-
Committer:
John Arbash Meinel
-
Date:
2008-08-29 17:38:47 UTC
-
mto:
This revision was merged to the branch mainline in
revision
3679.
-
Revision ID:
john@arbash-meinel.com-20080829173847-n3h7gwv6hep9glbn
If we read more than 50% of the whole index,
go ahead and buffer the whole thing on the next request.
This could be tuned (30%?, 75%?), but the old code could easily
get to the point where we would end up reading more than
1x the total bytes of the file.