* Detected file collision(s): * * /usr/bin/console * * Searching all installed packages for file collisions... * * Press Ctrl-C to Stop * * app-admin/conserver-8.1.18-r1:0::gentoo * /usr/bin/console * * Package 'dev-ruby/power_assert-1.0.1' NOT merged due to file ----------------------------------------------------------------- This is an unstable amd64 chroot image (named desktop-unstable_20170121-000413) at a hardened host acting as a tinderbox. ----------------------------------------------------------------- gcc-config -l: [1] x86_64-pc-linux-gnu-5.4.0 * llvm-config --version: 3.9.1 Available Java Virtual Machines: [1] icedtea-7 system-vm [2] icedtea-bin-7 [3] icedtea-bin-8 Available Python interpreters, in order of preference: [1] python3.4 [2] python2.7 (fallback) [3] pypy (fallback) [4] jython2.7 (fallback) Available Ruby profiles: [1] ruby21 (with Rubygems) * java-config: The following VMs are available for generation-2: *) IcedTea JDK 7.2.6.8 [icedtea-7] 2) IcedTea JDK 7.2.6.8 [icedtea-bin-7] 3) IcedTea JDK 3.2.0 [icedtea-bin-8] -----------------------------------------------------------------
Created attachment 461982 [details] emerge-info.txt
Created attachment 461984 [details] dev-ruby:power_assert-1.0.1:20170131-112815.log
Created attachment 461986 [details] emerge-history.txt
Created attachment 461988 [details] etc.portage.tbz2
I believe we've seen this problem a couple times before with ruby packages (#571186, #593068).
(In reply to Daniel M. Weeks from comment #5) yes, is conserver guilty ?
(In reply to Toralf Förster from comment #6) > (In reply to Daniel M. Weeks from comment #5) > yes, is conserver guilty ? No. In every case it's an experimental stub for the ruby package that's incorrectly installed and the ruby package in question needs the fix. The root cause appears to be that many ruby packages use Bundler to generate a template project or use it's skeleton as a standard specification. Bundler includes an interactive demo script named "console" in it's skeleton. When the ruby ebuild are made care isn't taken to remove these useless scripts before installation. With the pervasiveness of the problem (and since ruby packages could collide with each other for the same reason) it almost seems like one of the ruby eclasses should remove the file(s) by default and only include them when an ebuild has made an explicit exception.
Fixed with power_assert-1.0.1-r1
(In reply to Daniel M. Weeks from comment #7) > With the pervasiveness of the problem (and since ruby packages could collide > with each other for the same reason) it almost seems like one of the ruby > eclasses should remove the file(s) by default and only include them when an > ebuild has made an explicit exception. Yes, a more structural solution would probably be a good idea, I'll have a look at this.
