On Wed, 16 Jan 2002, Roth, Kevin P. wrote:

> Here ya go. I have no idea if it's good code, but it should at least give
> you a start. It works on *my* machine, of course I have very up-to-date
> OpenSSL code. Unknown whether this will work on other
> machines/build-environments/etc...

Many thanks. I consider it good code. I only had to tweak it very slightly to
make it report 0.9.5 properly on one of my test machines that (luckily) is
lagging a bit behind in the SSL version race. It also works for my 0.9.6a
version.

The modified patch I committed can be viewed here:

http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/curl/curl/lib/version.c.diff?r1=1.14&r2=1.15

> I did NOT alter it to use the "S" (status) indicator. I'll leave that up to
> you guys if that's desired. A status of "f" is release, "0" is
> initial-beta, and "1-e" would be in between.

We haven't shown that before, and I can't see any good reason to start doing
that now. Can you?

> I also didn't muck with any of the auto-config stuff. If a proper test
> needs to be added for the existance of SSLeay() (over-and-above the test
> that produces the value for USE_SSLEAY), could someone else please
> volunteer? I haven't a clue how to create those.

The SSLeay() function seems to be entirely undocumented, but the name
suggests that it has been around long enough for us to assume that it'l be
around on all OpenSSL libraries that work with curl.

(For people with older OpenSSL libraries than 0.9.5, then please note that I
am interested in patches that'll make libcurl work with those libraries
again, but I can't do that myself. Details in bug report #504163:
http://sourceforge.net/tracker/index.php?func=detail&aid=504163&group_id=976&atid=100976
.)

-- 
    Daniel Stenberg -- curl groks URLs -- http://curl.haxx.se/