"Derek Atkins" writes:
[snip]
> 106749 is marked a dupe_of 350675
> 350675 is marked a dupe of 570011
> 570011 blocks 570012
> 580012 is marked a dupe of 350675
>
> So 570011 is still open, so theoretically it should be "okay" and be in
> there "last". Maybe I need to swap the
As a follow-on to my last message, in those 185 messages that wont
linearlize, there are also circular dependencies. For example:
106749(350675) -> 350675(570011) -> 570011(570012) -> 570012(350675)
Most likely these are due to "depends_on" circular dependencies, not
necessarily duplicate-of