It seems to me that doing a bzr log <file> plus bzr diff -c <rev> <file> for all revisions is quite a bit faster than bzr log -v <file>, even though it produces a superset of the information. Is that what this bug is about?
Eg. (11551 is the only rev for this file, where it was added):
; time sh -c 'bzr log fileinbigrepo && bzr diff -c 11551 fileinbigrepo' > /dev/null
real 0m4.035s
user 0m3.826s
sys 0m0.203s
; time bzr log -v fileinbigrepo > /dev/null
real 0m19.138s
user 0m18.892s
sys 0m0.221s
It seems to me that doing a bzr log <file> plus bzr diff -c <rev> <file> for all revisions is quite a bit faster than bzr log -v <file>, even though it produces a superset of the information. Is that what this bug is about?
Eg. (11551 is the only rev for this file, where it was added):
; time sh -c 'bzr log fileinbigrepo && bzr diff -c 11551 fileinbigrepo' > /dev/null
real 0m4.035s
user 0m3.826s
sys 0m0.203s
; time bzr log -v fileinbigrepo > /dev/null
real 0m19.138s
user 0m18.892s
sys 0m0.221s