##// END OF EJS Templates
revlog: change generaldelta delta parent heuristic...
revlog: change generaldelta delta parent heuristic The old generaldelta heuristic was "if p1 (or p2) was closer than the last full text, use it, otherwise use prev". This was problematic when a repo contained multiple branches that were very different. If commits to branch A were pushed, and the last full text was branch B, it would generate a fulltext. Then if branch B was pushed, it would generate another fulltext. The problem is that the last fulltext (and delta'ing against `prev` in general) has no correlation with the contents of the incoming revision, and therefore will always have degenerate cases. According to the blame, that algorithm was chosen to minimize the chain length. Since there is already code that protects against that (the delta-vs-fulltext code), and since it has been improved since the original generaldelta algorithm went in (2011), I believe the chain length criteria will still be preserved. The new algorithm always diffs against p1 (or p2 if it's closer), unless the resulting delta will fail the delta-vs-fulltext check, in which case we delta against prev. Some before and after stats on manifest.d size. internal large repo old heuristic - 2.0 GB new heuristic - 1.2 GB mozilla-central old heuristic - 242 MB new heuristic - 261 MB The regression in mozilla central is due to the new heuristic choosing p2r as the delta when it's closer to the tip. Switching the algorithm to always prefer p1r brings the size back down (242 MB). This is result of the way in which mozilla does merges and pushes, and the result could easily swing the other direction in other repos (depending on if they merge X into Y or Y into X), but will never be as degenerate as before. I future patch will address the regression by introducing an optional, even more aggressive delta heuristic which will knock the mozilla manifest size down dramatically.

File last commit:

r18642:a40d608e default
r26117:4dc5b51f default
Show More
lsprof.py
109 lines | 3.6 KiB | text/x-python | PythonLexer
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 import sys
Joel Rosdahl
Remove unused imports
r6212 from _lsprof import Profiler, profiler_entry
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422
__all__ = ['profile', 'Stats']
def profile(f, *args, **kwds):
"""XXX docstring"""
p = Profiler()
Dirkjan Ochtman
updating lsprof.py from remote repository
r5992 p.enable(subcalls=True, builtins=True)
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 try:
Dirkjan Ochtman
updating lsprof.py from remote repository
r5992 f(*args, **kwds)
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 finally:
p.disable()
Dirkjan Ochtman
updating lsprof.py from remote repository
r5992 return Stats(p.getstats())
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422
class Stats(object):
"""XXX docstring"""
def __init__(self, data):
self.data = data
def sort(self, crit="inlinetime"):
"""XXX docstring"""
if crit not in profiler_entry.__dict__:
Peter Ruibal
use Exception(args)-style raising consistently (py3k compatibility)
r7008 raise ValueError("Can't sort by %s" % crit)
Alejandro Santos
compat: use 'key' argument instead of 'cmp' when sorting a list
r9032 self.data.sort(key=lambda x: getattr(x, crit), reverse=True)
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 for e in self.data:
if e.calls:
Alejandro Santos
compat: use 'key' argument instead of 'cmp' when sorting a list
r9032 e.calls.sort(key=lambda x: getattr(x, crit), reverse=True)
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422
def pprint(self, top=None, file=None, limit=None, climit=None):
"""XXX docstring"""
if file is None:
file = sys.stdout
d = self.data
if top is not None:
d = d[:top]
Dirkjan Ochtman
updating lsprof.py from remote repository
r5992 cols = "% 12s %12s %11.4f %11.4f %s\n"
hcols = "% 12s %12s %12s %12s %s\n"
Bryan O'Sullivan
lsprof: report units correctly
r16804 file.write(hcols % ("CallCount", "Recursive", "Total(s)",
"Inline(s)", "module:lineno(function)"))
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 count = 0
for e in d:
Dirkjan Ochtman
updating lsprof.py from remote repository
r5992 file.write(cols % (e.callcount, e.reccallcount, e.totaltime,
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 e.inlinetime, label(e.code)))
count += 1
if limit is not None and count == limit:
return
ccount = 0
Matt Mackall
profile: add undocumented config options for profiler output
r16263 if climit and e.calls:
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 for se in e.calls:
Mads Kiilerich
profiling: replace '+' markup of nested lines with indentation...
r18642 file.write(cols % (se.callcount, se.reccallcount,
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 se.totaltime, se.inlinetime,
Mads Kiilerich
profiling: replace '+' markup of nested lines with indentation...
r18642 " %s" % label(se.code)))
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 count += 1
ccount += 1
if limit is not None and count == limit:
return
if climit is not None and ccount == climit:
break
def freeze(self):
"""Replace all references to code objects with string
descriptions; this makes it possible to pickle the instance."""
# this code is probably rather ickier than it needs to be!
for i in range(len(self.data)):
e = self.data[i]
if not isinstance(e.code, str):
self.data[i] = type(e)((label(e.code),) + e[1:])
Dirkjan Ochtman
updating lsprof.py from remote repository
r5992 if e.calls:
for j in range(len(e.calls)):
se = e.calls[j]
if not isinstance(se.code, str):
e.calls[j] = type(se)((label(se.code),) + se[1:])
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422
_fn2mod = {}
def label(code):
if isinstance(code, str):
return code
try:
mname = _fn2mod[code.co_filename]
except KeyError:
Dirkjan Ochtman
lsprof: make profile not die when imported modules changes (issue1774)
r9314 for k, v in list(sys.modules.iteritems()):
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 if v is None:
continue
Augie Fackler
lsprof: use getattr instead of hasattr
r14959 if not isinstance(getattr(v, '__file__', None), str):
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422 continue
if v.__file__.startswith(code.co_filename):
mname = _fn2mod[code.co_filename] = k
break
else:
Benoit Boissinot
fix spaces/identation issues
r10339 mname = _fn2mod[code.co_filename] = '<%s>' % code.co_filename
Vadim Gelfer
add --lsprof option. 3x faster than --profile, more useful output....
r2422
return '%s:%d(%s)' % (mname, code.co_firstlineno, code.co_name)
if __name__ == '__main__':
import os
sys.argv = sys.argv[1:]
if not sys.argv:
print >> sys.stderr, "usage: lsprof.py <script> <arguments...>"
sys.exit(2)
sys.path.insert(0, os.path.abspath(os.path.dirname(sys.argv[0])))
stats = profile(execfile, sys.argv[0], globals(), locals())
stats.sort()
stats.pprint()