Mercurial > hgsubversion
view tests/comprehensive/test_verify_and_startrev.py @ 931:e1dbd9646d6a
svnwrap: use custom StringIO class in get_file()
The wrappers were calling ra.get_file() with a cStringIO object.
Empirically, svn 1.7.5 is writing 16kB blocks to the stream object, and
cStringIO reallocates its internal buffer and doubles its size whenever
it is filled. With large committed files this requires two large
memory blocks at the same time.
SimpleStringIO implements the mimimum StringIO interface used by
ra.get_file() but instead stores all the blocks and "join" them at the
end. It means more fragmentation but requires only one large block,
without overallocation. Also, 16kB blocks should be friendly to most
allocators.
In practice, this simple change let me convert a revision containing
multiple moderately large files, the largest being around 450MB, with a
32-bits Windows setup, python 2.7, swig svn 1.7.5, in stupid mode, while
it was previously aborting with "not enough memory". The same revision
still fails in replay mode.
author | Patrick Mezard <patrick@mezard.eu> |
---|---|
date | Sun, 16 Sep 2012 19:31:49 +0200 |
parents | 7f90bb48c9de |
children | f9014e28721b |
line wrap: on
line source
import os import pickle import sys import unittest # wrapped in a try/except because of weirdness in how # run.py works as compared to nose. try: import test_util except ImportError: sys.path.insert(0, os.path.dirname(os.path.dirname(__file__))) import test_util from mercurial import hg from mercurial import ui from hgsubversion import verify # these fixtures contain no files at HEAD and would result in empty clones _skipshallow = set([ 'binaryfiles.svndump', 'binaryfiles-broken.svndump', 'emptyrepo.svndump', 'correct.svndump', 'corrupt.svndump', ]) _skipall = set([ 'project_root_not_repo_root.svndump', ]) _skipstandard = set([ 'subdir_is_file_prefix.svndump', 'correct.svndump', 'corrupt.svndump', ]) def _do_case(self, name, stupid, layout): subdir = test_util.subdir.get(name, '') repo, svnpath = self.load_and_fetch(name, subdir=subdir, stupid=stupid, layout=layout) assert len(self.repo) > 0 for i in repo: ctx = repo[i] self.assertEqual(verify.verify(repo.ui, repo, rev=ctx.node(), stupid=True), 0) self.assertEqual(verify.verify(repo.ui, repo, rev=ctx.node(), stupid=False), 0) # check a startrev clone if layout == 'single' and name not in _skipshallow: self.wc_path += '_shallow' shallowrepo = self.fetch(svnpath, subdir=subdir, stupid=stupid, layout='single', startrev='HEAD') self.assertEqual(len(shallowrepo), 1, "shallow clone should have just one revision, not %d" % len(shallowrepo)) fulltip = repo['tip'] shallowtip = shallowrepo['tip'] repo.ui.pushbuffer() self.assertEqual(0, verify.verify(repo.ui, shallowrepo, rev=shallowtip.node(), stupid=True)) self.assertEqual(0, verify.verify(repo.ui, shallowrepo, rev=shallowtip.node(), stupid=False)) stupidui = ui.ui(repo.ui) stupidui.config('hgsubversion', 'stupid', True) self.assertEqual(verify.verify(stupidui, repo, rev=ctx.node(), stupid=True), 0) self.assertEqual(verify.verify(stupidui, repo, rev=ctx.node(), stupid=False), 0) # viewing diff's of lists of files is easier on the eyes self.assertMultiLineEqual('\n'.join(fulltip), '\n'.join(shallowtip), repo.ui.popbuffer()) for f in fulltip: self.assertMultiLineEqual(fulltip[f].data(), shallowtip[f].data()) def buildmethod(case, name, stupid, layout): m = lambda self: self._do_case(case, stupid, layout) m.__name__ = name bits = case, stupid and 'stupid' or 'real', layout m.__doc__ = 'Test verify on %s with %s replay. (%s)' % bits return m attrs = {'_do_case': _do_case} fixtures = [f for f in os.listdir(test_util.FIXTURES) if f.endswith('.svndump')] for case in fixtures: if case in _skipall: continue bname = 'test_' + case[:-len('.svndump')] if case not in _skipstandard: attrs[bname] = buildmethod(case, bname, False, 'standard') name = bname + '_stupid' attrs[name] = buildmethod(case, name, True, 'standard') name = bname + '_single' attrs[name] = buildmethod(case, name, False, 'single') # Disabled because the "stupid and real are the same" tests # verify this plus even more. # name = bname + '_single_stupid' # attrs[name] = buildmethod(case, name, True, 'single') VerifyTests = type('VerifyTests', (test_util.TestBase,), attrs) def suite(): all_tests = [unittest.TestLoader().loadTestsFromTestCase(VerifyTests)] return unittest.TestSuite(all_tests)