[DvBree]

Description

Replaces float->int conversion by truncation to floor/ceil operations. Truncation is incorrect for negative values: both 0.5 and -0.5 truncate to 0, which is undesirable.

For backwards compatibility with t_rex, we use floor for x-coordinates and ceil for y-coordinates. This was tested by comparing tiles to tiles generated by t_rex.

Type of change

• Bugfix

Checklist:

• I've double-checked the code in this PR myself
• I've left the code better than before (boy scout rule)
• The code is readable, comments are added that explain hard or non-obvious parts.
• I've expanded/improved the (unit) tests, when applicable
• I've run (unit) tests that prove my solution works
• There's no sensitive information like credentials in my PR