Repository information should be cached locally, at least per operation

What steps will reproduce the problem?
1. check version log of a file

What is the expected output? What do you see instead?
It works, but it should be faster.

Basically, one of my files got corrupted, so I wanted to check the boar version 
log to see what I had in the repository. Retrieving said log took pretty long, 
so I checked the server logs (see attachment). Basically, the same files are 
checked time and again, but especially the bloblists can be rather large, so 
downloading them more than once per operation doesn't seem to make much sense.

My server is just a machine at my home, a netbook with bitvise ssh server and 
Win7. I'm on a 10MBit/s upstream 50MBit/s downstream connection, but the ssh 
encryption bottlenecks this down to something around 300 kbyte/s in either 
direction. As you can see in the logs, there's a 9MByte bloblist that was 
downloaded 4 times in those transfers alone (and there were more before that), 
each time taking about 15 seconds. That really adds up and slows down 
operations in a way that isn't necessary.

Original issue reported on code.google.com by jannis.u...@stoppe.de on 14 Feb 2013 at 2:00

Attachments:

BoarMultiAccess.png

amy4li / boar

Repository information should be cached locally, at least per operation #95