Please drop the multiple KEYWORDS definitions as it confuses tools like ekeyword and NATTkA. The same pattern seems to be used in www-client/google-chrome*, www-plugins/chrome-binary-plugins, www-client/opera*, www-client/microsoft-edge-dev.
The KEYWORDS in these builds never change. Why would you need to run nattka or ekeyword on them?
(In reply to Mike Gilbert from comment #1) > The KEYWORDS in these builds never change. Why would you need to run nattka > or ekeyword on them? There are (broadly) two sets of > 1 KEYWORDS lines in tree: 1) dev-ros/* 2) Chromium ebuilds. If the Chromium project is never going to switch to the usual stabilisation workflow, and the QA team is happy, then I'm happy to have an exception for binary-only ebuilds or whatever (as I said on #gentoo-qa when sultan asked).
(In reply to Sam James from comment #2) > If the Chromium project is never going to switch to the usual stabilisation > workflow, and the QA team is happy, then I'm happy to have an exception for > binary-only ebuilds or whatever (as I said on #gentoo-qa when sultan asked). ... or preferably, if this is that static, just put it in the eclass.