Re: pgsql: Refactor MD5 implementations according to new cryptohash infrast - Mailing list pgsql-committers

From Michael Paquier
Subject Re: pgsql: Refactor MD5 implementations according to new cryptohash infrast
Date
Msg-id X9GdIu7QsbrBouXe@paquier.xyz
Whole thread Raw
In response to Re: pgsql: Refactor MD5 implementations according to new cryptohash infrast  (Michael Paquier <michael@paquier.xyz>)
List pgsql-committers
On Thu, Dec 10, 2020 at 12:18:57PM +0900, Michael Paquier wrote:
> And this has broken uuid-ossp.  The switch is quite easy to do as this
> module just needs to use cryptohash APIs for MD5.  Will fix in a
> minute..

So, I have reproduced the failure with --with-uuid=e2fs and now fixed
it with 525e60b.

FWIW, I got to wonder this morning whether we should have a sha1
option for cryptohash.c, and seeing how ugly is the dependency on
md5.c and sha1.c between pgcrypto and uuid-ossp, this may give an
extra reason to do so..
--
Michael

Attachment

pgsql-committers by date:

Previous
From: Michael Paquier
Date:
Subject: pgsql: Fix compilation of uuid-ossp
Next
From: Alexander Korotkov
Date:
Subject: Re: pgsql: Support subscripting of arbitrary types, not only arrays.