Re: [pbs-devel] [PATCH proxmox-backup 1/2] fix #3853: api: add force option to tape key change-passphrase

[Stefan Sterz <s.sterz@proxmox.com>]

On 2/8/22 16:26, Dominik Csapak wrote:
> On 2/7/22 17:14, Stefan Sterz wrote:
>> On 2/7/22 16:57, Stefan Sterz wrote:
>>> On 2/7/22 15:58, Thomas Lamprecht wrote:
>>>> On 07.02.22 13:48, Stefan Sterz wrote:
> [snip]
>>>>> +    // sanity checks for "password xor --force"
>>>>> +    if force && password.is_some() {
>>>>> +        bail!("password is not allowed when using force")
>>>>> +    }
>>>>> +
>>>>> +    if !force && password.is_none() {
>>>>> +        bail!("missing parameter: password")
>>>>> +    }
>>>> Above two if's could be written slightly shorter while IMO even 
>>>> improving readability
>>>>
>>>> match (force, password) {
>>>>      (true, Some(_)) => bail!("password is not allowed when using 
>>>> force"),
>>>>      (false, None) => bail!("missing parameter: password"),
>>>>      _ => (), // OK
>>>> }
>>> This does not work, because here password is moved into the match 
>>> expression. The borrow checker will complain about it being used 
>>> later on when trying to decrypt the key configuration. You could 
>>> clone password here, but this solution strikes me as rather 
>>> "inelegant".
>
> did not look at the rest of the patch really, but i think thats wrong..
>
> couldn't you simply use the reference instead? then the value
> will not be moved?
>
> match (force, &password) {
> ...
> }
>
>
> ?
Yes sorry, using a reference works too. The value is moved when the
tuple is constructed is what I meant. This can be avoided by cloning
it or as you mentioned using a reference.