Re: [PATCH] btrfs-progs: print-tree: fix chunk/block group flags output

[Qu Wenruo <wqu@suse.com>]

On 2021/10/12 21:59, Nikolay Borisov wrote:
> 
> 
> On 12.10.21 г. 15:24, Qu Wenruo wrote:
>>
>>
>> On 2021/10/12 19:56, Nikolay Borisov wrote:
>>>
>>>
>>> On 12.10.21 г. 14:42, Qu Wenruo wrote:
>>>>
> 
> <snip>
> 
>>
>> There is a special handling for SINGLE:
>>
>> +               /*
>> +                * Special handing for SINGLE profile, we don't output
>> "SINGLE"
>> +                * for SINGLE profile, since there is no such bit for it.
>> +                * Thus here we only fill @profile if it's not single.
>> +                */
>> +               if (strncmp(name, "single", strlen("single")) != 0)
>> +                       strncpy(profile, name, BG_FLAG_STRING_LEN);
>>
>> See, if we hit SINGLE profile, we won't populate @profile array.
> 
> Fair enough, I had misread the != 0 check ... However, I'm wondering,
> since this is only used during output, why don't we, for the sake of
> consistency, output SINGLE, despite not having an exclusive bit for it?
> The point of the human readable output is to be useful for users, so
> instead of me having to know an implementation detail that SINGLE is not
> represented by any bit, it will be much more useful if I see that
> something is single profile, no ?

On the other hand, this breaks the consistency of flags output.

We only output flags we have on-disk.

Showing another flag which doesn't have on-disk bit can also be 
confusing, and break the existing output format.

Thanks,
Qu
> 
>>
>> Thanks,
>> Qu
>>
>>>
>>> <snip>
>>>
>>
>