Created attachment 447658 [details] emerge --info file Following the failures to install dev-lang/ruby:2.1.9, I hand-crafted an ebuild and got dev-lang/ruby to install; however, rubygems fails to install and kill all the dependencies of dev-lang/ruby. Will attach "emerge --info' 'build.log' 'environment' and the relevant 'ebuild' for the failures. As with dev-lang/ruby, this blocks the installation of kde-plasma/plasma-meta on a fresh install.
Created attachment 447660 [details] build log file for dev-ruby/rubygems-2.2.5-r1
Created attachment 447662 [details] ebuild environment file
Created attachment 447664 [details] ebuild file for dev-ruby/rubygems-2.2.5-r1
dev-ruby/json also fails for the same reason rubygems fails: The implementation of 'each_ruby_configure' and '_ruby_each_implementation' fails to set/maintaion the {$RUBY} variable and the build fails because of a bad environment. Do you want a formal new bug for dev-ruby/json filled out? This is attempting to do the recent updates released to the unstable tree.
(In reply to G.Wolfe Woodbury from comment #4) > dev-ruby/json also fails for the same reason rubygems fails: > > The implementation of 'each_ruby_configure' and '_ruby_each_implementation' > fails to set/maintaion the {$RUBY} variable and the build fails because of a > bad environment. I think these are the result of your hand-crafted dev-lang/ruby ebuild. I'm inclined to close this as INVALID, but let's see if we can fix the dev-lang/ruby issues first and then retest this.
I am not using a "hand-crafted" ebuild when testing for the bug. The patch is a partial fix for one problem. I always test against the vanilla ebuilds.
I pulled the latest rubygems (2.6.7) and did an update. Enclosed are the einfo, ebuild and build log results. No change in status.
Created attachment 448910 [details] latest emerge --info
Created attachment 448912 [details] dev-ruby/rubygems ebuild used just pulled from portage
Created attachment 448914 [details] current build log (2.6.7)
(In reply to G.Wolfe Woodbury from comment #6) > I am not using a "hand-crafted" ebuild when testing for the bug. The patch > is a partial fix for one problem. I always test against the vanilla ebuilds. To state it more precisely: I've done enough QA and Testing in my time that when filing a bug, I will present the ebuild and sources "as is". Anything I do in terms of making my own versions is done in a VM and isolated from the stuff I file. It is hard enough to isolate and figure the precise dynamics of complex software, I would not want to make it any harder for the maintainers by making them deal with my changes as well.
Attaching the full build information from the latest attempt to install/update the rubygems package while updating kde and qt stuff. I am almost to the pooint of attempting to install ruby on a stable 64-bit virtualbox vm: if there are problems there (or not!) I will report them here. attaching: einfo.txt einfo-pqv.txt environment.txt rubygems-2.6.7.ebuild rubygems-2.6.7:20161012-234425.log screencap.txt
Created attachment 450048 [details] emerge --info for attempt on 2016/10/12 This batch of attachments is the full information for a build attempt on 2016/10/12
Created attachment 450050 [details] emerge -pqv output for attempt on 2016/10/12
Created attachment 450052 [details] environment dump from attempt 2016/10/12
Created attachment 450054 [details] ebuild as recorded on my system 2016/10/12
Created attachment 450056 [details] portage build log of attempt 2016/10/12
Created attachment 450058 [details] screen capture of error 2016/10/12 tail end of screen dump and error messages to CLI console doing the emerge.
Actually did a new install of dev-lang/ruby:2.3 on a new machine. ruby23 built and installed, but rubygems et alia did not install; failing the same way as documented. [Full diagnostics available if requested]
