Deliberately breaking software (was: cvs commit: src/sys/dev/vinum vinumioctl.c)

[Poul-Henning Kamp]

In message <20030505055618.GA84427 at wantadilla.lemis.com>, "Greg 'groggy' Lehey"
 writes:

>>>  Forgotten by: phk
>>
>> Nope, I knew about it.
>
>And you broke it anyway?  Why?

No, vinum disregarded how our disk system worked and implemented only
part of the API, that is why you needed a kludge option ("-v") in
newfs for vinum.  Compare this to ccd which did not need it.

>I don't recall that.

_That_ is not my problem.

>Anyway, you can't force people to maintain software.

I find it deeply ironic that you say this to distance yourself from
your responsibility as author of vinum, right after trying (and
before for that matter, you try again in this email) to force
me into fixing vinum for you.

But besides, we don't need to force people to maintain software:
Either the software is valuable enough for somebody that they will
maintain it, or it will rot away and be removed in due time.

>I relinquished my maintainership for Vinum simply because I
>couldn't work with you.

That's your choice.  By dropping vinum like a hot potato, you 
risk that nobody else picks it up (see above).

And I may add that you did nothing to make it easy for me to help
you with vinum.

>Judging by your present attitude, I don't
>think anything has changed.  If you make changes, you should ensure
>they don't break anything.

I make sure not to break vinum at the level I can test it.  Currently
the only test-case I have for vinum is "does it compile ?", and
since vinum is in LINT, that is an implict test for me.

>Your change broke things.  Don't blame the fact on other people.

Vinum never followed the rules in the piece of the kernel where was
stuffed in.  Considering that I tried very hard to teach you how
it had to behave, I don't think you can argue that I broke it and
that I have to fix it:  It was never unbroken to begin with.

Right now, you still do not use disk_create() to create proper
disk-device in vinum, and that means that features which work with
other disk devices in the system do not work with vinum.

When do you plan to fix that ?  Or should I simply tell people that
disk encryption does not, and will not ever work with vinum ?

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.