In case this isn't obvious from the last sentence, setting m33 or m44 to 0 or -0 does not result in is2D being set to false but it should.

I tested Firefox and it indeed outputs true, while Chrome outputs false.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/ac6d4866aa926c3edde2154eb7c207a21d19aae9

commit ac6d4866aa926c3edde2154eb7c207a21d19aae9
Author: Fredrik Söderquist <fs@opera.com>
Date: Tue Aug 22 05:49:33 2017

Setting DOMMatrix.m33 or m44 to 1 should preserve is2D

Per [1], setting a value other than one (1) should clear the is2D flag.
We had the logic reversed.

[1] https://drafts.fxtf.org/geometry/#dommatrix-attributes

Bug:  756789 
Change-Id: I37fcd4e20fedee6ba29bb164e81cdf324971f9a1
Reviewed-on: https://chromium-review.googlesource.com/623410
Commit-Queue: Dominic Cooney <dominicc@chromium.org>
Reviewed-by: Dominic Cooney <dominicc@chromium.org>
Cr-Commit-Position: refs/heads/master@{#496224}
[add] https://crrev.com/ac6d4866aa926c3edde2154eb7c207a21d19aae9/third_party/WebKit/LayoutTests/external/wpt/css/geometry-1/DOMMatrix-attributes.html
[modify] https://crrev.com/ac6d4866aa926c3edde2154eb7c207a21d19aae9/third_party/WebKit/Source/core/geometry/DOMMatrix.h

Is this fix going to make its way into Chrome 61 before release?  Is there anything I should do to increase the probability?

Based on prio and known size of affected it doesn't seem like an obvious candidate for merge (IMHO.) The risk should be low however.

That would be a shame as it causes our app to have severe rendering issues that would require a very messy workaround to avoid.

So, this is not a Pri-3 issue?

#9 - I do not advocate it or anything (and I think the fix, since it is very small, should be merged), but would a small polyfill-like code at the startup of your application not do the trick?
Something like -
(function ()
 {
  var m = new DOMMatrix;
  m.m44 = 1;
  if (m.is2D)
  {
   return;
  }
  // Work around the wrong is2D.
  const originalIs2D = Object.getOwnPropertyDescriptor(DOMMatrix.prototype.is2D);
  Object.defineProperty(
   DOMMatrix,
   "is2D",
   {
    ...originalIs2D,
    get()
    {
     return !originalIs2D.get.call(this);
    }
   });
 )());

This bug requires manual review: We are only 5 days from stable.
Please contact the milestone owner if you have questions.
Owners: amineer@(Android), cmasso@(iOS), ketakid@(ChromeOS), govind@(Desktop)

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

#11 Interesting idea but I don't see how that can work.  is2D is only wrong if you set m33 or m44.  With your get() a newly created DOMMatrix would give false for is2D and any DOMMatrix that should be 3d would give true.  Replacing the toString function also sounds like a reasonable approach but, since we do need both 2d and 3d to work, I don't think that would be possible via a polyfill approach either (except, maybe, by replacing more/all of DOMMatrix, if possible).  The messy workaround I alluded to involves adding a function that, for affected browsers, always creates a "matrix(...)" string from the 2d elements of the matrix even if is2D says false.  This could then be called from those places that absolutely require a 2d "matrix(...)" but this will involve checking and possibly tweaking quite a few bits of code.

#14 - sorry for the incomplete workaround - I do not know DOMMatrix that well, I was merely suggesting a general way of overriding that function with the right implementation when it is wrong.
Replacing DOMMatrix altogether with a full blown polyfill when the feature-without-bug detection fails sounds more appropriate, really.

Anyway, we will see what the release owners say, hopefully - they will approve the merge and this will be moot. :)

I would argue for merging. Small user group also means low chance of negatively affecting people, and right now we are clearly totally wrong. The patch is extremely local.

Approving merge to M61 branch 3163 based on comment #16 and per offline chat with  schenney@. Please merge ASAP. Thank you.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/526c288432fb14dbeb11bb01cb5dce36cbc737b4

commit 526c288432fb14dbeb11bb01cb5dce36cbc737b4
Author: Fredrik Söderqvist <fs@opera.com>
Date: Wed Aug 30 20:15:41 2017

Setting DOMMatrix.m33 or m44 to 1 should preserve is2D

Per [1], setting a value other than one (1) should clear the is2D flag.
We had the logic reversed.

[1] https://drafts.fxtf.org/geometry/#dommatrix-attributes

TBR=fs@opera.com

(cherry picked from commit ac6d4866aa926c3edde2154eb7c207a21d19aae9)

Bug:  756789 
Change-Id: I414985018282a98bb7edea67fc1fa060bd611ba8
Reviewed-on: https://chromium-review.googlesource.com/623410
Commit-Queue: Dominic Cooney <dominicc@chromium.org>
Reviewed-by: Dominic Cooney <dominicc@chromium.org>
Cr-Original-Commit-Position: refs/heads/master@{#496224}
Reviewed-on: https://chromium-review.googlesource.com/644326
Reviewed-by: Fredrik Söderquist <fs@opera.com>
Cr-Commit-Position: refs/branch-heads/3163@{#1004}
Cr-Branched-From: ff259bab28b35d242e10186cd63af7ed404fae0d-refs/heads/master@{#488528}
[add] https://crrev.com/526c288432fb14dbeb11bb01cb5dce36cbc737b4/third_party/WebKit/LayoutTests/external/wpt/css/geometry-1/DOMMatrix-attributes.html
[modify] https://crrev.com/526c288432fb14dbeb11bb01cb5dce36cbc737b4/third_party/WebKit/Source/core/geometry/DOMMatrix.h

Thank you, very much appreciated...

Tested the issue using #61.0.3163.73 on Mac 10.12.6 and Win 10 as per the steps mentioned in comment #0. Observed now the output is true. 

Please find the screen cast for the same. Hence adding verified labels.

Thanks!!