On 8/13/13 10:28 AM, Josef Bacik wrote:
> On Sun, Aug 11, 2013 at 09:54:41PM -0500, Eric Sandeen wrote:
>> On Aug 11, 2013, at 8:15 PM, Dave Chinner <david@xxxxxxxxxxxxx> wrote:
>>
>>> On Thu, Aug 08, 2013 at 12:40:55PM -0500, Eric Sandeen wrote:
>>>> On 8/8/13 3:17 AM, Jan Schmidt wrote:
>>>>> SUBDIRS =
>>>>>
>>>>> LLDLIBS = $(LIBATTR) $(LIBHANDLE) $(LIBACL)
>>>>>
>>>>> +OPT_LDLIBS = -lssl -lcrypto
>>>>
>>>> Hm, new deps. I guess it's not a huge problem, these should always
>>>> be available, right?
>>>
>>> Please, let's keep crypto out of xfstests if we can. That's just
>>> going to add a nightmare of US export compliance garbage to any
>>> distro that wants to package and ship this....
>>>
>> It's ssl that's (optionally) required for check summing I believe. But it's optional if done properly...
>>
>
> It doesn't build without -lcrypto for me, so either we have an external
> dependancy (which I'm fine with), or we add it to xfstests (which is what you
> wanted). Does a config option alleviate your concerns? And I it's just linking
> it in at build time, I don't think we have to work about export compliance as
> long as we aren't shipping actual crytpo stuff right? Anybody who is going to
> be running this stuff is going to have libssl installed, which installs
> libcrypto (at least on my box) so I don't think it's a huge deal to have us
> linking it. Thanks,
Yeah I think linking in a consumer is not at all the same as exporting the functionality.
Looks like my Makefile patch needs to add both -lssl & -lcrypto after all,
but again, yeah, just linking against it _if_ it's there.
Oh, and here's my m4/package_ssldev.m4, sorry about that:
AC_DEFUN([AC_PACKAGE_WANT_SSL],
[ AC_CHECK_HEADERS(openssl/md5.h, [ have_ssl=true ], [ have_ssl=false ])
AC_SUBST(have_ssl)
])
-Eric
> Josef
>
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
