Today at work I needed to clone a certain project with git-svn due to certain constraints and the command I used to do that is git svn clone -T --no-minimize-url URL_OF_THE_REPO
. Long story short... this has been PAINFULLY SLOW, and also keeps failing reliably because of some error in git... which naturally makes me continue the cloning by restarting the command, I can't just leave it and know it will complete the clone job when I come back. So making it a night job is out of the question.
1 I think it's because svn was not designed to work the way git works (with objects that are fetched that are compressed and sent in a single block). git-svn has to go revision by revision to replicate the content of the svn branches/tags into git equivalents.
2 and 3 are related: You can try asking for the fetch process to start from a revision that you know has already been retrieved so that git-svn doesn't have to start from scratch each time. you can do something like git svn fetch -r 10000:HEAD
.