git becomes too slow after many commits and going back in history several times. After a decent amount of modules installed I seem to have to run git repair' and git gc' a lot, and they take a lot of time.
One destructive way to resolve it is go back, in history, to your initial commit throw away (or move) your .git directory and start all over. Not a proper solution, though.
Probably git just stores too much and the loose objects can be cleaned up. I'm not sure how, though.
One way to limit the number of objects is also to commit less. One could instead of every file, only track files that are changed. that would make this handling more error-prone, though, as now the files that are added, but not tracked need to be listed somewhere. Abusing .gitignore for this seems not the right way.
git becomes too slow after many commits and going back in history several times. After a decent amount of modules installed I seem to have to run
git repair' andgit gc' a lot, and they take a lot of time.One destructive way to resolve it is go back, in history, to your initial commit throw away (or move) your .git directory and start all over. Not a proper solution, though.
Probably git just stores too much and the loose objects can be cleaned up. I'm not sure how, though.
One way to limit the number of objects is also to commit less. One could instead of every file, only track files that are changed. that would make this handling more error-prone, though, as now the files that are added, but not tracked need to be listed somewhere. Abusing .gitignore for this seems not the right way.