Re: 2.1.123 and fbcon.c

Theodore Y. Ts'o (tytso@mit.edu)
Tue, 29 Sep 1998 17:30:41 -0400


Date: Tue, 29 Sep 1998 11:07:13 -0700 (PDT)
From: Linus Torvalds <torvalds@transmeta.com>

The reason I'm disappointed is that vger in particular has been acting as
a "buffer" between me and bug-fixes, so that now we're in the situation
that there are obviously bugs, and there are obviously bug-fixes, but I
don't see it as such, I only see this humongous patch.

I don't know what it fixes, because vger has kept me out of the loop, and
quite frankly I don't have the time to look at several hundred kilobytes
of compressed patches by hand. And I refuse to apply patches that I don't
feel comfortable with.

Linus,

To be fair to the vger people, one of the problems which the
vger CVS tree is trying to fix is that sometimes you don't take patches
very quickly. There's been at least one set of tty patches which I had
to send you two or three times before you finally accepted it --- and
they were short patches (1-3 line changes in 4 files), and with a full
explanation of what it did. Heck, in the most recent case you and Alan
and I discussed different approaches for solving this problem before I
even started coding the patch.

Yet I had to resubmit the patch and the explanation at least
*THREE* times, over the course of 2-3 weeks, before I finally got a
response out of you. And this was for a utterly uncontroversial patch!
And this was not the first time I've had to resend patches 2 or 3 times,
either. Is Martin going to have to send you each driver update 2 or 3
times before they finally get accepted?!?

Now, I've had to do this enough times that I just simply
consider it par for the course, and the changes are important enough
that I'm willing to repeatedly resend you patches again and again. But
the risk is that other people who are not so persistent my give up, and
we therefore end up seeing patches get dropped.

So we have a problem, and perhaps vger isn't the best solution.
Maybe the linux-patches web page is a better solution. But before you
go slamming the vger folks because this patch batching effect which you
don't like, it might be nice for you to acknowledge that some of your
bandwidth constraints may have contributed to the problem, and they were
simply trying to find a way around it. If the vger CVS tree doesn't
work, fine, we can try to find another solution.

But before we can find an effective solution, it often helps to
admit that we have a problem.

- Ted

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/