Re: [PATCH v2 00/20] btrfs-progs: Rework of "subvolume list/show" and relax the root privileges of them

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2018/08/16 3:12, David Sterba wrote:
> Hi,
> 
> so I've merged the first bath to devel. There was quite some whitespace
> damage and style issues, but I fixed that as I went through the pathes.

Thanks for taking this.

> 
> I've left out all patches that introduce new options, this is the hard
> part and single letter options cannot be added just randomly. That's
> what led to the current state.

Ok, but could you consider to add -f (11th path, or --follow instead of
short option) which is useful when btrfs is used for root fs
(or please remove cli-tests/010 for the moment from devel branch).

> 
> The semantics of current options is preserved, with the slight exception
> to the option -a.

Devel branch also excludes 13th patch and behavior of -a is not changed,
is it intended? Without the patch, there is no way to get the same result
of current "btrfs subvol list" when mount point is not FS_TREE.

> 
>> * Future todo:
>> If this approach is ok, I'd like to update the output of "sub list" more like:
>>   - Remove obsolete field (i.e. top-level) from output
> 
> Yeah the top-level field can be removed from the user output. Current
> idea of the fields printed by default is: id, generation, type, path.
> The type is a representation of eg. read-write status,
> subvolume/snapshot, and maybe received, or if it is default.
> 
> The remaining fields are on-demand by other options and the future proof
> solution is to provide the single option to specify which columns to
> print.
> 
> The inline format, ie. the current one, is not very comprehensive so the
> table layout should be the default for user.
> 
> The path format needs to be selected by a more concise way than just -A
> and possibly more added options. Eg. --path
> 
> * absolute (shorthand abs)
> * realative to the current path (pwd)
> * toplevel relative (toplevel)
> 
> In order to make it possible to distinguish in the output without prior
> knowledge of which one was selected, I propose this formatting:
> 
> * absolute path always starts with / and is really absolute to the /
> * relative path to the current directory starts with ./
> * toplevel does not start with ./ nor /
> 
> There are patches touching the libbtrfsutil code, so I'd like to hear
> from Omar what he thinks.

Also, I want to get acknowledgement about 7th patch
(as I wrote in cover letter, I'm ok with original version too).

> 
> What's now in devel should not change and provide base for further
> patches, so please use it if you're going to implement something from
> the above
Understood. I will consider them if current patches in devel has been settled.

Regards,
Misono




[Index of Archives]     [Linux Filesystem Development]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux