It seems we always start counting from the start of the explicit grid when there is no <custom-ident> involved.

Comment on attachment 8675670 [details] [diff] [review] tests Review of attachment 8675670 [details] [diff] [review]: ----------------------------------------------------------------- Drive-by nit on test files -- please fix end-of-line whitespace in both the testcase & reference. Samples below, for the reference case. (and the testcase has whitespace in the same spots) ::: layout/reftests/css-grid/grid-placement-negative-lines-001-ref.html @@ +8,5 @@ > + <title>Reference: Placement involving negative line numbers</title> > + <link rel="author" title="Mats Palmgren" href="https://bugzilla.mozilla.org/show_bug.cgi?id=1215957"> > + <style type="text/css"> > +body,html { color:black; background:white; font-size:12px; padding:0; margin:0; } > + Nit: whitespace on blank line here @@ +52,5 @@ > +<x style="grid-row:4; grid-column-start:6">A -1</x> > +<x style="grid-row:4; grid-column-start:4">B -1</x> > +<x style="grid-row:5; grid-column-start:4">A -2</x> > +<x style="grid-row:6; grid-column-start:3">A -3</x> > +<x style="grid-row:6; grid-column-start:2">A -4</x> ...and end-of-line whitespace on these "A -1" etc. lines (and on other occurrences of "A -1")

I fixed the test nits. https://hg.mozilla.org/integration/mozilla-inbound/rev/bb0e04a7ab19 https://hg.mozilla.org/integration/mozilla-inbound/rev/d4e658173847

https://hg.mozilla.org/mozilla-central/rev/bb0e04a7ab19 https://hg.mozilla.org/mozilla-central/rev/d4e658173847

https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/bb0e04a7ab19 https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/d4e658173847

(In reply to Carsten Book [:Tomcat] from comment #6) > https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/bb0e04a7ab19 > https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/d4e658173847 backed this out on request from mats from b2g v2.5 in http://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/0869ace8c965