From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Wed, 5 Sep 2018 08:53:18 +0200 Subject: [Buildroot] [PATCH 3/3] qt5webkit: select leveldb package In-Reply-To: <01e51af4-335b-7b8b-7c03-c8c00960eb20@mind.be> References: <20180831202204.28093-1-gael.portay@savoirfairelinux.com> <20180831202204.28093-4-gael.portay@savoirfairelinux.com> <01e51af4-335b-7b8b-7c03-c8c00960eb20@mind.be> Message-ID: <20180905085318.4ae7a818@windsurf> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hello, On Tue, 4 Sep 2018 23:10:46 +0200, Arnout Vandecappelle wrote: > On 31/08/2018 22:22, Ga?l PORTAY wrote: > > + depends on BR2_INSTALL_LIBSTDCPP # leveldb > > I don't think we need to propagate this particular dependency - qt5webkit > already depends on libstdcpp indirectly through Qt5. I think we need at some point to define a real policy for this. When a dependency is redundant, should we: (1) Avoid duplicating it, to simplify things, but with the risk that if we remove the dependency at a higher-level, we forget to re-add back to the sub-options. (2) Always duplicate them, even if they are useless right now, so that we don't forget them in the future if the higher-level option drops this dependency. Drawback of this solution is that such dependencies don't get "exercised" by autobuilder testing so they are often wrong. I don't have a strong opinion, but we're doing (1) or (2) depending on the package developer and/or situation, and this isn't really nice. Best regards, Thomas -- Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com