Re: [pve-devel] [PATCH common] fix #3153: INotify: adding interface comment to inet6 section when this is the only section

[Lorenz Stechauner <l.stechauner@proxmox.com>]

On 25.06.21 13:18, Thomas Lamprecht wrote:
> On 25.06.21 12:48, Lorenz Stechauner wrote:
>> Signed-off-by: Lorenz Stechauner <l.stechauner@proxmox.com>
>> ---
>>   src/PVE/INotify.pm | 14 +++++++++++++-
>>   1 file changed, 13 insertions(+), 1 deletion(-)
>>
>> diff --git a/src/PVE/INotify.pm b/src/PVE/INotify.pm
>> index 562a243..ce1efd1 100644
>> --- a/src/PVE/INotify.pm
>> +++ b/src/PVE/INotify.pm
>> @@ -1156,6 +1156,12 @@ sub __read_etc_network_interfaces {
>>   	$d->{method} = 'manual' if !$d->{method};
>>   	$d->{method6} = 'manual' if !$d->{method6};
>>   
>> +	if ($d->{comments6}) {
>> +	    $d->{comments} = '' if !defined($d->{comments});
>> +	    $d->{comments} .= $d->{comments6};
>> +	    $d->{comments6} = undef;
> rather than setting it to `undef`, in which case the hash element still exists, use
> `delete`, which also returns the value.
>
>
> So this could be:
>
> if (my $comments6 = delete $d->{comments6}) {
>      $d->{comments} = ($d->{comments}) // '') . $comments6;
> }
>
> Shorter and also more correct/robust.
>
>> +	}
>> +
>>   	$d->{families} ||= ['inet'];
>>       }
>>   
>> @@ -1242,7 +1248,7 @@ sub __interface_to_string {
>>       my $done = { type => 1, priority => 1, method => 1, active => 1, exists => 1,
>>   		 comments => 1, autostart => 1, options => 1,
>>   		 address => 1, netmask => 1, gateway => 1, broadcast => 1,
>> -		 method6 => 1, families => 1, options6 => 1,
>> +		 method6 => 1, families => 1, options6 => 1, comments6 => 1,
>>   		 address6 => 1, netmask6 => 1, gateway6 => 1, broadcast6 => 1, 'uplink-id' => 1 };
>>   
>>       if (!$first_block) {
>> @@ -1733,6 +1739,12 @@ NETWORKDOC
>>   	    }
>>   	}
>>   
>> +	# if 'inet6' is the only family or is at least a family in use
>> +	if ($d->{families}[0] eq 'inet6') {
> shouldn't that include a length check like `scalar($d->{families}->@*) == 0` to be more safe/robust,
> as else inet6 could be just be first by chance (did not really checked the context though)

the length check would not hurt - true. But it does not matter, in which 
section (inet/inet6) the comments gets put. The reader method unifies 
these comments anyway. So the comment *might* gets put in the inet6 
section, event if there is another inet section. Should not matter, 
should it?

Will add a "lenght" == 1 to the if.

>
>> +	    $d->{comments6} = $d->{comments};
>> +	    $d->{comments} = undef;
> see above regarding use of delete
>
>> +	}
>> +
>>   	my $i = 0; # some options should be printed only once
>>   	$raw .= __interface_to_string($iface, $d, $_, !$i++, $ifupdown2) foreach @{$d->{families}};
>>       }
>>