Mail Archives: cygwin/2003/07/10/15:51:42
On Thu, Jul 10, 2003 at 02:33:09PM -0500, Brian Ford wrote:
> Is it true that I should not recompile anything under 1.5.0 unless
> all the dlls it uses have been recompiled first under same? If not all,
> how do I identify the subset?
By testing. It's save to use older DLLs if they don't expect any of
the changed datatypes as parameter or part of a parameter. This
part of the application is of course not 64 clean. However, for testing
purposes I've build OpenSSH using the current OpenSSL and it still
worked fine. Just as a prove of concept.
> So, 1.5.0 headers are not safe for development, except for dll package
> maintainers recompiling for the above purpose, and for those carefully
> checking from the bottom up?
Yes.
> Is it safe to use 1.3.22 headers with a 1.5.0 dll (I think so)? Obviously
> the converse is not true.
It's safe to use 1.3.22 headers *together* with 1.3.22 libcygwin.a as well
as it is safe to use 1.5.0 headers *together* with 1.5.0 libcygwin.a when
building a package, Headers and libcygwin.a are birds of a feather.
Mixing 1.3.x headers with 1.5.0 libcygwin.a and vice versa will very likely
create a special application which main purpose is to test how to get
segmentation violations.
Packages build with 1.5.0 headers and libcygwin.a will not run under
1.3.x. Packages build with 1.3.x headers and libcygwin.a will run
under 1.5.0 but they are not aware of the new datatype sizes.
Corinna
--
Corinna Vinschen Please, send mails regarding Cygwin to
Cygwin Developer mailto:cygwin AT cygwin DOT com
Red Hat, Inc.
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
- Raw text -