Re: [Gnu-arch-users] [NOT-BUG] tla 1.2.2rc1 doesn't accept {arch}/=partner-versions file.

[Tom Lord]

    > From: David Allouche <address@hidden>
    > On Wed, 2004-09-08 at 16:30 -0700, Tom Lord wrote:
    > >     > From: address@hidden (James Blackwell)

    > >     > In fact, let's up the ante a little bit. If you not only find a 
bug, but
    > >     > provide a test case that A) passes with 1.2.1 but B) fails with
    > >     > 1.2.2rc1, I'll _double_ the award! 

    > > Now you are taking the development philosophy:

    > >         Make only correctness-preserving/enhancing changes to the 
    > >         program.

    > > to a (eventually dangerous?) new level.

    > I am not sure I understand what you mean.

    > The intention does not seem to be "we must not fix old bugs in the
    > stable version" but more something like "we should not do anything but
    > fix regressions once rc1 is out".

    > I expect the various issues which have been raised (apply-delta w/o
    > default archive, explicit tag of files in {arch}) to be acceptable to
    > fix in the stable branch once 1.2.2 is out.


Yes, you did not understand what I mean!  :-)

Successive releases of arch should fix bugs and add features.   Either
way, no change to arch should break something that already works.

Another way to say that:   "all changes must be `correctness
preserving or enhancing'."

A new feature that doesn't break anything already there is
"correctness preserving".

A change that fixes bugs or edge cases is "correctness enhancing".

To be more precise, I guess I should say that the goal is for changes
which are "correctness preserving AND correctness enhancing".

Saying "or" instead of "and" was misleading.

Now, with that general goal, jblack has put $10 of his own money on
the line --- betting that he has admitted only changes which are, at
least, "correctness preserving".

Prove him wrong and you win $10!

Why dangerous?  Uh.... the politics of labor generally and,
specifically, the implied opportunities to "game" this reward system
to somebody else's detriment.

-t