[parted-devel] next to master; making a beta release soon; plans

[Jim Meyering]

Joel Granados wrote:

> On Tue, Sep 15, 2009 at 04:26:33PM +0200, Jim Meyering wrote:
>> Heads up:
>>
>> In the next couple of days I expect to rebase all of next and pull
>> it onto master, and then to begin working towards a near-term beta release.
>>
>> Things we should discuss first:
>>
>>   - how to handle the soon-to-be-deprecated FS-manipulation subcommands,
>>     mkfs, mkpartfs, check, cp.
>>
>>       Our options:
>>
>>         * warn-and-try-anyway: I would not advise this: the FS code is
>>           so old and bit-rotten, and I've made no attempt to adapt it
>>           to handle the possibility of >512-byte sectors.  In addition,
>>           there are relatively few tests that exercise it.
>>
>>         * simply fail, possibly suggesting to try parted-1.9.x
>>           I favor this one.
>
> +1 to this one.  Are you talking about introducing the failure in
> parted cli or parted lib?  and you have spoken about parted behavior but
> have not touched on what we are to do to the actual code.  Should we
> keep it or simply remove all traces of it.  If we are to fail, I think
> removing it is a worth while possibility.

Yes, definitely.
The reasons for removing functionality from the UI
(code is dangerously buggy) apply equally well to the API.
Actually removing the code will involve more work up front,
but will be better in the long run.

> This will mean that the
> parted cli user will see a fail of the type.  "mkpartfs -- unknown
> argument" and the project that uses parted lib will see a failure in the
> build when it does not find certain parts of the API.