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 >> $i; done
hg --time addremove -s75 --dry-run
before: real 264.560 secs (user 215.130+0.000 sys 49.410+0.000)
after: real 218.710 secs (user 172.790+0.000 sys 45.870+0.000)
from mercurial import parsers
from mercurial.node import nullid, nullrev
import struct
# This unit test compares the return value of the original Python
# implementation of parseindex and the new C implementation for
# an index file with and without inlined data
# original python implementation
def gettype(q):
return int(q & 0xFFFF)
def offset_type(offset, type):
return long(long(offset) << 16 | type)
indexformatng = ">Qiiiiii20s12x"
def py_parseindex(data, inline) :
s = 64
cache = None
index = []
nodemap = {nullid: nullrev}
n = off = 0
# if we're not using lazymap, always read the whole index
l = len(data) - s
append = index.append
if inline:
cache = (0, data)
while off <= l:
e = struct.unpack(indexformatng, data[off:off + s])
nodemap[e[7]] = n
append(e)
n += 1
if e[1] < 0:
break
off += e[1] + s
else:
while off <= l:
e = struct.unpack(indexformatng, data[off:off + s])
nodemap[e[7]] = n
append(e)
n += 1
off += s
e = list(index[0])
type = gettype(e[0])
e[0] = offset_type(0, type)
index[0] = tuple(e)
# add the magic null revision at -1
index.append((0, 0, 0, -1, -1, -1, -1, nullid))
return index, nodemap, cache
data_inlined = '\x00\x01\x00\x01\x00\x00\x00\x00\x00\x00\x01\x8c' \
'\x00\x00\x04\x07\x00\x00\x00\x00\x00\x00\x15\x15\xff\xff\xff' \
'\xff\xff\xff\xff\xff\xebG\x97\xb7\x1fB\x04\xcf\x13V\x81\tw\x1b' \
'w\xdduR\xda\xc6\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00' \
'x\x9c\x9d\x93?O\xc30\x10\xc5\xf7|\x8a\xdb\x9a\xa8m\x06\xd8*\x95' \
'\x81B\xa1\xa2\xa2R\xcb\x86Pd\x9a\x0b5$vd_\x04\xfd\xf6\x9c\xff@' \
'\x11!\x0b\xd9\xec\xf7\xbbw\xe7gG6\xad6\x04\xdaN\xc0\x92\xa0$)' \
'\xb1\x82\xa2\xd1%\x16\xa4\x8b7\xa9\xca\xd4-\xb2Y\x02\xfc\xc9' \
'\xcaS\xf9\xaeX\xed\xb6\xd77Q\x02\x83\xd4\x19\xf5--Y\xea\xe1W' \
'\xab\xed\x10\xceR\x0f_\xdf\xdf\r\xe1,\xf5\xf0\xcb\xf5 \xceR\x0f' \
'_\xdc\x0e\x0e\xc3R\x0f_\xae\x96\x9b!\x9e\xa5\x1e\xbf\xdb,\x06' \
'\xc7q\x9a/\x88\x82\xc3B\xea\xb5\xb4TJ\x93\xb6\x82\x0e\xe16\xe6' \
'KQ\xdb\xaf\xecG\xa3\xd1 \x01\xd3\x0b_^\xe8\xaa\xa0\xae\xad\xd1' \
'&\xbef\x1bz\x08\xb0|\xc9Xz\x06\xf6Z\x91\x90J\xaa\x17\x90\xaa' \
'\xd2\xa6\x11$5C\xcf\xba#\xa0\x03\x02*2\x92-\xfc\xb1\x94\xdf\xe2' \
'\xae\xb8\'m\x8ey0^\x85\xd3\x82\xb4\xf0`:\x9c\x00\x8a\xfd\x01' \
'\xb0\xc6\x86\x8b\xdd\xae\x80\xf3\xa9\x9fd\x16\n\x00R%\x1a\x06' \
'\xe9\xd8b\x98\x1d\xf4\xf3+\x9bf\x01\xd8p\x1b\xf3.\xed\x9f^g\xc3' \
'^\xd9W81T\xdb\xd5\x04sx|\xf2\xeb\xd6`%?x\xed"\x831\xbf\xf3\xdc' \
'b\xeb%gaY\xe1\xad\x9f\xb9f\'1w\xa9\xa5a\x83s\x82J\xb98\xbc4\x8b' \
'\x83\x00\x9f$z\xb8#\xa5\xb1\xdf\x98\xd9\xec\x1b\x89O\xe3Ts\x9a4' \
'\x17m\x8b\xfc\x8f\xa5\x95\x9a\xfc\xfa\xed,\xe5|\xa1\xfe\x15\xb9' \
'\xbc\xb2\x93\x1f\xf2\x95\xff\xdf,\x1a\xc5\xe7\x17*\x93Oz:>\x0e'
data_non_inlined = '\x00\x00\x00\x01\x00\x00\x00\x00\x00\x01D\x19' \
'\x00\x07e\x12\x00\x00\x00\x00\x00\x00\x00\x00\xff\xff\xff\xff' \
'\xff\xff\xff\xff\xd1\xf4\xbb\xb0\xbe\xfc\x13\xbd\x8c\xd3\x9d' \
'\x0f\xcd\xd9;\x8c\x07\x8cJ/\x00\x00\x00\x00\x00\x00\x00\x00\x00' \
'\x00\x00\x00\x00\x00\x00\x01D\x19\x00\x00\x00\x00\x00\xdf\x00' \
'\x00\x01q\x00\x00\x00\x01\x00\x00\x00\x01\x00\x00\x00\x00\xff' \
'\xff\xff\xff\xc1\x12\xb9\x04\x96\xa4Z1t\x91\xdfsJ\x90\xf0\x9bh' \
'\x07l&\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00' \
'\x00\x01D\xf8\x00\x00\x00\x00\x01\x1b\x00\x00\x01\xb8\x00\x00' \
'\x00\x01\x00\x00\x00\x02\x00\x00\x00\x01\xff\xff\xff\xff\x02\n' \
'\x0e\xc6&\xa1\x92\xae6\x0b\x02i\xfe-\xe5\xbao\x05\xd1\xe7\x00' \
'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01F' \
'\x13\x00\x00\x00\x00\x01\xec\x00\x00\x03\x06\x00\x00\x00\x01' \
'\x00\x00\x00\x03\x00\x00\x00\x02\xff\xff\xff\xff\x12\xcb\xeby1' \
'\xb6\r\x98B\xcb\x07\xbd`\x8f\x92\xd9\xc4\x84\xbdK\x00\x00\x00' \
'\x00\x00\x00\x00\x00\x00\x00\x00\x00'
def runtest() :
py_res_1 = py_parseindex(data_inlined, True)
c_res_1 = parsers.parse_index(data_inlined, True)
py_res_2 = py_parseindex(data_non_inlined, False)
c_res_2 = parsers.parse_index(data_non_inlined, False)
if py_res_1 != c_res_1:
print "Parse index result (with inlined data) differs!"
if py_res_2 != c_res_2:
print "Parse index result (no inlined data) differs!"
print "done"
runtest()