Content Pipeline Assets and Source Control (Git)

Currently I only have the original asset files added to Git without the compiled .xnb files, the latter ones get built into the output folder which is, as expected, ignored by Git.

But is this the intended way that the pipeline tool should be used?
Should both be added? Perhaps only the .xnb files be added?

Adding only the .xnb and having the original files in a separate repository seems somewhat logical to me but also cumbersome.

If you add xnbs to git: each time you will add an asset and build it, it will be indexed by git if you dont tell to handle it as binary content.
Each time you rebuild the assets because for ex, by changing a content model processor, it will also be marked as changed and you will loose readability of the changes between two commit. You can filter the view but it’s not a view ‘within an eye blink’
It will make your index grow with unnecessary data if not configured correctly and it will still make it grow
Git can handle many files but the purpose is to share code not all kind of data.

It’s only my point of view :wink:

The content files (.mgcb) work very well with Git. If that’s in your repository, others working on the project should be able to simply pull/clone and build it to get the XNBs. As such, you’d be able to see changes to your output files by just looking at the content file, so you likely won’t need the XNBs in source control.

But is this the intended way that the pipeline tool should be used?
Should both be added? Perhaps only the .xnb files be added?

yes
no
no