view tests/test-empty-group @ 11609:890ad9d6a169

log: slowpath: do not read the full changelog

When in the slowpath, we are examining _all_ changesets in revs.
We need to order reads so they happen increasingly for I/O performance.

Increasing windows were used to read changelog backwards in a windowed manner,
reading the changelog forward inside each window. But since no revision range
was specified, it was equivalent to reading the full changelog, even if a
single revision was passed to the commandline.

When --removed is used, we _need_ to scan all changesets, but if we're only
looking for file patterns, this is not necessary and we can stick to
the revspec that was given to us.

#!/bin/sh
#
#  A          B
#
#  3  4       3
#  |\/|       |\
#  |/\|       | \
#  1  2       1  2
#  \ /        \ /
#   0          0
#
# if the result of the merge of 1 and 2
# is the same in 3 and 4, no new manifest
# will be created and the manifest group
# will be empty during the pull
#
# (plus we test a failure where outgoing
# wrongly reported the number of csets)
#

hg init a
cd a
touch init
hg ci -A -m 0 -d "1000000 0"
touch x y
hg ci -A -m 1 -d "1000000 0"
hg update 0
touch x y
hg ci -A -m 2 -d "1000000 0"
hg merge 1
hg ci -A -m m1 -d "1000000 0"
#hg log
#hg debugindex .hg/store/00manifest.i
hg update -C 1
hg merge 2
hg ci -A -m m2 -d "1000000 0"
#hg log
#hg debugindex .hg/store/00manifest.i

cd ..
hg clone -r 3 a b
hg clone -r 4 a c
hg -R a outgoing b
hg -R a outgoing c
hg -R b outgoing c
hg -R c outgoing b

hg -R b pull a
hg -R c pull a