Regression with cliff 2.10.0
Bug #1741911 reported by
Julien Danjou
This bug affects 2 people
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Magnum |
Fix Released
|
Undecided
|
yatin | ||
cliff |
New
|
Undecided
|
Unassigned |
Bug Description
commit c6fc228a250aeb2
Author: Doug Hellmann <email address hidden>
Date: Wed Nov 22 15:09:18 2017 -0500
add support for legacy command name translation
Add add_legacy_
command search logic to try translating the old legacy name to the new
name.
Change-Id: I0c0cdbfcb1612e
Signed-off-by: Doug Hellmann <email address hidden>
broke https:/
$ pifpaf run mysql -- zsh
pifpaf: 'run mysql -- zsh' is not a pifpaf command. See 'pifpaf --help'.
Did you mean one of these?
help
Changed in magnum: | |
assignee: | nobody → yatin (yatinkarel) |
status: | New → In Progress |
To post a comment you must log in.
This is currently breaking the Magnum gate, see [1]. I've been looking at the code, and what I've found is:
Magnum inits DriverCommandMa nager using None as a namespace parameter [2]. With the cliff change, the _load_commands() method skips any loading, see [3], so this is causing the unit test error, which btw is also seen in the upstream gate.
I'm not sure if setting namespace as None is valid, so this could be fixed on the Magnum side or on the cliff side (if it's expected to be valid). Apparently, using None as a namespace is relatively widespread [4].
[1] -http:// logs.openstack. org/30/ 529330/ 2/check/ openstack- tox-py27/ 2e898ff/ job-output. txt.gz# _2018-01- 10_16_13_ 06_384428 /github. com/openstack/ magnum/ blob/master/ magnum/ cmd/driver_ manage. py#L87- L93 /github. com/openstack/ cliff/blob/ 2.10.0/ cliff/commandma nager.py# L55-L56 codesearch. openstack. org/?q= CommandManager% 5C(None& i=nope& files=& repos=
[2] - https:/
[3] - https:/
[4] - http://