Mail Archives: cygwin/2010/12/17/11:11:26
Am 17.12.2010 15:56, schrieb Rance Hall:
> Over the last several days I've been reading the thread that
> eventually migrated to a discussion of combining rebase and perlrebase
> as one.
>
> I've largely ignored it because I have never had a problem with cygwin
> that wasn't my fault.
>
> It seems more and more this rebase deal is coming up and I'm starting
> to be interested, for curiosity purposes if no other reason.
>
> The perlrebase maintainer even posted that he wrote the script because
> he was having a problem repeatedly.
>
> It seems as if rebase and friends are suggested to fix a wide array of
> strange problems, which I have never seen.
>
> So here are my questions:
>
> 1) What problem does rebase actually fix that seems to fix so many
> different strange problems?
Loading DLLs if processes fork children. To match Unix semantics, the DLL needs
to be loaded into the same place. If that doesn't work out because the child
pulls in libraries differently than the parent, the processes get killed. I've
frequently seen issues with Subversion, for instance.
> 2) Why do some people seem to have this problem often, and others
> almost never?
Depends on the characteristics of the software used. The way I used vim and
gcc, I've never had issues with those, but Git-SVN is peculiar...
Rebase has been discussed in lengths, the thing I see open is that rebase,
unlike peflagsall and perlrebase, does NOT make .dlls writable if they aren't
(with chmod, that is), so it chooses to fail.
I wonder if ASLR (that which you can enable with peflagsall) should become the
default on systems that support it. It seems to help quite a bit, too, on
systems that support it (for instance, Windows 7).
--
Matthias Andree
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
- Raw text -