Thread: Backend specific ifdefs in sha2.h

The #ifdef guards in sha2.h are using USE_SSL when they in fact are guarding
the inclusion of OpenSSL specific code.  This has never caused any issues as
there only is a single supported TLS backend in core so far, but since we’ve
spent a significant amount of energy on making the TLS backend non-hardcoded
it seems we should fix this too.  The Makefile around sha2.c/sha2_openssl.c is
already testing for openssl rather than ssl (which given src/Makefile.global
variables makes perfect sense of course).

cheers ./daniel

On Thu, Jun 13, 2019 at 09:32:28AM +0200, Daniel Gustafsson wrote:
> The #ifdef guards in sha2.h are using USE_SSL when they in fact are guarding
> the inclusion of OpenSSL specific code.  This has never caused any issues as
> there only is a single supported TLS backend in core so far, but since we’ve
> spent a significant amount of energy on making the TLS backend non-hardcoded
> it seems we should fix this too.  The Makefile around sha2.c/sha2_openssl.c is
> already testing for openssl rather than ssl (which given src/Makefile.global
> variables makes perfect sense of course).

Right, good catch.  I would not back-patch that though as currently
USE_SSL <=> USE_OPENSSL.  Any suggestions or thoughts from others?
--
Michael

> On 13 Jun 2019, at 10:29, Michael Paquier <michael@paquier.xyz> wrote:

> I would not back-patch that though as currently
> USE_SSL <=> USE_OPENSSL.

Right, there is no use in backporting of course.

cheers ./daniel

On Thu, Jun 13, 2019 at 10:31:23AM +0200, Daniel Gustafsson wrote:
> Right, there is no use in backporting of course.

And applied now, in time for beta2.
--
Michael