Performance implications of storing issues and comments in commits

[nafey]

Hi I have been following your video series and I think this is a very interesting project.

However, when you say that you want to store all the issues and comments on a separate branch are you not concerned about the performance implications of that? I have a very naive understanding of git but if I, say, double the number of commits in a repo would that not make every operation on the git repository that much slower?

[xeubie]

Most git operations are unaffected by the number of commits. Operations like status and commit only need to operate on the tree pointed to by HEAD. Also this would be an entirely separate branch so you would not see this data when running log for example. Of course, these objects involve extra data when doing an initial clone, but they are zlib-compressed and deduplicated via the packfile format. The entire point is to have that data locally, so this is not an avoidable cost. There are git repos with many hundreds of thousands of commits and they still can perform well.