Bug#584606: nmu: insighttoolkit_3.18.0-2
On Sat, 2010-06-05 at 08:34 -0500, Steve M. Robbins wrote:
> On Sat, Jun 05, 2010 at 10:36:56AM +0100, Adam D. Barratt wrote:
> > If so then you need a give-back, not a binNMU,
> > and only on those architectures where the package previously FTBFS and
> > therefore /can't/ currently be binNMUed.
>
> OK. I don't know what the distinction between a "give-back" and a
> "binNMU" is, so I'll take your word for it :-)
A give-back implies retrying a failed build, whereas a binNMU is a
completely new build of a previously successfully built package (i.e. a
binary-only NMU).
> I originally had the same thought: only rebuild on those architectures
> that previously failed. But I decided to ask for all architectures to
> ensure that my gccxml fix didn't accidentally break them in other
> ways. If it's not too much trouble, can you get insighttoolkit
> rebuilt on ALL architectures, please?
I've given it back on all of the failing architectures.
Several of the architectures on which it originally succeeded do not
have the spare capacity to justify a rebuild "just to check" -
particularly not when the build takes over two days on at least one
architecture, and over a week on another.
As a hopefully acceptable compromise, I've binNMUed it on i386 and
kfreebsd-amd64. That would give attempts on two-thirds of the available
architectures and, assuming no other issues, a complete set of builds.
Regards,
Adam
Reply to: