From mboxrd@z Thu Jan 1 00:00:00 1970 From: Loic Dachary Subject: libcrush C API documentation Date: Wed, 25 Jan 2017 19:23:09 +0100 Message-ID: <14e5037f-65c0-51d5-c9bd-0d1302baf144@dachary.org> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Return-path: Received: from relay2-d.mail.gandi.net ([217.70.183.194]:36287 "EHLO relay2-d.mail.gandi.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751513AbdAYSXN (ORCPT ); Wed, 25 Jan 2017 13:23:13 -0500 Sender: ceph-devel-owner@vger.kernel.org List-ID: To: "Adam C. Emerson" Cc: Ceph Development Hi Adam, I would be great if you could review the libcrush C API documentation merge request at http://libcrush.org/main/libcrush/merge_requests/1/diffs. I deliberately documented part of the external functions because some of them ( crush_make_uniform_bucket for instance ) are support functions for a higher level function ( crush_make_bucket for instance ). If you think a function should be documented and it is not, this is not because it's work in progress, it's because I may have misjudged it. The unsupported features are undocumented (for instance the tree and straw algorithms). They should probably be compiled out via a #if before the first standalone release of the library binary package. But it did not seem to be a good idea to dive into that while documenting. You can skip the review of the struct crush_map fields. Nothing was done except changing the existing comments to match Doxygen conventions. There is much to do to improve this. Explaining the tunables is not for the faint of heart. The format chosen for the documentation is Doxygen, as we discussed. Thanks for suggesting it :-) Cheers -- Loïc Dachary, Artisan Logiciel Libre