Bug#821805: kdevplatform: FTBFS in testing (ui_newclass.h: No such file or directory)
On Mon, Jun 06, 2016 at 11:10:03AM -0700, Lisandro Damián Nicanor Pérez Meyer wrote:
> > This would still be a bug, right? Otherwise we would have to create a
> > new control field for this, maybe "Build-CPUs: 2". I hope not!
>
> Yes, but I doubt it would be RC as it does not affects the official builds.
Feel free to downgrade to important if you wish.
I personally do not share the idea that a bug needs to happen in the
official buildds to be RC, and I'll try to elaborate:
Imagine a package which FTBFS with 1/2 of probability.
If the last build in the official autobuilders was successful, it just
means that you were lucky that time, it would not mean that the
package builds "fine".
This is of course a theoretical exercise, or a "mental experiment",
but I once did in fact find a package behaving like that:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=817033
I also found some funny case of a package which FTBFS when built under
stretch kernel (due to a bad interaction between the kernel and systemd).
Since autobuilders are still running jessie, the error would never
happen in an official autobuilder (yet). This is the bug if you are
curious, it made three different packages to FTBFS:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=823530
Back to our case, it could be that official autobuilders have more
than one CPU, but I would call that an accident, not something which
is "granted" or "promised". In other words, having two CPUs is
not build essential.
Thanks.
Reply to: