This is a table layout bug. The table cell frame doesn't handle border and padding correctly in a paginated context. For example, it didn't apply skip sides (by using `ApplySkipSides`) for the [`borderPadding`](https://searchfox.org/mozilla-central/source/layout/tables/nsTableCellFrame.cpp#818-820). Also, it unconditionally subtracts block-axis border/padding from available block-size for its children [here](https://searchfox.org/mozilla-central/rev/064b0f9501ad76802853b43f18e33d8713fd54d3/layout/tables/nsTableCellFrame.cpp#825), and [add it back](https://searchfox.org/mozilla-central/rev/064b0f9501ad76802853b43f18e33d8713fd54d3/layout/tables/nsTableCellFrame.cpp#934) after the cell size is known. This can go wrong even if the skip sides are applied. For example, if the table cell needs a second page, its bottom border/padding needs to be skipped on the first page. Hiro, did you see other reftest failures other than table layout under jgraham's wpt printing machinery. If yes, I think it's worth filing such bugs. (NI myself to revisit this later.)
Bug 1626129 Comment 4 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
This is a table layout bug. The table cell frame doesn't handle border and padding correctly in a paginated context. For example, it didn't apply skip sides (by using `ApplySkipSides`) for the [`borderPadding`](https://searchfox.org/mozilla-central/source/layout/tables/nsTableCellFrame.cpp#818-820). Also, it unconditionally subtracts block-axis border/padding from available block-size for its children [here](https://searchfox.org/mozilla-central/rev/064b0f9501ad76802853b43f18e33d8713fd54d3/layout/tables/nsTableCellFrame.cpp#825), and [add it back](https://searchfox.org/mozilla-central/rev/064b0f9501ad76802853b43f18e33d8713fd54d3/layout/tables/nsTableCellFrame.cpp#934) after the cell size is known. This can go wrong even if the skip sides are applied. For example, if the table cell needs a second page, its bottom border/padding needs to be skipped on the first page. Hiro, did you see other reftest failures other than table layout under jgraham's wpt printing machinery? If yes, I think it's worth filing such bugs. (NI myself to revisit this later.)