This is the mail archive of the
mailing list for the Cygwin project.
Re: rebase problem for cygcurl-2.dll still existing?!
- From: Nicholas Wourms <nwourms at netscape dot net>
- To: Robert Collins <robert dot collins at syncretize dot net>
- Cc: 'Jason Tishler' <jason at tishler dot net>, cygwin-apps at cygwin dot com
- Date: Wed, 17 Jul 2002 08:27:07 -0400
- Subject: Re: rebase problem for cygcurl-2.dll still existing?!
- References: <000701c22d8b$c7642b90$0200a8c0@lifelesswks>
Robert Collins wrote:
It goes against the way fork works.
[mailto:email@example.com] On Behalf Of Nicholas Wourms
Sent: Wednesday, 17 July 2002 10:09 PM
To: Jason Tishler
Subject: Re: rebase problem for cygcurl-2.dll still existing?!
Jason Tishler wrote:
Is that a deficiency of cygwin as a whole, or just related
to the way
Cygwin's fork() attempts to load DLLs in the child in the
my DLL was built?)
Other than the lack of someone writing the code, is there any
fork() can't automagically try another location? Is this
Or does it go against the way Cygwin/Windows works?
as in the parent. If it fails, then the child aborts.
Take a pointer foo with value bar. After fork, *foo must equal *foo
before the fork. If bar points into a dll, and the dll is mapped
somewhere else, then after fork *foo != *foo before the fork.
Thank you, Rob. That clears things up for me. :-)