MYSQL/BZR P3: "bzr log -rX..Y" should fail if X is not an ancestor of Y
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Bazaar |
Fix Released
|
Medium
|
Vincent Ladeuil |
Bug Description
This is a Sun/MySQL - Canonical escalation imported into Launchpad by the Canonical Support Team (internal case 6611 )
Original escalation: 2009-11-16 10:12 UTC
Hello. Another report, after 474807 filed recently.
My initial intention here was too see whether bzr can be used to tell if a revision Y is a child of X; I expected that "bzr log -rX..Y" would fail if Y is not a child of X. So I did experiments, below.
I grab this branch: https:/
and I do
bzr log -n0 --show-ids -r tag:clone-
I expected an error, as mysql-5.1.39 is not a child of clone-5.2. I got no error, and the displaid log contains lots of revisions, including tag:clone-5.2 and tag:mysql-5.1.39. I wonder if "no error" is normal, I expected to trigger "bzr: ERROR: Start revision must be older than the end revision" (but when I look at the code which can emit such error, I see it compares revnos; with revnos being dotted numbers, I cannot see how they can reliably be compared).
Then I tried the reverse order:
bzr log -n0 --show-ids -r tag:mysql-
I expected an error, as clone-5.2 is not a child of mysql-5.1.39. I got no error, and this time the displaid log contains tag:clone-5.2 but NOT tag:mysql-5.1.39, which sounds weird (how come the start bound is not included?).
Bazaar (bzr) 2.1.0dev3
from bzr checkout /home/mysql_
revision: 4782
revid: <email address hidden>
branch nick: bzr.dev
Related branches
summary: |
- "bzr log -rX..Y" gives output which does not contain X + MYSQL/BZR P3: "bzr log -rX..Y" gives output which does not contain X |
description: | updated |
visibility: | public → private |
summary: |
- MYSQL/BZR P3: "bzr log -rX..Y" gives output which does not contain X + MYSQL/BZR P3: "bzr log -rX..Y" should fail if X is not an ancestor of Y |
Changed in bzr: | |
status: | New → Confirmed |
importance: | Undecided → Medium |
visibility: | private → public |
Changed in bzr: | |
milestone: | none → 2.2.0 |
Changed in bzr: | |
milestone: | 2.2.0 → 2.2.0b1 |
status: | In Progress → Fix Released |
We have a regression here as the order of the revisions wasn't the criteria to decide how they were presented.
--forward was needed to reverse the default order.
I'll look into it.