On 12/3/05, mark benedetto king <email@example.com> wrote:
> > The only thing I don't understand is whether the patch to
> > subversion/libsvn_ra_dav/session.c is actually necessary. I don't
> > have a client side cert for any svn repo, so I couldn't test the SSL
> > cert password bits.
> I had the same question about the original patch. I haven't tried to
> store cert passwords in the keychain yet. I will commit the simple
> work first and then work through the SSL case.
I haven't tested the SSL code, but I noticed a small bug:
>+ svn_auth_save_credentials(state, pool);
> svn_auth_cred_ssl_client_cert_pw_t *pw_creds = creds;
Looks like we've got a declaration here after a statement. To allow
this code to parse as C89, we'll need to switch the order of these two
> 2.) Configury.
> +if test "$enable_keychain" = "yes"; then
> + AC_MSG_RESULT([yes])
> + SVN_KEYCHAIN_INCLUDES=""
> + SVN_KEYCHAIN_LIBS="-framework Security -framework CoreFoundation -framework CoreServices"
> + AC_DEFINE([SVN_HAVE_KEYCHAIN_SERVICES], , [Is Mac OS KeyChain support enabled?])
> + AC_SUBST(SVN_KEYCHAIN_INCLUDES)
> + AC_SUBST(SVN_KEYCHAIN_LIBS)
> + AC_MSG_RESULT([no])
Shouldn't SVN_KEYCHAIN_INCLUDES and SVN_KEYCHAIN_LIBS be substituted
unconditionally? In Makefile.in, I see that @SVN_KEYCHAIN_INCLUDES@
was used inside INCLUDES on all platforms.
David James -- http://www.cs.toronto.edu/~james
Received on Sat Dec 3 16:45:35 2005