[igt-dev] ✓ Fi.CI.BAT: success for tests/kms_chamelium: add hdmi-audio test (rev2)

[Ser, Simon]

On Tue, 2019-05-14 at 12:02 +0300, Petri Latvala wrote:
> On Mon, May 13, 2019 at 05:17:53PM +0300, Ser, Simon wrote:
> > On Mon, 2019-05-13 at 15:55 +0300, Petri Latvala wrote:
> > > On Mon, May 13, 2019 at 03:09:04PM +0300, Arkadiusz Hiler wrote:
> > > > (kms_chamelium:2919) igt_chamelium-CRITICAL: Chamelium RPC call
> > > > failed: RPC failed at server.  <type
> > > > 'exceptions.Exception'>:method
> > > > "GetAudioFormat" is not supported
> > > > 
> > > > Welp, at least there are no regressions. I guess our Chameliums
> > > > need
> > > > to be updated...
> > > > 
> > > > Anyway, everything (sans the HAX) is:
> > > > Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler at intel.com>
> > > 
> > > Hold on, recording the conversation from the office to email:
> > > 
> > > These fails need to be skips instead.
> > 
> > Actually, calling HasAudioSupport will only work for DisplayPort
> > tests.
> > For HDMI tests, the old Chamelium devices support audio capture but
> > are
> > missing some API calls (GetAudioChannelMapping, GetAudioFormat).
> > 
> > These new functions fail if no audio capture has been started.
> > Thus,
> > we'd need to call them, expect them to fail, and check the error
> > message. It'd be a bit hacky.
> 
> Isn't it just comparing the fault code to
> XMLRPC_NO_SUCH_METHOD_ERROR?

Ohh, missed that, it's much better. Thanks!

> > Is it really that important?
> 
> Yes, failure should always mean it takes a kernel change to get it to
> pass.

Ack.