(Thanks :dholbert for helping me track this down) It all starts from [`ReflowOverflowContainerChildren`](https://searchfox.org/mozilla-central/rev/5f2c1701846a54c484d7dd46a291b796f5a67cac/layout/generic/nsContainerFrame.cpp#1168-1169) - Which is a special abs-position reflow case outside of the usual reflow. This feeds to [here](https://searchfox.org/mozilla-central/rev/5f2c1701846a54c484d7dd46a291b796f5a67cac/layout/generic/ReflowInput.cpp#2189-2202), where we try to calculate the containing block size manually by calling `ComputeContainingBlockRectangle`. In that function, I see a [TODO](https://searchfox.org/mozilla-central/rev/5f2c1701846a54c484d7dd46a291b796f5a67cac/layout/generic/ReflowInput.cpp#2099) for continuation-related workaround. This seems like a pretty good place to investigate. Bug 154892 links from that comment. Another one of interest, again thanks to :dholbert - Bug 546559. Given the low bug numbers, this is a longstanding issue (That may still interfere with use cases for anchor positioning mentioned in this ticket).
Bug 1927861 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.
(Thanks :dholbert for helping me track this down) It all starts from [`ReflowOverflowContainerChildren`](https://searchfox.org/mozilla-central/rev/5f2c1701846a54c484d7dd46a291b796f5a67cac/layout/generic/nsContainerFrame.cpp#1168-1169) - Which is a special abs-position reflow case outside of the usual reflow. Note that we pass `Nothing` for containing block. This feeds to [here](https://searchfox.org/mozilla-central/rev/5f2c1701846a54c484d7dd46a291b796f5a67cac/layout/generic/ReflowInput.cpp#2189-2202), where we try to calculate the containing block size manually by calling `ComputeContainingBlockRectangle`. In that function, I see a [TODO](https://searchfox.org/mozilla-central/rev/5f2c1701846a54c484d7dd46a291b796f5a67cac/layout/generic/ReflowInput.cpp#2099) for continuation-related workaround. This seems like a pretty good place to investigate. Bug 154892 links from that comment. Another one of interest, again thanks to :dholbert - Bug 546559. Given the low bug numbers, this is a longstanding issue (That may still interfere with use cases for anchor positioning mentioned in this ticket).