[openssl-dev] Question about dynamically loadable engines on Cygwin / Mingw

Jeremy Farrell jeremy.farrell at oracle.com
Tue Feb 16 01:39:12 UTC 2016


On 15/02/2016 23:16, Richard Levitte wrote:
> In message <20160215.185953.117619649162395329.levitte at openssl.org> on Mon, 15 Feb 2016 18:59:53 +0100 (CET), Richard Levitte <levitte at openssl.org> said:
>
> levitte> In message <56C210E7.5080804 at oracle.com> on Mon, 15 Feb 2016 17:54:47 +0000, Jeremy Farrell <jeremy.farrell at oracle.com> said:
> levitte>
> levitte> jeremy.farrell> It sounds good, except shouldn't it be "capi.so" for
> levitte> jeremy.farrell> cygwin, like the other mainstream POSIXy
> levitte> jeremy.farrell> implementations? The point of cygwin is that it's
> levitte> jeremy.farrell> POSIX not Windows, and it generally follows common
> levitte> jeremy.farrell> practices of POSIXy OSes for things which aren't
> levitte> jeremy.farrell> specified by POSIX. It seems that it would be simpler
> levitte> jeremy.farrell> all round (for users as well as development) to treat
> levitte> jeremy.farrell> it the same as "normal" UNIX-like OSes except when it
> levitte> jeremy.farrell> absolutely has to be treated differently.
> levitte>
> levitte> In practice, it really doesn't matter, it all comes down what the DSO
> levitte> module supports, and the way I'm coding this, Configure will decide.
> levitte> So it's a preference and nothing else.  Me, I don't particularly care,
> levitte> but if it disturbs the Cygwin community to see one .dll too many, I'm
> levitte> ready to make the necessary changes (it's literally one line of code
> levitte> to change).
>
> I had myself a look around in my little installation, and tried these
> two commands:
>
>      find /usr/lib -name '*.so'
>
>      find /usr/lib -name '*.dll'
>
> The first one didn't even return a screenfull (in my 25 line terminal
> screen), and the overwhelming majority was OpenSSL 1.0.2 engines ;-)
>
> The second one, on the other hand, gave a *lot* more output.  All
> aspell, babl, gawk, gegl, perl(!) loadable modules are named with
> the FOO.dll naming convention (there are a few packages, a minority,
> that have named them cygFOO.dll)...  and the list goes on.
>
> So looking at how things seem to be done normally, I feel confident
> that FOO.dll is a sane choice, even though not strictly POSIX in its
> file name extension.  And like I said, it matters very little for any
> user, the goal is to allow this kind of command line:
>
>      openssl s_server -engine FOO
>
> No extension, just a name for the user to worry about.

Thanks Richard - it was just a thought, and clearly not a very helpful 
one. The rest of the proposal looks like a good improvement to me.

-- 
J. J. Farrell - not speaking for Oracle

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mta.openssl.org/pipermail/openssl-dev/attachments/20160216/0665ed72/attachment.html>


More information about the openssl-dev mailing list