1511109 - moz-phab doesn't link up revisions correctly when I submit a second patch for a bug before the existing patch

Status: RESOLVED DUPLICATE of bug 1481539

(Conduit :: moz-phab, defect)

Description

[Daniel Holbert [:dholbert]]

Filing this for an issue I ran into with moz-phab for on bug 1337655.

tl;dr: I had a single patch there (submitted w/ moz-phab), and I realized I needed a second patch before it (not sure if ordering matters for the bug but this was my situation). So I created the second patch before the first (with a rebase), and submitted both with moz-phab, and it did not link them up into a "stack" (no depends-on relationship).

(I'm not 100% sure about these STR, but this is what happened as I recall it):
 1. create a commit for a given bug. Let's say its cset ID is $A
 2. moz-phab submit $A
 3. realize you need to make some changes before $A, so do that:
    hg up -r $parent-of-A
    [make changes and commit, which we'll call $B]
    hg evolve --all
  This rewrites $A as a child of $B. A now has a new commit hash, we'll call that $C.

 4. Submit your new patch stack:
    moz-phab submit $B:$C

ACTUAL RESULTS:
You end up with two phabricator revisions that aren't linked as a "stack" (no depends-on relationship).

EXPECTED RESULTS:
My phabricator revisions should've been linked to each other with depends-on.


For reference, my phabricator revision for $A (and eventually its rewrite $C) is here:
 https://phabricator.services.mozilla.com/D13298
and my phabricator revision for $B is here:
 https://phabricator.services.mozilla.com/D13370

Note that if I'd skipped step 2 and had only done a single moz-phab command at the very end, they would've been linked correctly.

Comment 1

[Daniel Holbert [:dholbert]]

in my case, I set the parent revision manually so that I could land the patch-stack. I did that here:
 https://phabricator.services.mozilla.com/D13298#330668

...and here's the earlier edit made by moz-phab that I think should've set the "parent revision" for me, but did not:
 https://phabricator.services.mozilla.com/D13298#328446

Comment 2

[Mark Côté [:mcote]]

Yup this is a known deficiency (because it's hard ;).  It's on our short list of moz-phab improvements.

Comment 3

[Daniel Holbert [:dholbert]]

OK, thanks!