Arun Isaac <arunisaac@systemreboot.net> writes:

> Test 92 of tar 1.30 fails on my system. I had to disable tar tests to get my
> cmake to build. I am not entirely sure if the test failure has got something
> to do with insufficient memory on my system. So, I'd like somebody to
> reproduce the problem before patching it.

That test fails on one of my build nodes with 64GB RAM as well.

There is a bug report about it upstream:
<https://lists.gnu.org/archive/html/bug-tar/2017-12/msg00026.html>.

However, the suggested fix does not work for me (I think it requires
running autoreconf, but haven't tried that yet).

> I'm a little confused now. Shouldn't I rebase the core-updates branch against
> master in order to have rhash? Wouldn't this disrupt the evaluation of
> core-updates on the build farm? Sorry I'm not very clear about how we handle
> core-updates.

We don't rebase 'live' branches.  The best way is to merge master, or
you could cherry-pick the patches.  I can merge it later if you're not
comfortable doing it :-)