From: Frank W. <fwi...@gm...> - 2009-04-13 13:41:49
|
On Mon, Apr 13, 2009 at 3:01 AM, Mayne, Peter <pet...@hp...> wrote: > Looking at mercurial/util.py, it contains (leaving out lots of stuff): > > if os.name == 'nt': > def samestat(s1, s2): > return False > else: > samestat = os.path.samestat > > The last line is where the problem is: Jython has no os.path.samestat. I suppose either there needs to be a check for os.name=='java' as well to sort out the Jython-specific things, instead of assuming only 'nt' and not 'nt'; or Jython could acquire os.path.samestat (and whatever else needs to work on that side of the equation). > > I've just found http://fwierzbicki.blogspot.com/2009/02/progress-with-mercurial-on-jython.html, so the Mercurial on Jython thing seems to be in good hands. :-) Jython and mercurial are indeed collaborating, sorry I am late replying to this thread... The biggest issue is that Jython lacks buffer() and a bzip2 implementation. buffer() will hopefully get some love in the coming 2.6 series (we may even port it back to 2.5.1) but buffer() did not make the cut for 2.5.0. The samestat bug report is very useful though, we should indeed either check on mercurial's side, or see how hard it would be to grow a samestat in Jython (easier than it used to be...) -Frank |