On Fri, Apr 4, 2014 at 8:38 PM, Dylan Baker <[email protected]> wrote:
> On Friday, April 04, 2014 18:57:12 Ilia Mirkin wrote:
>> On Fri, Apr 4, 2014 at 6:39 PM, Dylan Baker <[email protected]>
>> wrote:
>> > diff --git a/framework/summary.py b/framework/summary.py
>> > index 6d2a7bb..51923ad 100644
>> > --- a/framework/summary.py
>> > +++ b/framework/summary.py
>> >
>> > @@ -423,7 +423,8 @@ class Summary:
>> > info=value.get('info', 'None'),
>> > traceback=value.get('traceback', 'None'),
>> > command=value.get('command', 'None'),
>> >
>> > - dmesg=dmesg,
>> > + images=value.get('images', None),
>> > + dmesg=value.get('dmesg', None),
>>
>> dmesg was messed up for a time, the computed dmesg value normalizes
>> runs that were made during the "bad" timeperiod.
>
> Huh?
You replaced
dmesg=dmesg
with
dmesg=value.get('dmesg', None)
The variable dmesg is defined as
dmesg = value.get('dmesg', 'None')
if isinstance(dmesg, list):
dmesg = "\n".join(dmesg)
Which is necessary to deal with the short period of time when dmesg
was stored as an array of strings in the output (when you initially
redid the dmesg impl), which could have been saved (in fact, I'm
fairly sure some of the results files at
http://people.freedesktop.org/~imirkin/piglit/ are affected). Please
undo that so that I can keep using piglit-summary-html on the saved
results files and not end up with stupid-looking dmesg.
Hope that clears things up.
-ilia
_______________________________________________
Piglit mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/piglit
