[PATCH 1/2] inflate_codes(): Add fix from upstream gzip to prevent false CRC error

Rob Landley rob at landley.net
Tue Feb 16 19:08:14 UTC 2010


On Sunday 14 February 2010 10:38:23 Joakim Tjernlund wrote:
> Rob Landley <rob at landley.net> wrote on 2010/02/14 10:05:42:
> > > Applied, thanks!
> >
> > The last gplv2-compatible release of gzip was 1.3.12, released in 2007. 
> > Both 1.3.13 and 1.4 are GPLv3-only.  The above is a diff from 1.3.13 to
> > 1.4, including the name of the new variable "delta".
> >
> > So it looks like you've just introduced gplv3-only code into BusyBox,
> > which conflicts with the license on the Linux kernel, including code like
> > ifenslave.c which came from the kernel's Documentation directory.
>
> Oops, but on the other hand there is nothing unique about that snippet. I
> would probably have written it the same way, possibly with another variable
> name. You could rename it "dist" (for distance) if you like.

So you consider the Free Software Foundation to be more sane than SCO, then.

I don't.

> > There's a reason the Linux kernel uses the zlib deflate implementatin,
> > which is bsd licensed.
>
> That and it is better structured.

Back in 1996 I wrote my own deflate implementation in Java under OS/2, vaguely 
based on the info-zip source code and the old dos ZIP deflate description file.  
(Then the 1.1 JTK came out with inflate/deflate as part of the standard library, 
and I never got around to the decompression side...)  It's been quite a while, 
but I have some vague residual understanding of the protocol. :)

However, the linux kernel has lib/zlib_deflate and lib/zlib_inflate and they 
look fairly straightforward.  They're already maintained and optimized by the 
single largest open source development community on the planet (the linux-
kernel guys), under a license that's guaranteed to remain compatible with 
BusyBox.

I'd say the kernel's version is probably the best one for us to use.  I can 
take a stab at porting it over if Denys is interested...

Rob
-- 
Latency is more important than throughput. It's that simple. - Linus Torvalds


More information about the busybox mailing list