Re: [Qemu-devel] [PATCH 00/28] Memory API for 1.6: fix I/O port endianness mess

[Anthony Liguori]

Paolo Bonzini <address@hidden> writes:

> Il 25/07/2013 12:23, Benjamin Herrenschmidt ha scritto:
>> On Thu, 2013-07-25 at 10:38 +0100, Peter Maydell wrote:
>>> On 25 July 2013 10:00, Benjamin Herrenschmidt <address@hidden> wrote:
>>>> That's fine, I know you can fix stuff :-) I'm just really annoyed that
>>>> upstream qemu remained broken for so long (and still is) while the whole
>>>> thing derailed into a mostly pointless discussion on endianness and
>>>> nobody (including Alexey) hollered loud enough that the breakage was
>>>> fairly extensive
>>>
>>> I think that for the minor architectures we just have to make
>>> sure that we do yell loudly when things are broken, because
>>> the nature of things is that people won't notice. Maybe we should
>>> have a qemu-urgent list to parallel qemu-trivial for compile
>>> fixes, reversions of bad breakage, etc, to try to keep them
>>> out of the general noise ?
>> 
>> Or I teach Alexey to yell louder (along with some French :-)
>> 
>> Not sure if a mailing list is useful. A tag might be enough
>> [REGRESSION] ? [URGENT] ? Those kind of subject tags tend to stand out
>> pretty well on mailing lists.
>
> I've definitely seen them already in the past on qemu-devel, so I concur
> that a mailing list is not useful.

This is a technical problem, not a social one.  One something is broken,
it's a hell of a lot harder to fix than it is to prevent something from
breaking in the first place.

If you want to prevent minor architecture regressions, add unit tests.

Regards,

Anthony Liguori

>
> Paolo