view tests/test-merge6 @ 5339:058e93c3d07d

I have spotted the biggest bottleneck in "bdiff.c". Actually it was
pretty easy to find after I recompiled the python interpreter and
mercurial for profiling.

In "bdiff.c" function "equatelines" allocates the minimum hash table
size, which can lead to tons of collisions. I introduced an
"overcommit" factor of 16, this is, I allocate 16 times more memory
than the minimum value. Overcommiting 128 times does not improve the
performance over the 16-times case.

#!/bin/sh

cat <<EOF > merge
import sys, os
print "merging for", os.path.basename(sys.argv[1])
EOF
HGMERGE="python ../merge"; export HGMERGE

mkdir A1
cd A1
hg init
echo This is file foo1 > foo
echo This is file bar1 > bar
hg add foo bar
hg commit -m "commit text" -d "1000000 0"

cd ..
hg clone A1 B1

cd A1
rm bar
hg remove bar
hg commit -m "commit test" -d "1000000 0"

cd ../B1
echo This is file foo22 > foo
hg commit -m "commit test" -d "1000000 0"

cd ..
hg clone A1 A2
hg clone B1 B2

cd A1
hg pull ../B1
hg merge
hg commit -m "commit test" -d "1000000 0"
echo bar should remain deleted.
hg manifest --debug

cd ../B2
hg pull ../A2
hg merge
hg commit -m "commit test" -d "1000000 0"
echo bar should remain deleted.
hg manifest --debug