Re: Bug#154829: Is a bug "grave" if the package is unusable for at least two architectures (Was: Bug#154829: Aido problems on 64 bit architectures)
On Thu, Jun 12, 2003 at 04:06:36PM -0700, Mike Markley wrote:
> On Fri, Jun 13, 2003 at 12:08:49AM +0200, Wouter Verhelst <wouter@grep.be> wrote:
> > I agree with Andreas here. I feel that 'major impact, does not render
> > unusable to everyone' should be interpreted as 'the bug makes things go
> > all whacky, making it hard to do useful stuff with the program, although
> > it's probably still useful to some users in its current state, with the
> > bug'. Not something such as "if there's any user on this planet who can
> > still use the program, it's not a grave bug", since in that case, grave
> > bugs would be reserved for packages that segfault at startup on *all*
> > architectures...
>
> My logic is that it segfaults on startup on very few architectures,
> unless someone has information to the contrary (so far I've heard about
> two). I don't contend that, say, one user being able to use it means
> that it's not grave; I do contend that *most* users can use it.
Your logic is flawed. *Most* users are most certainly i386.
By saying that it does work for *most* users, then a package
for Architecture: any that segfaults on all archs except i386,
could not have grave bugs against it...
Futhermore, what's the point of having support for Arch X if program
segfaults on it. The bug should either be fixed or supported archs
specifically listed. Grave bugs are bugs that prevent a package
from working at all -- segfaults on startup most definately qualify.
[segfaults that occur when user tries to access some non-critical
feature of the program, would be important IMHO].
Just imagine if binutils was accepted to stable even if
it segfaults on hppa or something but works fine on i386 and others. :)
- Adam
Reply to: