!!! Fetched file: eric-3.3.1.tar.gz MD5 FAILED! Removing corrupt distfile... I don't now why the md5 changed, but I made the ebuild myself so the md5 is not the problem. I just wanted to backup the tarball, which was deleted accidentally by myself. The problem: Portage is fetching and deleting the tarball again and again. After ~20 times I detected and stopped it, but unattended installations in conjunction with such a problem will stress mirrors unneccessarily. I'm using portage-2.0.49-r21 here.
Maybe, the tar ball was changed in upstream.. If you regenerate md5 and it occurs again, feel free to reopen this bug.
I don't think this bug report is invalid. Couldn't you add a variable (maybe configurable via make.conf) to limit the number of attempts to fetch a tarball?. I mean, if you are fetching the tarball a couple of times from different servers and it's md5 is always the same, but doesn't match the digest, it's likely that the problem is not the tarball.
If the MD5 of the tarball doesn't match the digest there is a problem. There is no way for portage to decide if the digest is wrong or the tarball, even if all mirrors have a tarball with the same MD5, as the maintainer could have uploaded a wrong tarball. With your suggestion we wouldn't even notice such an error (ok we would probably notice it, but it would certainly take longer).
I want a alway_stop_after_N_trials option. Is my english that bad?
