[virt-tools-list] [PATCH virt-manager] virtManager, host: warn user for unapplied changes

[Giuseppe Scrivano]

Cole Robinson <crobinso at redhat.com> writes:

>> @@ -377,11 +387,15 @@ class vmmHost(vmmGObjectUI):
>>      def page_changed(self, src, child, pagenum):
>>          ignore = src
>>          ignore = child
>> +        self.confirm_changes()
>>          if pagenum == 1:
>> +            self.repopulate_networks()
>>              self.conn.schedule_priority_tick(pollnet=True)
>>          elif pagenum == 2:
>> +            self.repopulate_storage_volumes()
>>              self.conn.schedule_priority_tick(pollpool=True)
>>          elif pagenum == 3:
>> +            self.repopulate_interfaces()
>>              self.conn.schedule_priority_tick(polliface=True)
>> 
>
> What's this for?

this is to fix the case when the user switches from page A to page B and
choose No when asked to save changes.  Next time page A is visited,
these repopulate_* force the configuration to be reloaded.


>>      def refresh_resources(self, ignore=None):
>> @@ -485,7 +499,7 @@ class vmmHost(vmmGObjectUI):
>>          except Exception, e:
>>              self.err.show_err(_("Error launching network wizard: %s") % str(e))
>>  
>> -    def net_apply(self, src_ignore):
>> +    def net_apply(self):
>>          net = self.current_network()
>>          if net is None:
>>              return
>> @@ -532,15 +546,28 @@ class vmmHost(vmmGObjectUI):
>>              self.disable_net_apply()
>>  
>>      def disable_net_apply(self):
>> -        self.active_edits = []
>> +        for i in [EDIT_NET_NAME, EDIT_NET_AUTOSTART, EDIT_NET_QOS]:
>> +            if i in self.active_edits:
>> +                self.active_edits.remove(i)
>>          self.widget("net-apply").set_sensitive(False)
>
> This is going to get out of date fast. I'd say do something like:
>
> EDIT_NET_IDS = (
> EDIT_NET_FOO,
> ...) = range(5)
>
> EDIT_STORAGE_IDS = (
> ...) = range(100, 101)
>
> And then use those named tuples to implement just one disable_apply handler
> that doesn't list the EDIT_* bits explicitly.

ok, thanks, I will make this change and repost a new version.

Regards,
Giuseppe