Text formatting does not specify culture

[tom-englert]

String formatting in the analyzer does not specify a culture, therefore some tests fail on systems with non-English culture with e.g. "expected: 2.0, actual: 2,0"

I have added a workaround in ClousotMain.Analyze() by setting the main threads culture to invariant. However this is very bad practice, and should be fix in long term.

[SergeyTeplyakov]

Yep, totally agree. It should be automated way to switch to culture agnostic formatting. I think R# could do this.

[tom-englert]

Currently it's hard to use R# because the formatting of the code is not yet cleaned up.
Once we have the code properly formatted, we could add a R#.sln.DotSettings file so everyone is using the same style.

[SergeyTeplyakov]

I meant that even in this case it is possible to run only one fix across all projects. I'll try to do this.

[sharwell]

Once we have the code properly formatted, we could add a R#.sln.DotSettings file so everyone is using the same style.

I'm not using ReSharper. I'm planning to add a proposal that we use StyleCopAnalyzers for style checking. It has no dependencies other than Roslyn, is free for everyone, is extremely fast, and we've implemented the ability to fix most violations with 3 clicks or less.

As a bonus, a few members of the team working on StyleCopAnalyzers have actually been using CodeContracts.sln for profiling passes, leaving me quite confident that it will be a wonderful solution for the needs of this project.

[tom-englert]

I don't mind using StyleCopAnalyzers, I think it does not contradict with making the live of R# users easier 😄
Hopefully we get the StyleCop-cleaned code soon.