tests/test-symlink-os-yes-fs-no.py

memctx: fix memctx manifest file hashes

When memctx is asked for a manifest, it constructs one by merging the p1
manifest, and the changes that are on top. For the changes on top, it was
previously using p1.node() as the file entries parent, which actually returns
the commit node that the p1 linkrev points at! Which is entirely incorrect.

The fix is to use p1.filenode() instead, which returns the parent file node as
desired.

I don't know how to execute this or make it have a visible effect, so I'm not
sure how to test it.  It was noticed because asking for the linkrev is an
expensive operation when using the remotefilelog extension and this was causing
performance regressions with commit.

import os, sys, time
from mercurial import hg, ui, commands, util

TESTDIR = os.environ["TESTDIR"]
BUNDLEPATH = os.path.join(TESTDIR, 'bundles', 'test-no-symlinks.hg')

# only makes sense to test on os which supports symlinks
if not getattr(os, "symlink", False):
    sys.exit(80) # SKIPPED_STATUS defined in run-tests.py

u = ui.ui()
# hide outer repo
hg.peer(u, {}, '.', create=True)

# clone with symlink support
hg.clone(u, {}, BUNDLEPATH, 'test0')

repo = hg.repository(u, 'test0')

# wait a bit, or the status call wont update the dirstate
time.sleep(1)
commands.status(u, repo)

# now disable symlink support -- this is what os.symlink would do on a
# non-symlink file system
def symlink_failure(src, dst):
    raise OSError(1, "Operation not permitted")
os.symlink = symlink_failure

# dereference links as if a Samba server has exported this to a
# Windows client
for f in 'test0/a.lnk', 'test0/d/b.lnk':
    os.unlink(f)
    fp = open(f, 'wb')
    fp.write(util.readfile(f[:-4]))
    fp.close()

# reload repository
u = ui.ui()
repo = hg.repository(u, 'test0')
commands.status(u, repo)

# try cloning a repo which contains symlinks
u = ui.ui()
hg.clone(u, {}, BUNDLEPATH, 'test1')