As bug #1477199 describes, the wrong comment below is all over the
unit test code of OSC.
# DisplayCommandBase.take_action() returns two tuples
There is no such class named DisplayCommandBase in OSC. It is in cliff.
All OSC command classes inherit from the base classes in cliff,
class Command, class Lister and class ShowOne. It is like this:
Object
|--> Command
|--> DisplayCommandBase
|--> Lister
|--> ShowOne
take_action() is an abstract method of class Command, and generally is
overwritten by subclasses.
* Command.take_action() returns nothing.
* Lister.take_action() returns a tuple which contains a tuple of columns
and a generator used to generate the data.
* ShowOne.take_action() returns an iterator which contains a tuple of
columns and a tuple of data.
So, this problem should be fixed in 3 steps:
1. Remove all DisplayCommandBase comments for tests of classes inheriting
from class Command in cliff as it returns nothing.
2. Fix all DisplayCommandBase comments for tests of classes inheriting
from class Lister in cliff. Lister.take_action() returns a tuple and
a generator.
3. Fix all DisplayCommandBase comments for tests of classes inheriting
from class ShowOne in cliff. ShowOne.take_action() returns two tuples.
This patch finishes step 3 in all but identity tests. There are too many
such comments in identity tests. So fix them all in another patch.
Change-Id: I1afe4852069d25d562a9448ec2bf2cff58955052
Partial-bug: #1477199
As bug #1477199 describes, the wrong comment below is all over the
unit test code of OSC.
# DisplayCommandBase.take_action() returns two tuples
There is no such class named DisplayCommandBase in OSC. It is in cliff.
All OSC command classes inherit from the base classes in cliff,
class Command, class Lister and class ShowOne. It is like this:
Object
|--> Command
|--> DisplayCommandBase
|--> Lister
|--> ShowOne
take_action() is an abstract method of class Command, and generally is
overwritten by subclasses.
* Command.take_action() returns nothing.
* Lister.take_action() returns a tuple which contains a tuple of columns
and a generator used to generate the data.
* ShowOne.take_action() returns an iterator which contains a tuple of
columns and a tuple of data.
So, this problem should be fixed in 3 steps:
1. Remove all DisplayCommandBase comments for tests of classes inheriting
from class Command in cliff as it returns nothing.
2. Fix all DisplayCommandBase comments for tests of classes inheriting
from class Lister in cliff. Lister.take_action() returns a tuple and
a generator.
3. Fix all DisplayCommandBase comments for tests of classes inheriting
from class ShowOne in cliff. ShowOne.take_action() returns two tuples.
This patch finishes step 2 in all but identity tests. There are too many
such comments in identity tests. So fix them all in another patch.
Change-Id: I00f38d12f55abe20fa708f6349073da658622f8d
Partial-bug: #1477199
As bug #1477199 describes, the wrong comment below is all over the
unit test code of OSC.
# DisplayCommandBase.take_action() returns two tuples
There is no such class named DisplayCommandBase in OSC. It is in cliff.
All OSC command classes inherit from the base classes in cliff,
class Command, class Lister and class ShowOne. It is like this:
Object
|--> Command
|--> DisplayCommandBase
|--> Lister
|--> ShowOne
take_action() is an abstract method of class Command, and generally is
overwritten by subclasses.
* Command.take_action() returns nothing.
* Lister.take_action() returns a tuple which contains a tuple of columns
and a generator used to generate the data.
* ShowOne.take_action() returns an iterator which contains a tuple of
columns and a tuple of data.
So, this problem should be fixed in 3 steps:
1. Remove all DisplayCommandBase comments for tests of classes inheriting
from class Command in cliff as it returns nothing.
2. Fix all DisplayCommandBase comments for tests of classes inheriting
from class Lister in cliff. Lister.take_action() returns a tuple and
a generator.
3. Fix all DisplayCommandBase comments for tests of classes inheriting
from class ShowOne in cliff. ShowOne.take_action() returns two tuples.
This patch finishes step 1 in all but identity tests. There are too many
such comments in identity tests. So fix them all in another patch.
Change-Id: I9849baa8141ea8af2042a69afd540b77ce6ae6bd
Partial-bug: #1477199
Users could specify the last image (name or ID) of the previous page with
--marker option to control the start image of the output.
Change-Id: Idca0235ee83b1226b00c89cf3d38500fa898b7d0
Closes-Bug: #1540988
This option is quite useful if there are too many images.
Change-Id: If6a901c27c5da2d1f4412e8fa9ba3bed3b72fdd9
Co-Authored-By: Tang Chen <chen.tang@easystack.cn>
Partial-Bug: #1540988
Since --page-size has never worked, there is no paginate logic needs
to be implemented in "image list" command. So remove the unnecessary
loop.
And also, the marker is not necessary because --marker option has not
been implemented. Will add it back when implementing --marker option.
Change-Id: I71fea1502f92f447a49697edb52e8e82f336772f
Partial-Bug: #1540988
test cases
columns and datalist has been set in each test case in image
and object which is not necessary. This patch abstract it out
and remove all redundant code.
Change-Id: Ie6aa3fa27ab2a468c67da31209107517259631c2
Related-Bug: 1532384
* image create and image set now use --project to specify an alternate
project to own the image
* --owner is still silently accepted but deprecated, add warning messages
* --project and --owner are mutually exclusive to prevent precedence issues
Closes Bug: 1527833
Change-Id: Iccb1a1d9175ef9b5edcd79d294607db12641c1f0
This change allows admins to deactivate and reactivate their
images. Currently this has to be done with the REST api or the
glanceclient.
This change introduces `--deactivate` and `--activate` for the `image
set` command.
This requires glanceclient 1.2.0. Which got bumped here:
https://review.openstack.org/#/c/257512/
Change-Id: I476c44a0343cdc92d58ddc93fb06470242de2345
Depends-On: I2c370c6bf6ff664d94d756cc76aaa983fbdb8869
Closes-Bug: 1516661
Introduce FakeImage to improve the current image unittest framework
with following two advantages:
1. generate more than one faking images
2. all faking images generated by random
Change-Id: Ide326fa2a047ddeea478bef97000083617a0b744
Implements: blueprint improve-image-unittest-framework
Actually, the status column is useful for the user.
So, it`s better to output this info by default (or, user had
to specify the extra option, i.e. --long).
Change-Id: Id2a9f86f0de5310f8f5ff9a46bf1b7411094b519
Closes-Bug: #1519181
This adds --owner to `image create`. This is backwards compatable with v1.
Change-Id: I9e79cf880c91a1386419db729818d23dfe632179
Depends-On: I8d572a070bbb04dccdd051b8e0ad199c5754746e
image_member doesn't have a _info attribute, glanceclient returns
warlock object instead of a Resource object.
Change-Id: If6e7c4bd404454bd6cbe8c111879c1afa1380211
Closes-Bug: #1509054
the rest of OSC set commands do not show the resource after it has
been updated. unless the update fails then we report back a failure,
otherwise the user should assume everything went fine.
Change-Id: I2bd4188450c3853b4a1bc25f80fc9450cda32bdd
This adds --tag to the v2 version of `image set`. This
is another step to compatability between the osc image api.
Added merge of tags into existing tags and handling duplicates,
and tests for same.
Co-Authored-By: Steve Martinelli <stevemar@ca.ibm.com>
Change-Id: Ie800fcbf8bbc0978c54ace3278750a18023e8ce4
Make the Image v2 image set command meet at the intersection of the v1
image set command and the v2 image create command:
* Add visibility to the deadopts list and remove the option
* Put the options in the same order as v1 image set
* Make the help text match
* Add --properties
* Move the additional options that do not appear in either v1 image set or
v2 image create after --property as they are really pre-defined properties
* Add tests for v2 image set to match v1 and then some
* Put the SetImage class in v2/image.py in alphabetical order
Change-Id: I102b914e8ad09a014f6fdd846c5766b6c2eaadb8
Currently `image set` uses the new api, where other parts of osc the old
api is used. This deprecates the v2 api in favour of the v1 to maintain
the same commands across osc. However the functionality now remains
there as people could now be using this functionality.
This also adds the --unprotected argument, as in the previous version if
--protected was not supplied it would just make the argument
--unprotected without the users explicit consent.
The patch also fixes the documentation for image set as it was outdated.
Change-Id: I990d20332c80165102badef7ac94ddbeb7824950
Closes-Bug: 1498092
We have it for v1, but v2 is the future. There are two differences,
things in v2 do not go into a properties dict, and the actual image
data needs to get uploaded as a second step.
Closes-Bug: 1405562
Co-Authored-By: Niall Bunting <niall.bunting@hp.com>
Co-Authored-By: Sean Perry <sean.perry@hp.com>
Change-Id: If7b81c4a6746c8a1eb0302c96e045fb0f457d67b
Openstack image create command updates existing image (with same name) by
default. That might be confusing since glance allows create multiple
images with same names and may lead to unwanted image update by image
create command.
Image update code was moved from image create action to image set action.
BackwardsIncompatibleImpact
Change-Id: I1686c6544c366262efab9e33c066d5f8a667f707
Closes-Bug: #1461817
This patch includes functional tests for image set and it includes
a change to use the OSC utils.format_dict method to format the
properties. This will give a more user friendly format to the image
commands and it gives a more consistent testable format to the
output. Instead of:
{u'a': u'b', u'c': u'd'}
The user will see:
a=b, c=d
Change-Id: Ib396316586ffc5dbab231064d5b6dc9425507934
Enable user to update the following image properties from OSC:
container-format, disk-format, size
Closes-Bug: #1446362
Change-Id: Id9f40f15702e8f14f0327a37fcc7d7971338c258
Image v2 uses warlock objects rather than the usua Resource objects
so we need to test for those. This adds a subset of the Image v2
schema that should be enough to test for proper warlock image handling.
Depends-On: Ic95db2f63d9f5f37e29f0d7e048397da311fbf8c
Change-Id: Ib89cce87f110a554f40e726718e31d39b500a6ae
Handle the paginatiion for image list. We were sorting the
data here, so nothing lost for the generator.
Change-Id: I2d7d4b3d5c9f650953f309c971ac53b64f6f7f77
Add sort support to image list by sorting items in the client side.
The parameter syntax follows this spec[1].
[1] https://review.openstack.org/#/c/145544/
Change-Id: I42b487d18f00f937db1938daa46487cea2a896ab
Closes-Bug: #1410251
* Hides previously broken --page-size option
* Adds --property to image list for filtering on properties
* Adds Visibility, Protected, Owner, Properties/Tags to --long output
* Adds api.utils.simple_filter() for selecting matches out of a list
of objects
* Adds tests for all of the above
* Updates image docs
There are additional filtering options to be added in later reviews.
Change-Id: I32feff0ad61aae749b33621c817658d7dc90c3aa
Closes-bug: 1401902
image list for v1 and v2:
* Add --public|--private to command parsers
* Implement local public/private filtering for v1 image_list()
* Pass public/private filter to server for v2 image_list()
Change-Id: Ie7c24ea2d1bf2b3b1b7fa342eb45fee45894634d
This is part1, add support for these objects:
compute.server
imagev1.image
imagev2.image
network.network
volume.volume
volume.backup
volume.snapshot
Closes-Bug: #1400597
Change-Id: Ice21fee85203a8a55417e0ead8b509b8fd6705c1
The current tests for image do not run v2 image code, changing
that portion also made it's only test fail.
I opted to change the image delete code and not the test, since
passing the object ID is more in line with the rest of the
project code.
Change-Id: I62e13c063a5d68279dbbf31e59266db6285d73bf
Add disk_format, container_format, size and status to image list
command. Added tests as well.
Change-Id: I8e3822c6d46d0020fc706955c026549f6c635587
Closes-Bug: #1348475
Refactor image create and set commands to properly handle properties.
This is particularly tricky with exclusive booleans as in this case leaving
both choices off the command line should NOT assume a default value but
leave the existing value unchanged.
Properties were not being updated correctly in the 'image set' command.
Refactor it to use the same pattern as in other SetXxx commands.
Add tests for arg handling.
Change-Id: I123a64c9b4feecab25a3e2013cc047f55b1c9967
Add ability to create an image from a volume.
* Added --volume command to image create
* Added --force option to image create
* Added block to access volume manager in image create
* Tests added for the volume option
Change-Id: I3910a2b5e04acd0d15dd230747ba6ebca07aa316
Closes-Bug: #1207615
