[parted-devel] [PATCH] Document the lack of support for ext3 filesystems.

[Jim Meyering]

Joel Andres Granados <jgranado at redhat.com> wrote:
> Jim Meyering wrote:
>> Thanks for the patch.
>> Deprecating ext3 support is a good first step.
>
> yep, its all good.  My idea is not to take it away completely though.  The more I think about this issue the more I realize why it was put there in the first place.
> Reasons that I think we should support ext3:
> 1. There is a library, makes things easier and makes stuff less prone to bugs.

As long as the bit-rotten ext* code remains, using Parted to
manipulate file systems is like playing russian roulette.
The least we can do is to warn users of the risk.

> 2. I like the idea of creating an ext{3,4} filesystem with the same
> tool that I do the partitioning.  Its just less hassle.

Of course, that would be best.  But highest priority
for me is to clean up parted's partition management code.

> 3. I would like parted to have at least filesystem knowledge.  This would allow parted to be more informative to the user when a partition/disk is being handled.  Parted could say something like "dear user, the end size of the partition is smaller than the filesystem you have inside, do you really want to destroy this filesystem?"
>
> Are we taking away all traces of filesystem support?

I merely mentioned the long-term possibility.
Of course, before taking such drastic measures,
we'd have to announce the plan, then make a release
of the tool that warns when using the offending functionality.
Only once that release has made it into wide distribution
(and assuming no one has found time to make parted use
the libraries) would we actually remove functionality.

>> However, I think we should be warning that as much
>> file-system manipulation as possible should be done
>> by tools other than parted.  E.g., rather than using parted's
>> mkpartfs command, just use mkpart to create the empty(unformatted)
>> partition, and then use an external tool like mkfs.ext2 to create the
>> file system.
>
> I'll put what you suggested in the messages.
>
>>
>> I know from bug-fixing experience that the ext2-FS-manipulating code in
>> parted is far too old to be considered reliable, and all you need to do
>> is choose slightly unlucky or unusual parameters (possibly due merely
>> to the layout of the target partition) in order to provoke
>> data corruption.
>>
>> As I said, this patch is a fine first step, and
>> I'll commit it once you make a few small changes:
>>
>> [this wording appears in two files]
>>> +...  Tools like resize2fs (8) and mke2fs (8) are
>>> +suggested for managing these types of filesystems.
>>
>> Use an active voice.  I.e., instead of the above, say e.g.,
>
> right active voice.  Shouldn't be difficult.
> Do you have a reason for choosing one besides the other?  just curious ;)

It's universally preferred, e.g., in writing style guides.
I've heard that it's also mentioned in some Red Hat
documentation style guides, but don't have a pointer.

>>     Use tools like resize2fs(8) and mke2fs(8)
>>     to manage these types of filesystems.
>>
>> Also, while I know at least one "please" was preexisting,
>> please remove all of the ones in the lines changed or
>> added by your patch.
>
> mmmm, don't like the magic word....
> I'll remove every trace of niceness in the documentation ;)

No need to go overboard ;-)
Just seemed like those particular ones added nothing.
Felt a little inappropriate.