findrenames: Optimise "addremove -s100" by matching files by their SHA1 hashes.
We speed up 'findrenames' for the usecase when a user specifies they
want a similarity of 100% by matching files by their exact SHA1 hash
value. This reduces the number of comparisons required to find exact
matches from O(n^2) to O(n).
While it would be nice if we could just use mercurial's pre-calculated
SHA1 hash for existing files, this hash includes the file's ancestor
information making it unsuitable for our purposes. Instead, we calculate
the hash of old content from scratch.
The following benchmarks were taken on the current head of crew:
addremove 100% similarity:
rm -rf *; hg up -C; mv tests tests.new
hg --time addremove -s100 --dry-run
before: real 176.350 secs (user 128.890+0.000 sys 47.430+0.000)
after: real 2.130 secs (user 1.890+0.000 sys 0.240+0.000)
addremove 75% similarity:
rm -rf *; hg up -C; mv tests tests.new; \
for i in tests.new/*; do echo x >> $i; done
hg --time addremove -s75 --dry-run
before: real 264.560 secs (user 215.130+0.000 sys 49.410+0.000)
after: real 218.710 secs (user 172.790+0.000 sys 45.870+0.000)
#!/bin/sh
"$TESTDIR/hghave" svn svn-bindings no-outer-repo || exit 80
fixpath()
{
tr '\\' /
}
svnupanddisplay()
{
(
cd $1;
svn up;
svn st -v | fixpath | sed 's/ */ /g'
limit=''
if [ $2 -gt 0 ]; then
limit="--limit=$2"
fi
svn log --xml -v $limit | fixpath | sed 's,<date>.*,<date/>,' | grep -v 'kind="'
)
}
echo "[extensions]" >> $HGRCPATH
echo "convert = " >> $HGRCPATH
hg init a
echo a > a/a
mkdir -p a/d1/d2
echo b > a/d1/d2/b
echo % add
hg --cwd a ci -d '0 0' -A -m 'add a file'
"$TESTDIR/svn-safe-append.py" a a/a
echo % modify
hg --cwd a ci -d '1 0' -m 'modify a file'
hg --cwd a tip -q
hg convert -d svn a
svnupanddisplay a-hg-wc 2
ls a a-hg-wc
cmp a/a a-hg-wc/a && echo same || echo different
hg --cwd a mv a b
echo % rename
hg --cwd a ci -d '2 0' -m 'rename a file'
hg --cwd a tip -q
hg convert -d svn a
svnupanddisplay a-hg-wc 1
ls a a-hg-wc
hg --cwd a cp b c
echo % copy
hg --cwd a ci -d '3 0' -m 'copy a file'
hg --cwd a tip -q
hg convert -d svn a
svnupanddisplay a-hg-wc 1
ls a a-hg-wc
hg --cwd a rm b
echo % remove
hg --cwd a ci -d '4 0' -m 'remove a file'
hg --cwd a tip -q
hg convert -d svn a
svnupanddisplay a-hg-wc 1
ls a a-hg-wc
chmod +x a/c
echo % executable
hg --cwd a ci -d '5 0' -m 'make a file executable'
hg --cwd a tip -q
hg convert -d svn a
svnupanddisplay a-hg-wc 1
test -x a-hg-wc/c && echo executable || echo not executable
echo % executable in new directory
rm -rf a a-hg a-hg-wc
hg init a
mkdir a/d1
echo a > a/d1/a
chmod +x a/d1/a
hg --cwd a ci -d '0 0' -A -m 'add executable file in new directory'
hg convert -d svn a
svnupanddisplay a-hg-wc 1
test -x a-hg-wc/d1/a && echo executable || echo not executable
echo % copy to new directory
mkdir a/d2
hg --cwd a cp d1/a d2/a
hg --cwd a ci -d '1 0' -A -m 'copy file to new directory'
hg convert -d svn a
svnupanddisplay a-hg-wc 1
echo % branchy history
hg init b
echo base > b/b
hg --cwd b ci -d '0 0' -Ambase
"$TESTDIR/svn-safe-append.py" left-1 b/b
echo left-1 > b/left-1
hg --cwd b ci -d '1 0' -Amleft-1
"$TESTDIR/svn-safe-append.py" left-2 b/b
echo left-2 > b/left-2
hg --cwd b ci -d '2 0' -Amleft-2
hg --cwd b up 0
"$TESTDIR/svn-safe-append.py" right-1 b/b
echo right-1 > b/right-1
hg --cwd b ci -d '3 0' -Amright-1
"$TESTDIR/svn-safe-append.py" right-2 b/b
echo right-2 > b/right-2
hg --cwd b ci -d '4 0' -Amright-2
hg --cwd b up -C 2
hg --cwd b merge
hg --cwd b revert -r 2 b
hg resolve -m b
hg --cwd b ci -d '5 0' -m 'merge'
hg convert -d svn b
echo % expect 4 changes
svnupanddisplay b-hg-wc 0