Fwd: I'd like to reopen discussion on the Windows path problem

Lee Cantey lcantey at gmail.com
Wed Apr 25 16:14:28 CDT 2007

Wrong reply list.

---------- Forwarded message ----------
From: Lee Cantey <lcantey at gmail.com>
Date: Apr 25, 2007 2:12 PM
Subject: Re: I'd like to reopen discussion on the Windows path problem
To: TK Soh <teekaysoh at gmail.com>

I think it's the right fix, or at least a workable one.  I use it when
building the installer since it's too painful not to.  I agree that we
should take it since it's been out there awhile and if something
better comes along or we need to tweak it then that's fine too.

BUT, I've been having problems since the switch to demandimport.  I
finally managed to get py2exe to actually include hgext and the .pyd
extensions but the .pyd files can't be loaded at runtime.tro

I'd been planning on making a post about the experience but figure now
is as good as time as any to bring it up. :P


On 4/25/07, TK Soh <teekaysoh at gmail.com> wrote:
> On 1/2/07, Lee Cantey <lcantey at gmail.com> wrote:
> > On 1/2/07, Matt Mackall <mpm at selenic.com> wrote:
> > >
> > > What happens when you explicitly put c:\hg in PYTHON_PATH because you
> > > want to run a copy out of the repo?
> > >
> >
> > You should still get that copy since it will be inserted into sys.path
> > after the directory that the script is running from.  You'll still
> > have the problem locating the supporting dynamic libraries unless you
> > add them also.
> Just wondering what the latest status on this issue. I still having
> running the latest hg inside the crew repo.

More information about the Mercurial mailing list