[Top][All Lists]

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

Re: [GNU-linux-libre] The "Free" Kernel In Debian Squeeze

From: Richard Stallman
Subject: Re: [GNU-linux-libre] The "Free" Kernel In Debian Squeeze
Date: Mon, 27 Dec 2010 13:18:56 -0500

    > I think what we want is a program that will modify a git repository by
    > meta-deleting a certain range of code.

    Yup.  That's relatively easy.  The difficulty is in maintaining the
    ability to merge changes from one repository to the other.

To merge changes from the original repository to our repository, it
would suffice to re-convert the new version of the original
repository.  You can then merge from that into our repository.

That might be ridiculously slow, but I am sure it could be optimized.
Think of that as the reference implementation.  Even that reference
implementation might be good enough if we don't need to do it often.

Merging in the other direction not much harder.  The only difficulty
is that, in order to add an old change to a repository, it is necessary
to specify what the old change was.

But it isn't our problem.  We can leave it to be implemented by
someone who wants it.

    - rewrite the revision history, so that new hashes (that identify the
    commits) are computed for each revision => repositories become

That's what I am suggesting.  I have demonstrated it is no real

I think you are being hampered by the feeling that this ought to be
trivial.  You know it can be achieved with a moderate amount of work,
but you don't want to do it, because you feel at some level that it
shouldn't be necessary.

It were nice if this were trivial.  But we have the choice to do the
moderate amount of work or fail to fix the problem.

Isn't it better to do the moderate amount of work?

Richard Stallman
President, Free Software Foundation
51 Franklin St
Boston MA 02110

reply via email to

[Prev in Thread] Current Thread [Next in Thread]