On 05/11/2017 01:33 PM, Adam Williamson wrote:
On Thu, 2017-05-11 at 17:33 +0100, Peter Robinson wrote:
>> For this particular Firefox example, what is the core problem that you're
>> trying to fix here? Is it the fact that Firefox excluded many arches from
>> builds? From my QA POV, since it excluded arm, it's a blocker, since arm is
> Well it causes most of the world to fail to compose on numerous
> architectures. EG Workstation on both ARMv7 and aarch64 is very usable
> and accelerated on a number of devices but now completely fails to
> compose, Firefox is the default browser there.
The point is that as the criteria stand, we could fix that without
fixing Firefox, by simply changing the default browser on the image. We
don't *technically* have to make Firefox build again in order to
resolve the blocker issue.
Some people have an instinctive feeling that this is 'bad' and the
criteria must somehow require that we fix Firefox. Stephen is proposing
something less than that; that at least we should never vary the
'default browser' (and other 'applications') between arches.
Well,
I was specifically trying to avoid making a rule that applied to
Firefox specifically (because as I mentioned, there's nothing stopping
us from deciding that Chromium or Epiphany or even elinks should be the
default on all systems). My reasoning was that this would be sort of a
back-door to forcing this to be fixed, since the SIGs would apply
pressure not to be forced to switch their other arches to move away from
a package just because it doesn't build (currently) on one of them.
I'm open to a better approach here, but I want to try to encode into the
criteria somewhere that the same deliverable should have the same
default behaviors on all arches for which it is built.