[Object Analyzer] Fix display of Mtx holder

[Dragorn421]

This fixes displaying Mtx holders in the object analyzer

(didn't test with several matrices, only 1, I assume it should work fine though)

example with gLinkChildDekuShieldMtx

Mtx gLinkChildDekuShieldMtx = { 
    0x0000FFFF , 0          , 0xFFFFFFFF , 0xFFFF0000 ,
    0          , 0xFFFF0000 , 0x0258FF9C , 0x00320001 ,
    0x21B4023D , 0          , 0xF136FE0A , 0x00710000 ,
    0xFD5421A5 , 0xF1140000 , 0          , 0          ,
}; 

Sources:
Matrix_MtxToMtxF in oot decomp
gSPMatrix in the sdk

[Thar0]

Printing Mtx as floats is a nice change, I have a couple of suggestions on the implementation.

[Thar0]

Printing the floats only up to the precision necessary for the fixed point range should make things nicer, i.e. Matrix_CheckFloats elects to print the floats as 12.6f in printf notation, although I think it should really be 13.6 as 12.6 fails to account for negative floats with 5-digit integer part (such large numbers likely will never come up in practice however). I'm not sure what the equivalent is in C# interpolated strings.

[Dragorn421]

In decimal,
the smallest positive s16.16 number is 1/0x10000 = 0.0000152587890625
the smallest s16.16 number is 0x80000000/0x10000 = -32768
and the largest is 0x7FFFFFFF/0x10000 = 32767.9999847412109375

so that's 5.16f maybe? and I don't know how it works in C# either. Microsoft probably thought it was funny to make up their own format strings

[Dragorn421]

Rereading this,

the smallest positive s16.16 number is 1/0x10000 = 0.0000152587890625

absolutely doesn't mean we need 16 digits, but only 5 (for the decimal part)

idk still about this, it's not like there isn't enough space in Z64Utils to show whatever useless digits