Here's what Phlogiston sees today:

phlogiston=# select id, status, points from task_on_date_recategorized where category like 'NewWiki%' and date = '2016-09-20' o
rder by id;                                                                                                                    
   id   |   status   | points 
--------+------------+--------
 141068 | "open"     |      8
 141149 | "open"     |      8
 142138 | "resolved" |     40
 142642 | "resolved" |      8
 142643 | "open"     |      8
 142644 | "resolved" |      1
 142645 | "open"     |      8
 142646 | "resolved" |      1
 142854 | "resolved" |      1
 142861 | "open"     |      8
 143577 | "resolved" |      1
 143578 | "open"     |      1
 143582 | "resolved" |      1
 143917 | "resolved" |      1
 143920 | "resolved" |      1
 143954 | "resolved" |      1
 144323 | "open"     |      8
 144621 | "resolved" |      1
 144653 | "open"     |      1
 144654 | "open"     |     40
 145417 | "open"     |      8
(21 rows)

With the new Task Graph, it's possible to compare that directly to Phabricator. Results:

1. All of the stories mentioned in the original Description of this bug appear to be correct in Phlogiston.

2. I manually compared the two lists. They match except for:

2a) T141069 and T143709. I think they got caught by the External rule, which maybe should move down in the rules list since it's less specific. Probably by default all Parent rules should go first, then Column, then Intersection, then Project by Name or ID or Wildcard. (suggesting this as a tentative best practice for manually setting up the rules, not something to automate in Phlog).

2b) T144651 is a TR0 bug, so is in that category. Which is more useful, to see that in TR0 or here?

2c) T146182, just filed today.

So I think CNR?

I think External should trump the tree, actually (because work done by others should be counted as such, not against the milestone it blocks).

OK, declaring this done.