[NetKAN] Tolerate Malformed Author URLs from KS/SD

[dbent]

From KSP-CKAN/NetKAN#2980. This PR changes the datatype of KerbalStuff's website and source_code properties into String from Uri, since we have no guarantee that they're actually well-formed Uri strings. Now when "normalizing" the URIs (renamed the method from Escape to Normalize since we were doing a lot more in it than merely escaping characters) we do a final check to make sure the string is a well formed URI. If it is, we proceed as normal, if it isn't we spit out a warning and return null, preventing it from being added to the metadata (the SafeAdd extension method will ignore null inputs).

@politas @plague006 @pjf

[mheguy]

Trying to run the debug output leads to a problem with windows not having curl, and I don't have access to a linux machine to properly build a netkan.exe. If anyone would link one, I'd happily start using/testing it.

KerbalStuff is/was having issues and that was causing netkan to try to use curlsharp (causing my issues).

Here's a debug build if anyone else wants to run it for a couple days and see if any issues emerge: http://www.mediafire.com/download/8q7m9d4fs5x73a4/Debug.zip

Tagging @Olympic1 as a person likely to test this.

[Olympic1]

Tried out the debug build from @plague006 with KSP-CKAN/NetKAN#2980 and netkan.exe succesfully created a .ckan file without problems. Looking at the debug screen it checks the link and removes the broken link.

Will test further for a couple of days with existing and new netkans.

[mheguy]

This PR seems moot with recent events, but I have to imagine the changes can, will, and should be applied to spacedock.

[dbent]

Yes, I'll have to update it for Spacedock.

[dbent]

Updated for SpaceDock

[Olympic1]

Weird, it only added the Kerbalstuff ones, merging manually then.