852d7b564d
virtinst: Remove name field validate_cb usage
...
Either openly do this at callers, or wrap it in a validate() call.
Doing this at set time has always had issues, like the self.replace hack
2018-09-04 14:55:04 -04:00
2ac54ac001
guest: Drop self.installer and start_install wrapper
...
This changes all the callers to invoke start_install directly on the
Installer object. We still stash the installer instance inside the
guest object in create.py, just for simplicity
2018-09-04 14:55:04 -04:00
c154bbacd4
guest: Fold add_default_devices into set_defaults
...
It doesn't have any need anymore to be separate
2018-09-04 14:55:04 -04:00
f1bbc5f67a
guest: move set_install_defaults to installer class
...
It's only called explicitly in one place, so update the caller
2018-09-04 14:55:04 -04:00
8a6a491306
guest: Drop self.autostart and self.replace
...
Require callers to pass autostart as an argument to start_install
self.replace isn't even used for guest installs
2018-09-04 14:55:04 -04:00
9e05c56397
guest: Drop self.domain
...
The only user is virt-install, make its usage explicit
2018-09-04 14:55:04 -04:00
fdbb1f9e5a
installer: Absorb all guest start_install handling
...
Really it makes more conceptual sense that the Installer class
should handle start_install and friends, but that's difficult
because the layering is all mixed up.
Start to untangle it by moving all those routines to the Installer
class and adding wrapper APIs to Guest to maintain compat for now.
2018-09-04 14:55:04 -04:00
1abbdfc551
installer: Drop isinstall parameter
...
Separate the install from postinstall config case more clearly,
so the installer is only ever altering bits for the install-time
case, and Guest handles actually writing postinstall bits
2018-09-04 14:55:04 -04:00
068c8aedd2
installer: Remove ContainerInstaller and ImportInstaller
...
Fold their non-logic into the base Installer class
2018-09-04 14:55:04 -04:00
0293fa492c
virt-install: Centralize installer building and validation
...
The ordering was strange here and forced us to duplicate checks
for cli validation
2018-09-04 14:55:04 -04:00
256ca7f3fb
virt-install: Simplify install method validation a bit
...
And drop the --pxe --nonetwork collision check, it's near pointless
2018-09-04 14:55:04 -04:00
0ed8946a5a
xmlbuilder: Remove the default_cb and default_name logic
...
It was always very hard to follow the logic chain with default_*
settings, having it explicit in set_defaults() function calls is
much easier to follow and more flexible for the future
2018-09-04 14:55:04 -04:00
1e76496fa8
storage: remove default_cb usage
2018-09-04 14:55:04 -04:00
221dd00cb9
devices: disk: remove default_cb usage
2018-09-04 14:55:04 -04:00
fff3c88f9c
devices: disk: Absorb Guest bus/target default setting
...
Simplifies cli stuff as a nice side effect
2018-09-04 14:55:04 -04:00
4e7652b548
domain: cpu: Add set_defaults
2018-09-04 14:55:04 -04:00
cf65bef6f5
domain: clock: add set_defaults
2018-09-04 14:55:04 -04:00
d5bff2c987
guest: Remove default_cb usage
...
This lets us simplify things in create.py a bit, since we don't
need to manually encode a UUID
2018-09-04 14:55:04 -04:00
e6219504ce
domain: os: Add set_defaults
2018-09-04 14:55:04 -04:00
c1443f8a80
guest: Remove fallback xen loader/emulator handling
...
This has been provided by libvirt caps for a long time, and we
don't even test it
2018-09-04 14:55:04 -04:00
468f5e67fa
domain: pm: Add set_defaults
2018-09-04 14:55:04 -04:00
5517033c79
domain: seclabel: Add set_defaults
2018-09-04 14:55:04 -04:00
9a5b903086
domain: features: Add set_defaults
...
And move all the Guest logic there.
We need to add an xmlbuilder helper _prop_is_unset to be able to
determine if a bool property is unset, in our quest to remove
default_* xmlbuilder bits
2018-09-04 14:55:04 -04:00
640a068650
devices: controller: Add set_defaults
2018-09-04 14:55:04 -04:00
6215f790f5
devices: filesystem: Add set_defaults
...
Requires a bit of rework in virt-manager handling, which had lots
of explicit "default" handling
2018-09-04 14:55:04 -04:00
ef83565e48
devices: hostdev: Add set_defaults
2018-09-04 14:55:04 -04:00
0f918a4bd5
addhw: Move interface model list to virtinst
...
So we are keeping the default and model list logic closer together,
as its similar
2018-09-04 14:55:04 -04:00
64dc334538
devices: interface: Add set_default function
...
Move all the Guest network default logic to it, and drop a now
redundant test
2018-09-04 14:55:04 -04:00
7b5f1a304d
guest: Publicize _os_object as osinfo
2018-09-04 14:55:04 -04:00
347143982a
devices: Add set_defaults for memballon, redirdev, rng, smartcard
2018-09-04 14:55:04 -04:00
35c0cf352f
guest: Move graphics defaults handling to DeviceGraphics
2018-09-04 14:55:04 -04:00
e29980959c
devices: tpm: Add set_defaults
2018-09-04 14:55:04 -04:00
882e9d1dee
virt-xml: Call set_defaults on created XML objects
2018-09-04 14:55:04 -04:00
221edf7d86
devices: char: Remove default_cb in favor of add_defaults
2018-09-04 14:55:04 -04:00
31bda5bbb6
devices: watchdog: Add set_defaults
2018-09-04 14:55:04 -04:00
31d74ff0eb
devices: input: Add set_defaults
2018-09-04 14:55:04 -04:00
698c72e60d
devices: sound: Add set_defaults
2018-09-04 14:55:04 -04:00
40de2c4762
guest: For implied devices, explicitly call set_defaults
...
These can be added after we initially call set_defaults on all devices,
so require a manual set_defaults call
2018-09-04 14:55:04 -04:00
1b0f325813
devices: video: Add set_defaults
2018-09-04 14:55:04 -04:00
d25ce81439
guest: Force qxl/vga more often
...
Letting libvirt choose gives us cirrus, which is really bad.
So for x86 for it more
2018-09-04 14:55:04 -04:00
fdd1cc5ca8
guest: Make spice default device additions explicit
...
Don't lump it in with setting video defaults
2018-09-04 14:55:04 -04:00
7738b8f407
devices: panic: Add set_defaults
...
Move default handling out of Guest and default_cb callbacks, and
make implement it in a standard set_defaults function
2018-09-04 14:55:04 -04:00
c3515fa469
guest: Drop multifunction address magic
...
Seems sufficiently obscure that if someone is manually specifying
addresses we can require them to pass multifunction as well
2018-09-04 14:55:04 -04:00
41daf1227c
choosecd: Emit cdrom-chosen signal after the dialog has been closed
...
This fixes the testChooseCDHotplug UI test: the cdrom-chosen signal
opens an alert and it does not get focus because it's being opened
before the dialog is closed. As a result, the tests timeout waiting for
the alert to get focus. Emitting the signal after the closure of the
dialog fixes this issue.
Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
2018-09-04 13:30:14 -04:00
3c62685bf1
uitests: Use Ctrl_L+Alt_L+Z as a grab keys when testing prefs
...
Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
2018-09-04 13:30:14 -04:00
568a39f1ee
uitests: Increase click offset for window drag
...
On KDE the testManagerWindowCleanup test clicks right on the border and
the drag operation fails, bringing down the test. Clicking a bit lower
fixes this issue.
Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
2018-09-04 13:30:14 -04:00
fb7b295e21
uitests: Fix selection of items in long menus that don't fit on screen
...
Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
2018-09-04 13:30:14 -04:00
895668fa9a
uitests: Fix onscreen to return True only when full widget is on screen
...
Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
2018-09-04 13:30:14 -04:00
bf823fdb3c
details: Pin network-model label directly to combo to simplify tests
...
This approach is used in majority of other places and allows the
combobox items to be selected directly instead of typing text in tests.
Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
2018-09-04 13:17:19 -04:00
c81d722a19
uitests: Improve robustness of click_combo_entry()
...
Signed-off-by: Povilas Kanapickas <povilas@radix.lt>
2018-09-04 13:17:19 -04:00
Cole Robinson