From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([209.51.188.92]:43728) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1hN1v1-0008WR-IC for guix-patches@gnu.org; Sat, 04 May 2019 17:10:04 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1hN1v0-0001DG-Hz for guix-patches@gnu.org; Sat, 04 May 2019 17:10:03 -0400 Received: from debbugs.gnu.org ([209.51.188.43]:38921) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1hN1v0-0001Cb-EH for guix-patches@gnu.org; Sat, 04 May 2019 17:10:02 -0400 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1hN1v0-0005Mo-24 for guix-patches@gnu.org; Sat, 04 May 2019 17:10:02 -0400 Subject: [bug#34807] [PATCH 1/2] Add (guix lzlib). Resent-Message-ID: From: Ludovic =?UTF-8?Q?Court=C3=A8s?= References: <20190310180209.11578-1-mail@ambrevar.xyz> <8736ne3855.fsf@gnu.org> <87pnp2f7gr.fsf@ambrevar.xyz> <8736lxdxn6.fsf@ambrevar.xyz> <87lfzm7fdz.fsf@gnu.org> <878svm5xic.fsf@ambrevar.xyz> Date: Sat, 04 May 2019 23:09:40 +0200 In-Reply-To: <878svm5xic.fsf@ambrevar.xyz> (Pierre Neidhardt's message of "Sat, 04 May 2019 12:23:23 +0200") Message-ID: <87ef5e0vvv.fsf@gnu.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-patches-bounces+kyle=kyleam.com@gnu.org Sender: "Guix-patches" To: Pierre Neidhardt Cc: 34807@debbugs.gnu.org Hello! Pierre Neidhardt skribis: > Right on time, I just finished it! > > - I've been in touch with Antonio, Lzip's maintainer, for more than a > week and now I'm confident that I have a decent understanding of the > library. > > - Your m4 suggestion didn't work. I've included a comment. We need to > fix it before merging. I'm not the right person for this job I'm > afraid :p Ludo? Sure, I can do it. > - The convenience functions do not support multi-member archives. > Multi-member archives are mostly useful for parallelization, but we > don't use that in Guix, so it's OK. Should it be required some day, > we would need to implement it, which requires a little bit more work. > I've documented all that. > > - The implementation of lzread! is subpar because I understood a > subtlety a bit too late. But that's alright, it does not affect > performance nor reliability. > > - I've included 11 tests covering all your suggestions. > > - I haven't strace'd the Guile process. The code regarding ports is > identical to zlib.scm, so it's unlikely there would be an issue in > this area. I have never done this before, so out of curiosity, how do > you run a specific Guix tests without going through `make'? > > Next steps? :D This looks all good to me! I was about to apply it and add the Autoconf machinery, but I thought we could also make it a separate project that could be beneficial to other Guilers out there (like we did with Guile-Gcrypt and Guile-Git). Incidentally that would also avoid the need for adding the =E2=80=98%liblz= =E2=80=99 variable in (guix config), which simplifies things a bit. WDYT? If you want to take that route, I=E2=80=99m happy to help with the Autotools machinery (or you could use =E2=80=98hall=E2=80=99 from the =E2=80=98guile-= hall=E2=80=99 package to do that for you.) If you don=E2=80=99t feel like taking that route (or at least not yet ;-)), that=E2=80=99s OK for me too, I don=E2=80=99t feel strongly either way. Thoughts? Thank you! Ludo=E2=80=99.