This is the mail archive of the
mailing list for the Cygwin project.
Re: [RFC] an alternative to rebasing [recall]
- From: Steven O'Brien <steven dot obrien2 at ntlworld dot com>
- To: cygwin at cygwin dot com
- Date: Tue, 6 Aug 2002 20:22:03 +0100
- Subject: Re: [RFC] an alternative to rebasing [recall]
> On Tue, Aug 06, 2002 at 09:38:52AM +0100, Steven O'Brien wrote:
> > So rebase is the only solution available right now.
> Are you giving up? Or, will you try to find another solution?
I have no new ideas to investigate at present. The problem is that the
cygwin algorithm in dll_init.cc is able to force windows to use a higher
relocation address, but cannot force it to relocate lower than the one
it finds by default. I was hoping that by always loading from low
address to high then we could side-step this problem, and indeed in many
cases this does work. But sometimes windows relocates to a higher
address than it did in the parent, and we are just not able to correct
If anyone knows where I can find a specification of the windows
relocation algorithm, then maybe I can find a mechanism to implement
true fork() semantics, or prove once and for all that it just cannot be
done. Until then, or some other flash of inspiration, I will be working
on some of the other (many) outstanding issues with my gnome port.
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html