From mboxrd@z Thu Jan 1 00:00:00 1970 From: Amir P Subject: Re: [PATCH] gnu: link-grammar: New module. Date: Mon, 24 Oct 2016 18:19:02 +0300 Message-ID: Mime-Version: 1.0 Content-Type: multipart/alternative; boundary=001a11402228db5fa7053f9deb0f Return-path: Received: from eggs.gnu.org ([2001:4830:134:3::10]:41380) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1byh2Q-0006Sc-Iq for guix-devel@gnu.org; Mon, 24 Oct 2016 11:19:47 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1byh2P-0000BA-HS for guix-devel@gnu.org; Mon, 24 Oct 2016 11:19:46 -0400 Received: from mail-lf0-x235.google.com ([2a00:1450:4010:c07::235]:39809) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1byh2P-00009o-9k for guix-devel@gnu.org; Mon, 24 Oct 2016 11:19:45 -0400 Received: by mail-lf0-x235.google.com with SMTP id t133so1579731lff.6 for ; Mon, 24 Oct 2016 08:19:44 -0700 (PDT) List-Id: "Development of GNU Guix and the GNU System distribution." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org Sender: "Guix-devel" To: guix-devel@gnu.org --001a11402228db5fa7053f9deb0f Content-Type: text/plain; charset=UTF-8 On Sun, 23 Oct 2016 17:20:11 -0400 Leo Famulari wrote: > If sqlite and zlib are only used when building, but not when running > link-grammar, then they should be native-inputs. > > Otherwise, we will need to figure out how to make link-grammar retain > references to these libraries. The link-grammar library doesn't use the zlib library directly or indirectly. What happens is that zlib.h is needed for the compilation of its sat-solver code because it is included in a header file of the minisat library which is indirectly included by the sat-solver code. So zlib (actually zlib.h) is only used when building. However, if the sqlite library+headers exist when using "configure", HAVE_SQLITE will get defined, which will bring in code that can use sqlite (sqlite3_open etc. are referenced in the result link-grammar library). As this code is currently not fully functional, a solution may be to "configure" link-grammar without the presence of the sqlite library+headers. A better solution may be to add a link-grammar "configure" option like "--enable-sqlite". If you think it is indeed better, I will open an issue there for adding it - to be included in the next version - 5.3.12. In any case, 5.3.12 is needed to overcome additional problems in 5.3.11 (some of them are already fixed in the repository, and for some others pull-requests are still pending). --001a11402228db5fa7053f9deb0f Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
On=C2=A0Sun, 23 Oct 2016 17:20:11 -0400= =C2=A0Leo Famulari wrote:
> If sqlite and zlib are only used when building, but no= t when running
> link-grammar, then they should be native-inputs.
>
> Otherwise, we will need to= figure out how to make link-grammar retain
> references to these libraries.

= The link-grammar library doesn't use the zlib library directly or indir= ectly.
What happens is that zlib.h is nee= ded for the compilation of its sat-solver code
because it is included in a header file of the minisat library which = is indirectly
included by the sat-solver = code. So zlib (actually zlib.h) is only used when building.

However, if the s= qlite library+headers exist when using "configure",
HAVE_SQLITE will get defined, which will bring in code= that can use sqlite
(sqlite3_open etc. a= re referenced in the result link-grammar library).
As this code is currently not fully functional, a solution may be= to "configure"
link-grammar wi= thout the presence of the=C2=A0sqlite library+headers.
A better solution may be to add a link-grammar "configur= e" option like "--enable-sqlite".
If you think it is indeed better, I will open an issue there for ad= ding it - to be
included in the next vers= ion - 5.3.12.
In any case, 5.3.12 is need= ed to overcome additional problems in 5.3.11 (some=C2=A0
of them are already fixed in the repository, and for some o= thers pull-requests are
still pending).
--001a11402228db5fa7053f9deb0f--