers a browser extension for seamless coverage viewing on GitHub. Try it in Chrome or Firefox today!

[DeathPhoenix22]

Hi,
I've added tests. It took me several hours to figure out a way to find a difference between a Dummy image and an Image. I finally decided to add-up all pixels. This is due to the fact that the dummy Image is a valid image with all correct properties :)

I've added an extra test that validates the checksum strategy.

Hopefully, this suites the testing needs.

[AlmasB]

Looks great, many thanks.

[AlmasB]

I am not sure how big argb values are, so perhaps Long is more suitable. Though I suppose identical values will overflow in the same way, so Int should work too.