1009627 - 2.4% svg-asap regression on windows 7 found May 10th from rev 02eb695c8640 on inbound (fx32)

Status: RESOLVED WONTFIX

(Core :: Graphics, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

Here is a graph:
http://graphs.mozilla.org/graph.html#tests=[[281,132,25],[281,131,25]]&sel=1399390072952,1399994872953&displayrange=7&datatype=running

I did some retriggers to help reduce any effects of the noise:
https://tbpl.mozilla.org/?tree=Mozilla-Inbound&fromchange=299a5a0a5458&tochange=81f84748793b&jobname=Windows%207%2032-bit%20mozilla-inbound%20talos%20svgr

And we have a winner:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?changeset=02eb695c8640

We had some tresize regressions in bug 1008965 for this same patch.  We also see a similar pattern of an improvement the day before and then a slight regression (this bug).  So the overall result is we have a net win- the question is- do we understand this regression and is was it expected.

If this is expected and we this is clean up from the earlier win, then lets resolve as wontfix- we will have this documented.

Comment 1

[Matt Woodrow (:mattwoodrow)]

I think this is the same as bug 1008965.

Neither are actually related to the changes from the previous day, they were likely wins from around a year ago when we switched to Moz2D for windows.

We did by regressing some correctness though, and bug 1008695 fixed that, taking us back to where we were.

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

my concern is we are introducing regressions in trunk (as fixes to other problems) on code that has already shipped.

Luckily in this case there was a bigger win on May 8th from bug 938395, so the net result is still a win.

Comment 3

[Joel Maher ( :jmaher ) (UTC -8)]

I noticed bug 997336 landed on aurora as well, have all our fixes made it there as well?