Change: Improve number parsing

[leezer3]

This branch attempts to unify & improve the number parsing code.

Most places parsing a Vector simply called the NumberFormats.TryParseDoubleVb6 method repeatedly on each component of the vector, and had their own Interface.AddMessage call with an appropriate error message.

This PR creates several new helper functions in the NumberFormats file which perform the function in a unified way. It also adds an unconditional number parser, which adds an appropriate error,

Benefits:

• Single set of methods for parsing vectors.
• Gets rid of a bunch of duplicated code.

Drawbacks / Issues

• This unfortunately changes the structure of some error messages. Most users don't worry about these, but may confuse developers slightly? New: Error X is invalid in AddVertex - Coordinates at line 2 in B:\test\FCMB 2000\3D\Cabina\AgullaPressioB.csv Old: Error Invalid argument vX in AddVertex at line 2 in file B:\test\FCMB 2000\3D\Cabina\AgullaPressioB.csv
• Should these be moved to be vector constructors?
• Slightly nasty hack to enable the messages to be added by the API.
• More places can use this, e.g. TrainEditor2
• Some bits in the panel parsers could use Vector2 more often.

TEST

• Are there any places where an array length check is now required?
• Potentially dicey code in the Panel.xml parser- Some angles don't appear to be converted to radians. Double check what's going on here....

@s520 any thoughts? Ended up running this up as I looked in the Panel2 parser in preparation for adding the windscreen properties & decided I could get rid of some of the duplicated code.

[s520]

I'm busy this week, so I'd like to see it next week.

[leezer3]

No trouble.

There are a lot more places and similar bits that can be added (color parser for one, used in the objects, panels etc), & I'd consider this low priority cleanup work at best.

[leezer3]

This PR is now about 50% of the way towards replacing the parse calls.

A lot of the rest of these use the negative response to set some sort of default value. (where the standard double / int default of 0.0 doesn't apply) Need to think about the correct way to do this....