Emerging sci-electronics/kicad-20130518 fails with many compiler errors when attempting to build with gcc-5.1. Reproducible: Always Build log and output of `emerge --info '=sci-electronics/kicad-20130518::gentoo'` attached to bug. Output of: `emerge -pqv '=sci-electronics/kicad-20130518::gentoo'`: [ebuild N ] sci-electronics/kicad-20130518 USE="-debug" LINGUAS="en -de -es -fr -hu -it -ja -pl -pt -ru -zh_CN"
Created attachment 405466 [details] build.log xz-compressed
Created attachment 405468 [details] emerge --info output
sci-electronics, this seems up your alley.
This package is building against a bundled copy of boost and the error is coming from there. I noticed we're already applying a patch to cmake to let it find the system copy but it looks like that isn't enough. From http://www.kicad-pcb.org/display/DEV/Known+System+Related+Issues "There has previously been problems with internal bugs in boost, whereby at one point in time the developers decided to include download, patch and build of boost in the KiCad build framework. Many users have been reluctant to understand this fact. Once a build option to skip this download and build was added, such that you can save some initial build time and actually use the boost you already have installed on your system. This downloading depends inside the KiCad installer framework is being slowly removed again, because it has been acknowledged that is is a bad idea, because it clutters the KiCad installer. Instead such things should be an external project. If you have boost 1.56+ this skipping seems fairly safe now a days. Use the option -DKICAD_SKIP_BOOST=ON when cmake'ing." The version we have in portage doesn't support that option though.
Upstream is preparing for a new release, the first one in more than two years. I will have to massively overhaul the ebuild at that point and will look into that and other things. I will probably not backport any fixes to the current ebuild since it's so old it will no longer make sense to use it. That said, I'm very unlikely to fix any bug with a version of gcc that's not keyworded for any architecture, and especially ~amd64 since it's what my dev box uses. If you're using Kicad with an unkeyworded and/or masked version of gcc I'll be happy to take your patches though. And if you're a dev, feel free to commit your fixes directly.
Considering the only reason it's not keyworded is to give you time to fix these bugs, I don't think that's a valid argument. Fair enough though.
New 4.0.0-rc1 for kicad has landed into the portage tree. See what gcc-5.1 says to that.
purged from portage
