Refactor the UID class

[qmfrederik]

In the end, a UID object is an reference to another object, by id. The value of UID is always positive (because it is an index). The maximum allowed value is ulong.MaxValue.

UIDs have no (meaningful) names; they are just numbers.

With this in mind:

• UID object can store their value as an ulong instead of a byte[] array, preventing the allocation of an array
• We can deprecate constructors which take signed values (such asint instead of uint)
• We can remove the name property alltogether.

This helps, you have guessed it, performance, as it's easier to track a ulong than it is to track a byte[]:

• 27% performance increase on Windows
• 34% performance increase on Linux

[qmfrederik]

Windows

Before

Method	Mean	Error	StdDev	Gen 0	Gen 1	Allocated
ReadLargePropertylistTest	2.543 ms	0.0336 ms	0.0315 ms	175.7813	85.9375	1.06 MB

After

Method	Mean	Error	StdDev	Gen 0	Gen 1	Allocated
ReadLargePropertylistTest	1.844 ms	0.0235 ms	0.0220 ms	117.1875	54.6875	720.21 KB

Linux

Before

Method	Mean	Error	StdDev	Gen 0	Gen 1	Allocated
ReadLargePropertylistTest	3.240 ms	0.1291 ms	0.3766 ms	171.8750	85.9375	1.06 MB

After

Method	Mean	Error	StdDev	Median	Gen 0	Gen 1	Allocated
ReadLargePropertylistTest	2.128 ms	0.0416 ms	0.0671 ms	2.096 ms	109.3750	54.6875	720.21 KB