[Info-vax] An HP SSL mystery

[Stephen Hoffman]

On 2015-06-23 04:18:23 +0000, Steven Schweda said:

>    The quick (obvious?) work-around seems to do the job:
>       define SSLROOT SSL$ROOT
> 
>    So, is this a long-standing bug in HP SSL, or am I missing something 
> obvious (again), or what?  (I noticed nothing relevant in the HP SSL 
> release notes.)
> 
>    Observed with HP AXPVMS SSL V1.4-502 and HP I64VMS SSL V1.4-502, and 
> in some older 1.4-3?? version (which I updated to -502, to no avail).

I'd guess that the porting scripts omitted that step, or somebody 
following the scripts missed that particular step.  Based on some 
off-hand comments from HP folks and based on the turn-around times for 
the ports, it appears that HP is using some combination of scripts and 
manual steps to port over the code.

That this was shipped implies that whoever was porting OpenVMS had both 
versions of SSL installed.

I've not reversed the executables to see if there are any significant 
divergences from the original OpenSSL source code, beyond the symbol 
name changes.  But I've not noticed any, nor encountered any documented 
divergences.

We're due for a change here, though.  Updates for this OpenSSL 0.9.8 
series that HP is using end at the end of 2015; this year.   HP has not 
commented on what they're planning to do, with the end of support for 
their upstream.  Logically, they'll move forward to the 1.0.2 series, 
but I've seen no plans nor public comments.   As for VSI plans?   
Locally, I've been moving to the 1.0.1 and 1.0.2 series and to more 
current ports, where the customer will accept that.


-- 
Pure Personal Opinion | HoffmanLabs LLC