https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=274099

--- Comment #25 from John Hein <jcfyecr...@liamekaens.com> ---
(In reply to Muhammad Moinur Rahman from comment #23)
(In reply to jwdevel from comment #24)
Yes, the build with the 's|00907000|00000000|' substitution is still using
include files from /usr/include.  Try adding this little patch to print out the
search path for opensslv.h:

@@ -889,6 +890,8 @@ class PyBuildExt(build_ext):
         # look for the openssl version header on the compiler search path.
         opensslv_h = find_file('openssl/opensslv.h', [],
                 inc_dirs + search_for_ssl_incs_in)
+        print('opensslv_h: %s' % str(opensslv_h))
+        print('inc_dirs: %s' % str(inc_dirs))
         if opensslv_h:
             name = os.path.join(opensslv_h[0], 'openssl/opensslv.h')
             if host_platform == 'darwin' and is_macosx_sdk_path(name):

Then do a build and look for opensslv_h in the build output.

And, yes, even if your build succeeds, you may be linking with the ports ssl
library, but using the base version of opensslv.h at compile time.  Will that
work?  It might - as long as the two opensslv.h files are compatible "enough". 
But it also may lead to subtle breakage.  It's better to compile with the right
header that matches the library that is being linked.

The 's|00907000|00000000|' version hack should probably be removed in favor of
the attachment 245256 patch in comment 16 (that patch does not remove the
version hack, but that should probably be done).  The ssl version hack is quite
fragile and dependent on a specific version that may not be right with the
passage of time.

Despite the attitude in comment 18 from a couple years ago to not do anything
to ports that may soon be removed, I think applying the more correct change
should almost always be the right course of action, especially when it is
fairly trivially implemented and verifiable.  Even if it is removed "soon",
this is low effort and improves the lives of users in the mean time.

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.

Reply via email to