Add support and tests for AssemblyInformationalVersion configurability

[MarkZuber]

No description provided.

[JakeGinnivan]

Thanks for rebasing this @MarkZuber

Just wondering if it would be better to instead call it assembly-informational-override then specify a GitVersion variable. For instance assembly-informational-override: NuGetVersion, but you can also go assembly-informational-override: FullSemVer

Another option is to have assembly-informational-format: {NuGetVersion}, sha: {Sha}, then you can format however you want. I would suggest maybe using https://www.nuget.org/packages/netfx-System.StringFormatWith/ if you wanted to go down this road, if that doesn't do the job then have a look at http://haacked.com/archive/2009/01/14/named-formats-redux.aspx/ for some code to use.

[asbjornu]

I'd prefer a string interpolation based solution with named variables. :)

[MarkZuber]

OK. I've looked at StringFormatWith and it works great except the code
pulled in from the nuget package is older .net and has a few conflicts as
is. It's one file, so i pulled that code in (with appropriate
attributions) into the gitversioncore project.

all tests currently pass and i'm writing some additional tests now to cover
a few additional formatting cases. hope to submit an updated pull request
later today.

thanks for the feedback and input.

-mark

On Mon, Oct 5, 2015 at 12:19 AM, Asbjørn Ulsberg [email protected]
wrote:

I'd prefer a String.Format-based solution with named variables. :)

—
Reply to this email directly or view it on GitHub
#660 (comment).

[MarkZuber]

Quick ping. Does everything look acceptable? Would love to get this into mainline and get a real drop. Please let me know if there are any other changes you'd like me to make. thanks!

[pascalberger]

Does this also fix #368?

[JakeGinnivan]

Why can't we use the VersionVariables class for this?

[MarkZuber]

We need this class in order to generate the VersionVariables class since the types we're using int he formatting need to come from a constructed object. And if we exposed the VersionVariables object as what's valid for formatting, then someone could do {InformationalVersion} and get themselves into an infinite loop. So this gives us type safety of what's allowed to be formatted as well as getting the formatting done to construct the VersionVariables type. I agree it's redundant on most fields (since VersionVariables now just becomes a passthrough/holder for most of the values calculated in the SemanticVersionFormatValues class. Open to suggestions to minimize here if you have them.

[JakeGinnivan]

Looking really good @MarkZuber, just a few comments on there. Thanks!

[MarkZuber]

Ping... I think i've answered/fixed all of the issues requested on the feedback. Would love to get this merged so we can pick up an official drop to use on our build servers. thanks!