Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Polarhome OpenSSL kit missing a header file?

82 views
Skip to first unread message

Mark Daniel

unread,
Apr 28, 2012, 9:49:09 AM4/28/12
to
I was exploring the use of the pre-built Polarhome OpenSSL kit (thanks
to all concerned)

http://www.polarhome.com/openssl/

and the openssl-101b.zip (at least) appears to be missing an include file:

#include <openssl/srtp.h> /* Support for the use_srtp extension */
.^
%CC-F-NOINCLFILEF, Cannot find file <openssl/srtp.h> specified in
#include directive.
at line number 1363 in file SSLROOT:[include]ssl.h;1

$ dir openssl:srtp.h
%DIRECT-W-NOFILES, no files found
$

which also does not appear in the kit build log:

http://www.polarhome.com/openssl/INSTALL101.LOG

A quick look in ssl.h doesn't show any obvious macro for
including/excluding this functionality so it's not necessarily under the
builder's control. Any useful comment would be welcome.

TIA, Mark Daniel.

Steven Schweda

unread,
Apr 28, 2012, 1:32:20 PM4/28/12
to Steven M. Schweda
> and the openssl-101b.zip (at least) appears to be missing an include file:
>
> #include <openssl/srtp.h> /* Support for the use_srtp extension */
> .^
> %CC-F-NOINCLFILEF, Cannot find file <openssl/srtp.h> specified in
> #include directive.
> at line number 1363 in file SSLROOT:[include]ssl.h;1

When you do _what_, exactly? And what is "SSLROOT" when
you're doing this?

I haven't tried any OpenSSL 1.0.1 stuff since a
pre-release snapshot from "20101217", so I know nothing, but
if I were looking for working stuff, and if I had a C
compiler, then I'd start at openssl.org. When the
polarhome.com fellow says "TESTED", he doesn't always mean
what most people mean when they say "tested", and building it
yourself means that you get to enable some options (like zlib
support, for example) which might be missing from a pre-built
kit.

Mark Daniel

unread,
Apr 28, 2012, 4:12:57 PM4/28/12
to
On 29/04/12 3:02 AM, Steven Schweda wrote:
>> and the openssl-101b.zip (at least) appears to be missing an include file:
>>
>> #include<openssl/srtp.h> /* Support for the use_srtp extension */
>> .^
>> %CC-F-NOINCLFILEF, Cannot find file<openssl/srtp.h> specified in
>> #include directive.
>> at line number 1363 in file SSLROOT:[include]ssl.h;1
>
> When you do _what_, exactly? And what is "SSLROOT" when
> you're doing this?

There's sufficient information in the report for anyone not distracted
with composing notes of condescension. Give some of us some credit for
having some experience. The world-weary, oh here's another one to deal
with, is a bit thin with overuse.

> I haven't tried any OpenSSL 1.0.1 stuff since a
> pre-release snapshot from "20101217", so I know nothing, but
> if I were looking for working stuff, and if I had a C
> compiler, then I'd start at openssl.org. When the
> polarhome.com fellow says "TESTED", he doesn't always mean
> what most people mean when they say "tested", and building it
> yourself means that you get to enable some options (like zlib
> support, for example) which might be missing from a pre-built
> kit.

For anyone interested I grabbed an SRTP.H off the net and dropped it
into SSLROOT:[INCLUDE] (OPENSSL:) and the code compiles and runs fine.
Should have thought to do this before bothering with c.o.v.

My thanks again to Polarhome for making these kits available, and it
looks like the kit build procedure needs a tweak.

Mark Daniel

unread,
Apr 28, 2012, 9:58:12 PM4/28/12
to
On 29/04/12 5:42 AM, Mark Daniel wrote:
8< snip 8<
> For anyone interested I grabbed an SRTP.H off the net and dropped it
> into SSLROOT:[INCLUDE] (OPENSSL:) and the code compiles and runs fine.
> Should have thought to do this before bothering with c.o.v.
>
> My thanks again to Polarhome for making these kits available, and it
> looks like the kit build procedure needs a tweak.

An unrelated wrinkle when attempting feedback:

Recipient address: openssl-...@polarhome.com
Reason: Remote SMTP server has rejected address
Diagnostic code: smtp;550 5.1.1 <openssl-...@polarhome.com>...
User unknown
Remote system: dns;mail.polarhome.com

Steven Schweda

unread,
Apr 28, 2012, 11:19:06 PM4/28/12
to Steven M. Schweda
> [...] condescension. Give some of us some credit for
> having some experience. [...]

I'm sorry. You're right, of course. I should remember
that the less useful information in the problem report, the
more expert the source. I'll try to assume more next time.

> [...] I grabbed an SRTP.H off the net [...]

That URL should provide all the info one might wish when
looking for the replacement file. (Another reason to get a
source kit, and build it yourself, I claim.)

Note that the usual OpenSSL self-test procedure normally
works from the source+build tree, not the installation target
tree, so it's not well suited to find a problem like this
in the installation procedure itself. If I still had
convenient access to VMS patches, then I'd probably spend
more time working on OpenSSL for VMS, but I'm trying not to
care more about it than the folks at HP.

Steven Schweda

unread,
May 2, 2012, 3:39:28 PM5/2/12
to Steven M. Schweda
Has anyone reported this problem to OpenSSL.org (or even
to Polarhome.com)? That'd probably be a good idea if you
care about this stuff. (I haven't seen anything on the
OpenSSL development e-mail forum.)

Mark Daniel

unread,
May 2, 2012, 4:36:53 PM5/2/12
to
Your moniker is on the polarhome page Steve

http://www.polarhome.com/openssl/

and this seems to be a polarhome issue, as well as

http://groups.google.com/group/comp.os.vms/msg/b646dabe62beb103?hl=en

Steven Schweda

unread,
May 2, 2012, 5:53:51 PM5/2/12
to Steven M. Schweda
> Your moniker is on the polarhome page Steve

And your point is? (Not my Web site. Not my program.)

Back when I still had VMS patch access, I did more work on
software like this for VMS. (Before OpenSSL 1.0.0e, some of
the VMS stuff was pretty raggedy. _Saying_ that you can make
a 64-bit library is not the same as actually being able to
make one, for example. A suitable Web search should be able
to find other examples.) But, as I've said before, working
on software for VMS (as a hobbyist VMS user) has been more
satisfying than it is now. My back is available, but HP
seems disinterested in scratching it. I find it hard to stay
more interested in VMS than its vendor is.
0 new messages