From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.gentoo.org (woodpecker.gentoo.org [140.211.166.183]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id A617953AA for ; Sun, 2 Apr 2023 18:48:52 +0000 (UTC) User-agent: mu4e 1.10.1; emacs 29.0.60 From: Sam James To: distributions@lists.linux.dev Subject: Possible issues with arpack-3.9.0 Date: Sun, 02 Apr 2023 19:46:13 +0100 Message-ID: <87zg7qt90v.fsf@gentoo.org> Precedence: bulk X-Mailing-List: distributions@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: multipart/signed; boundary="=-=-="; micalg=pgp-sha512; protocol="application/pgp-signature" --=-=-= Content-Type: text/plain Hi folks, I came across this a week or so ago and it came back to mind because of the openblas kerfuffle [0]. arpack-3.9.0 will in some cases give deterministic failure as the randomness in the retry mechanism seems busted. This can cause e.g. igraph's test suite to fail. References: - https://github.com/opencollab/arpack-ng/issues/401 - https://github.com/opencollab/arpack-ng/issues/410 Ths discussion is quite long and the second bug linked tries to summarise it. It's complicated and for now I've just yanked arpack-3.9.0 in Gentoo (masked it). [0] https://lore.kernel.org/distributions/CAE6_+UeiCzpyA9-XVgX1x5Ucjccng0rdw-+WCZbtMnLX8VOKgQ@mail.gmail.com/T/#m6ed52b95d975ae3b047f5704e6f2cc9914e112fe best, sam --=-=-= Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iOUEARYKAI0WIQQlpruI3Zt2TGtVQcJzhAn1IN+RkAUCZCnOEF8UgAAAAAAuAChp c3N1ZXItZnByQG5vdGF0aW9ucy5vcGVucGdwLmZpZnRoaG9yc2VtYW4ubmV0MjVB NkJCODhERDlCNzY0QzZCNTU0MUMyNzM4NDA5RjUyMERGOTE5MA8cc2FtQGdlbnRv by5vcmcACgkQc4QJ9SDfkZBqkwD/WE1V+PDmcA60Z+Jq7jb4nghlEcjnOL8tFEJF lFLA0KEA/12ndmL8ByBf2OYn4+ORC/SvmIoCiP7WIuia622gNPYD =q0gX -----END PGP SIGNATURE----- --=-=-=--