[Mesa-dev] AppVeyor: Build failed: mesa 39

Jose Fonseca jfonseca at vmware.com
Thu Dec 3 23:16:20 PST 2015

On 04/12/15 03:21, Roland Scheidegger wrote:
> Am 04.12.2015 um 03:58 schrieb AppVeyor:
>>    Build mesa 39 failed
>>    <https://ci.appveyor.com/project/jrfonseca-fdo/mesa/build/39>
>> Commit 51140f452a by Roland Scheidegger <mailto:sroland at vmware.com> on
>> 12/4/2015 2:42 AM:
>> draw: fix clipping of layer/vp index outputs\n\nThis was just plain
>> broken. It used always the value from v0 (for vp_index)\nbut would pass
>> the value from the provoking vertex to later stages - but only\nif there
>> was a corresponding fs input, otherwise the layer/vp index would
>> get\nlost completely (as it would try to interpolate the (unsigned)
>> values as\nfloats).\nSo, make it obey provoking vertex rules (drivers
>> relying on draw will need to\ndo the same). And make sure that the
>> default interpolation mode (when no\ncorresponding fs input is found)
>> for them is constant.\nAlso, change the code a bit so constant inputs
>> aren't interpolated then\ncopied over later.\n\nFixes the new piglit
>> test gl-layer-render-clipped.\n\nv2: more consistent whitespaces fixes
>> for function defs, and more tab killing\n(overall still not quite right
>> however).\n\nReviewed-by: Brian Paul <brianp at vmware.com>\nReviewed-by:
>> Jose Fonseca <jfonseca at vmware.com>
>> Configure your notification preferences
>> <https://ci.appveyor.com/notifications>
> "Failed to provision build worker virtual machine in Google Compute
> Engine cloud. Try restarting build later."
> I wonder if that's something happening frequently - if so these messages
> will get annoying really fast. Would it be possible to skip them if
> there's no log (thus indicating it wasn't really a code problem)?
> Roland

It never happened for me with Apitrace and other pet projects I have on 

But Mesa has a high commit frequency, so whatever rare problems there 
may, it will become slightly less rare.  Lets wait and see, if this 
happens too frequently I'll try to figure a way to reduce the spam.


More information about the mesa-dev mailing list