[Bug 257744] math/lapacke: relies on buggy base system ranlib behaviour
Date: Tue, 10 Aug 2021 21:07:14 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=257744
--- Comment #2 from commit-hook@FreeBSD.org ---
A commit in branch main references this bug:
URL:
https://cgit.FreeBSD.org/ports/commit/?id=6fd5f57d0613a9f6d1816a912f21728b0d12435a
commit 6fd5f57d0613a9f6d1816a912f21728b0d12435a
Author: Ed Maste <emaste@FreeBSD.org>
AuthorDate: 2021-08-10 16:01:23 +0000
Commit: Ed Maste <emaste@FreeBSD.org>
CommitDate: 2021-08-10 21:03:05 +0000
math/lapack: remove superfluous and broken ranlib invocation
FreeBSD's base system ar and ranlib have a bug where they exit with
status 0 (success) even in case of fatal errors. This hid the fact that
math/lapacke was invoking ranlib on a non-existent file. (Presumably the
ranlib invocation was correct when introduced, but broken by some
rework of lapack's upstream build system).
Use of ranlib is generally unncessary, assuming the -s flag is passed to
ar (as is typical, and as done here), so just delete the invocation.
See PR 257599 and review D31402 for the ar/ranlib base system bug.
PR: 257599, 257744
Reviewed by: jrm
Tested by: jrm
Approved by: kevans (ports), portmgr (implicit, blanket: build fix)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D31492
math/lapack/files/static.mk | 1 -
1 file changed, 1 deletion(-)
--
You are receiving this mail because:
You are the assignee for the bug.