On Thu, Nov 19, 2020 at 09:49:05PM +0100, Magnus Hagander wrote:
> Ugh, that's pretty terrible.
I have spent some time testing this part this morning, and I can see
that genhtml does not complain with your patch. It looks like in the
case of 2771fce the tool got confused because the same function was
getting compiled twice for the backend and the frontend, but here you
only get one code path compiled depending on the option used.
+#else /* not OPENSSL or WIN32 */
I think you mean USE_OPENSSL or just OpenSSL here, but not "OPENSSL".
pg_strong_random.c needs a pgindent run, there are two inconsistent
diffs. Looks fine except for those nits.
--
Michael