Add the ability to load IFP animations without compression

[FileEX]

Summary

This PR adds an optional uncompressedAnimations argument to the engineLoadIFP function, allowing you to specify which animations should be loaded without compression.

bool engineLoadIFP(string ifp_file / string raw_data, string  custom_block_name [, table uncompressedAnimations = nil ] )

Motivation

Currently, all IFP animations are loaded as compressed by default, regardless of whether compression is necessary. While testing PR #4555, @MegadreamsBE discovered several issues when using cutscene animations, including broken facial bones (fixed in #4565) and offset problems. It was determined that the issue occurs because all animations are loaded as compressed by default, whereas GTA loads some cutscene animations in a special way as uncompressed (the uncompress section in .cut files).

Test plan

@MegadreamsBE tested this.

Thanks to @Merlin for helping with the crash investigation.

Checklist

• Your code should follow the coding guidelines.
• Smaller pull requests are easier to review. If your pull request is beefy, your pull request should be reviewable commit-by-commit.

[qaisjp]

I haven't review the code but

@MegadreamsBE tested this.

Could you go into more detail into what the actual tests were? E.g. sample resources that they tested against

[FileEX]

I haven't review the code but

@MegadreamsBE tested this.

Could you go into more detail into what the actual tests were? E.g. sample resources that they tested against

This branch master...FileEX:mtasa-blue:megadreams/object-animations_with_clumps was used for testing, because currently MTA doesn’t include object animations or the ability to replace clump-type models (#4555 and #4052).

Test resource:
anims2.zip

I personally don’t know cutscene animations very well, but MegaDreams is familiar with them, and the test resource comes from him.

[Lpsd]

It seems like the ReadKr* methods are practically doing the same thing, couldn't an interface help cut down code-duplication here?

[FileEX]

That's true, it could be done better, but in my opinion, that's not the topic of this PR. This PR simply adds a new argument to the function, rather than refactoring the existing code.

[MegadreamsBE]

I haven't review the code but

@MegadreamsBE tested this.

Could you go into more detail into what the actual tests were? E.g. sample resources that they tested against

Two-fold:
I run a resource to test a specific IFP which contains uncompressed animations. The fix made it work correctly, before the coordinates were wrong. FileEX linked it in a previous comment.

I ran multiple cutscene IFPs, which rely on uncompressed animations with this new code through my Sphene project. No crashes nor issues were observed.

I can vow for the stability of this change.