Using catalyst 1.1.0 and an old stage3 tarball (my own), my new stage1 build kept failing with errors something like "--newuse isn't a valid option". Reproducible: Always Steps to Reproduce: Use a seed stage with an older portage, and the build will fail with a cryptic error message. That is, the error message is from the embedded portage, but it's not trivial to diagnose. Expected Results: Catalyst clearly depends on the version of portage in the seed stage. It should check the version number against an internal minimum at run-time. (This can probably even be grokked from the ebuild at build-time, so you wouldn't have to keep track of it twice.)
zhen: can you script up some foo for catalyst to ensure that at least portage 2.0.51 is used in the seed stages? If not, I'll just do it in the bash, which I think will be a bit less elegant.
bash foo it - catalyst stays out of the chroot. Use portageq ;)
Way to pass the buck, zhen... ;] I was actually hoping that you could have caught it in catalyst in the python, as it could have saved a bit of repetition. Unpack source -> check portage version -> bail It would have only had to have been added once and could have just been applied to all targets. Anyway, I used my nasty bash-fu and showed it who was boss.
Catalyst 1.1.2 is now out!