Fix SA1308CodeFixProvider creating empty identifier and add regression test

[jamesqo]

Had to close #2337 because I submitted from the wrong branch. Sorry about that.

SA1308CodeFixProvider has a branch to prevent it from creating an empty identifier if the variable name consists solely of the offending prefix. However, if it consists only of multiple prefixes, e.g. m_s_t_, then it will try to rename the variable to the empty string.

This PR changes the logic to be more flexible, and adds regression tests. It also converts some existing tests for that analyzer to use Theory instead of Fact.

[codecov]

Codecov Report

Merging #2338 into master will increase coverage by 0.02%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #2338      +/-   ##
==========================================
+ Coverage   96.79%   96.81%   +0.02%     
==========================================
  Files         598      732     +134     
  Lines       84340    97117   +12777     
  Branches     3211     3834     +623     
==========================================
+ Hits        81639    94028   +12389     
- Misses       1785     2176     +391     
+ Partials      916      913       -3

Impacted Files	Coverage Δ
...eCop.Analyzers.Test/NamingRules/SA1308UnitTests.cs	100% <100%> (ø)	⬆️
...ers.CodeFixes/NamingRules/SA1308CodeFixProvider.cs	100% <100%> (+5.55%)	⬆️
...nalyzers.Test/Helpers/DiagnosticVerifier.Helper.cs	80.46% <0%> (-9.27%)	⬇️
...leCop.Analyzers.Test/Helpers/MetadataReferences.cs	94.11% <0%> (-5.89%)	⬇️
...nalyzers/SpecialRules/SA0002InvalidSettingsFile.cs	95.65% <0%> (-4.35%)	⬇️
...Analyzers/Settings/ObjectModel/StyleCopSettings.cs	98.07% <0%> (-1.93%)	⬇️
...zers/StyleCop.Analyzers/Settings/SettingsHelper.cs	76.13% <0%> (-1.25%)	⬇️
...zers/Settings/ObjectModel/DocumentationSettings.cs	99.2% <0%> (-0.8%)	⬇️
...ReadabilityRules/SA1115ParameterMustFollowComma.cs	93.22% <0%> (-0.24%)	⬇️
...SA1110OpeningParenthesisMustBeOnDeclarationLine.cs	75.22% <0%> (-0.02%)	⬇️
... and 221 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 841c0ac...b7e8a57. Read the comment docs.

[sharwell]

❓ I noticed from the code coverage report that this continue statement isn't hit by the current tests. Is it possible to reach this if statement and still have numberOfCharsToRemove be 0?

[jamesqo]

@sharwell I think I added this to maintain previous behavior if the name was 1 char, or 2 chars and not one of the prefixes; without this check, we would no longer hit the continue below where we had before. However, if either of those were the case, then clearly the analyzer wouldn't have reported this in the first place. I'll remove this if block.

[sharwell]

Awesome. BTW, I asked because I was using this pull request to demonstrate to someone how code coverage tools can be used to ask more specific questions during code review.