https://blogs.gentoo.org/ago/2020/07/04/gentoo-tinderbox/ Issue: sci-libs/brial-1.2.11 fails tests (DASH-SYSTEM). Discovered on: amd64 (internal ref: dash_tinderbox) NOTE: (DASH-SYSTEM) in the summary means that the bug was found on a machine that runs /bin/sh as dash but this bug MAY or MAY NOT BE related to dash.
Created attachment 848177 [details] build.log build log and emerge --info
I'm guessing this one has something to do with the Unable to unshare: ENOSPC (for FEATURES="ipc-sandbox network-sandbox... >>> Test phase: sci-libs/brial-1.2.11 but can we see the failed test log to be sure?
I thought I had posted something here. We have a reported issue about broken tests with boost 1.81.0+. https://github.com/BRiAl/BRiAl/issues/53 I am going to do a new release upstream to fix it. There is one thing that bugs me. The broken tests are about hashes that changed with boost versions. It seems wrong to me to do it that way as it will break every so often with a new release of boost. What is even the point of that testing?
(In reply to François Bissey from comment #3) > I thought I had posted something here. We have a reported issue about broken > tests with boost 1.81.0+. > https://github.com/BRiAl/BRiAl/issues/53 > > I am going to do a new release upstream to fix it. There is one thing that > bugs me. The broken tests are about hashes that changed with boost versions. > It seems wrong to me to do it that way as it will break every so often with > a new release of boost. What is even the point of that testing? You commented on the m4ri bug (#890431) instead. I just looked at these tests and I agree. I don't see any value in comparing the hash of an object with a magic number, especially if those hashes change every now and then. Even if the hashes were truly stable, how can you look at the test and tell that comparing the hash of some polynomial to 44023709327539078 is doing the right thing? I think they can be dropped, or replaced by some property that is actually guaranteed to hold. For example, you might call stableHash() twice if it's not cached, and compare the two for equality. Or check that different objects have different hashes.
brial 1.2.12 is now released. I have removed all hash testing of given values but I introduced a few hash comparison between some objects where we know whether they should match or not. I'll do a PR for the main tree later today.
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=08fd3aa913284cddc9f861e9b27b0087455d827d commit 08fd3aa913284cddc9f861e9b27b0087455d827d Author: François Bissey <frp.bissey@gmail.com> AuthorDate: 2023-01-15 00:55:17 +0000 Commit: Michael Orlitzky <mjo@gentoo.org> CommitDate: 2023-01-16 23:20:11 +0000 sci-libs/brial: add 1.2.12 Closes: https://github.com/gentoo/gentoo/pull/29111 Closes: https://bugs.gentoo.org/890425 Signed-off-by: Michael Orlitzky <mjo@gentoo.org> Signed-off-by: François Bissey <frp.bissey@gmail.com> sci-libs/brial/Manifest | 1 + sci-libs/brial/brial-1.2.12.ebuild | 36 ++++++++++++++++++++++++++++++++++++ 2 files changed, 37 insertions(+)
