We've tried out commitstream with our gitlab repo recently and what stood out the most to me is:
- it relies completely on developers putting an assetID in the commit message.
- tight coupling that it creates between git and versionOne
Because this is a manual process there will be mistakes - mistyped assetID, forgotten assetID etc. and no way to fix those mistakes. Not to mention the extra effort and distraction to find the correct assetID while the developer wants to just make a commit and move on.
There is also the issue of VersionOne recognizing the assetID correctly. What if a commit comment includes something that looks like an assetID, but isn't!

Why doesn't VersionOne let users associate commits to assets right from VersionOne and maintain those mappings within VersionOne?

This way,
- developers don't need to be aware of assetIDs at the commit time
- git repos are not full of assetIDs that don't really belong in git commits. Some might disagree, but nothing stop them from putting the assetIDs in the commit comment if they want to.
- Because the commit to assetID mapping is done in VersionOne, the mistakes can always be corrected.