[Mesa-dev] [PATCH] travis: Initial import of travis instructions.

[Michel Dänzer]

On 03.12.2015 06:01, Jose Fonseca wrote:
> On 02/12/15 03:39, Michel Dänzer wrote:
>> On 02.12.2015 07:06, Jose Fonseca wrote:
>>> On 28/11/15 21:06, Emil Velikov wrote:
>>>> On 25 November 2015 at 07:20, Jose Fonseca <jfonseca at vmware.com> wrote:
>>>>
>>>>> BTW, I setup Mesa with Appveyor (like Travis for Windows)
>>>>>
>>>>>     https://ci.appveyor.com/project/jrfonseca/mesa
>>>>>
>>>>> I'll try to get that going and commited too.
>>>>>
>>>> As a person who has broken the Windows build on an occasion or two,
>>>> yes please.
>>>>
>>>>> The nice thing about Appveyor is that it can clone the git from
>>>>> anywhere,
>>>>> not just GitHub.
>>>>>
>>>>>
>>>>> Would it be OK to have email notifications to mesa-commits?
>>>>>
>>>> I'm wondering if mesa-dev wouldn't be more suitable. Not many of us
>>>> (or is it just me?) don't follow mesa-commits.
>>>
>>> I'm OK either way.
>>>
>>>
>>> Whichever list we choose, we need to whitelist emails from
>>> no-reply at appveyor.com somehow.   I don't know the procedure for that.
>>
>> A list administrator can add a subscription for that address with
>> delivery of list posts disabled.
> 
> Great.
> 
> Per http://lists.freedesktop.org/mailman/listinfo/mesa-dev it looks
> you're one of the admins!  Could you please add no-reply at appveyor.com?

Done.


>>>> One thing no one mentioned is that one can automatically feed the
>>>> results from Travis-CI/github into Coverity. I know Vinson has been
>>>> doing the latter, although I'm not sure how much of it is automated.
>>>
>>> Yes, but Coverity recommend (ie ask) not to do so on every commit.
>>>
>>> Clang and MSVC have static analyzers too.  I've been running them.  But
>>> to be honest, running them all the time is not enough -- it takes effort
>>> to actually act on issues, resolving them / whitelisting them.  And I'm
>>> afraid I don't had the time.  The bulk is false negatives.
>>
>> FWIW, I think you mean false positives here.
> 
> I know it should be false _something_, but I always hesitate on which to
> use. :)

False positive means that the static analyzer reports a problem, but in
fact there is no problem. False negative means there is a problem, but
the static analyzer doesn't report it.


-- 
Earthling Michel Dänzer               |               http://www.amd.com
Libre software enthusiast             |             Mesa and X developer