Correct, we are building on a CentOS 7.3 virtual machine.
We eventually discovered (I don't remember how) that the package builds if
you copy it to /tmp and run rpmbuild with sudo. The copy to /tmp was to get
around our root-squash which prevented us from doing it in place.
I can reproduce this reliably using our CentOS image (I have not tried a
base from-ISO install), so let me know if there are any tests you'd like me
to perform.
Doug Penner
IT - 3vGeomatics
On Wed, Aug 29, 2018 at 8:13 AM, Ed J via RT <
bug-ExtUtils-MakeMaker@rt.cpan.org> wrote:
Show quoted text> <URL:
https://rt.cpan.org/Ticket/Display.html?id=126039 >
>
> On Thu Aug 23 16:32:26 2018, dpenner@3vgeomatics.com wrote:
> > This looks like the same error in ticket #102706
> >
> > Perhaps it came back?
>
> Hi Doug,
>
> The test in question sets the destination file to be read-only, then tries
> to write to it, to check that Mkbootstrap() will die. This does not happen
> on Cygwin, so the fix for the error in #102706 was to not even try that on
> Cygwin.
>
> I gather you're building an RPM on a CentOS box (ie not Cygwin). Is the FS
> it's using read-only?
>
> What do other Toolchain folks think about making this test do its thing in
> a tempdir()? That looks more idiomatic than relying on being able to write
> to files in t.
>
RT for rt.cpan.org