When running the mochitests in content/base/test, there are a lot of assertions of the form: ###!!! ASSERTION: We placed a float where there was no room!: 'psd->mX - mTrimmableWidth <= psd->mRightEdge', file /home/dbaron/builds/bisect-mozilla-central/mozilla/layout/generic/nsLineLayout.cpp, line 345 These are a regression from the float manager landing (bug 191448). I tested this by observing that the assertions were not present with a build from http://hg.mozilla.org/mozilla-central/rev/b087a09f82aa but are present with a build from http://hg.mozilla.org/mozilla-central/rev/aa32889429db with the patches from http://hg.mozilla.org/mozilla-central/rev/63fc1585fe00 , http://hg.mozilla.org/mozilla-central/rev/c162278affad , and http://hg.mozilla.org/mozilla-central/rev/2a4999efa837 patched in.

Note that there are some comments about this bug in bug 439204.

I think they're mainly on 416317-2.html, but maybe some on 416317-1.html.

I think the difference is that in the old code, GetAvailableSpace conditioned the band width to be >= 0, whereas now we're ending up with negative available space widths.

Landed on 1.9.1 along with the patch that caused the regression: http://hg.mozilla.org/releases/mozilla-1.9.1/rev/110384dabb87