but not too early. Because of an ExtJS bug/limitation, it can only happen after
the window is rendered, so use an afterrender listener. Without setting the
mask there, the window will be active already before the storage selectors
change listener triggers, which can only happen after the storage selectors
store is loaded.
Made noticable by the new "filling in defaults" behavior, but the issue was
already present earlier, where the compression selector for PBS storages would
be disabled late, after the window was already active.
Also move the setValue call into the afterrender listener, so ordering is easy
to verify/more stable.
Reported-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
so that mailto and mode are not overwritten if the first /vzdump/defaults call
failed, but a later one succeeds.
Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
could be triggered when there are no backup storages at all configured or if
the 'Backup now' button is clicked before the storage selector from the guests
'Backup' tab could load its store.
Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
Note that it's not possible to use read_vzdump_defaults() and storage_config(),
because they auto-converts maxfiles already.
Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
The change not to pass the 'upgrade' parameter in the frontend was made in
953f6e9bb3 (the commit doesn't talk about it, it's
likely an accidental squash of two changes)
Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
The switch to 'cmd' was made by commit af39a6f09651e15d1c83536e25493a2212efd7d3
in the pve-xtermjs repo and is included in 4.7.0
Signed-off-by: Fabian Ebner <f.ebner@proxmox.com>
everywhere where Pool.Allocate was unnecessarly used it was replaced
with Pool.Audit.
`/cluster/resources` now returns pool infomation for guests only if
the requesting user has the Pool.Audit permission on the pool.
`/pool/` now returns only pools where the requesting user has the
Pool.Audit permission.
Signed-off-by: Lorenz Stechauner <l.stechauner@proxmox.com>
Without this the check and the, through the 'install' target
triggered, incremental lint target triggered a full eslint run.
Makes it similar to what PBS did from the beginning of eslint
inclusion..
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
preparatory steps for better make dependencies when executing the
lint check.
People can still just make the .js targets directly or use the
install target.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
there was a spurious leftover single quote("'"), that prevented
the correct css class to be inserted
(e.g., fa-exclamation' instead of fa-exclamation)
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
on extjs 7.0, this triggers when the handler is not available anymore,
so use the beforedestroy event instead
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
treelist technically allows deselection and this happens on destruction
in extjs 7.0 so simply return early to avoid errors accessing
components that do not exist anymore
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
else with extjs 7.0, the first render does not happen yet, because
the grid is not visible at that time
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
remove vmid from data part, it is already contained in object part.
this is accomplished by adding the parameter $excluded to
build_influxdb_payload().
Signed-off-by: Lorenz Stechauner <l.stechauner@proxmox.com>
while the function in '.on' and '.un' are identical, they are not the
*same* function, thus the '.un' does not really remove the listener,
and we have leftover references to the grid which means it will never
really garbage-collected
instead, use '.mon' on the grid, which automatically cleans up the
listeners
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
otherwise the grid is confused about the row height (since it is
variable here) and renders the scrollbars weirdly
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
match returns 'null' if the regex does not match, which is not
destructurable. so we have to save the match and check if it valid
Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
