Re: 2.1.22 name space analysis

Hans Lermen (
Wed, 29 Jan 1997 01:28:02 +0100 (MET)

On 28 Jan 1997, David Hinds wrote:
> Keith Owens ( wrote:
> : On Mon, 27 Jan 1997 13:13:48 -0500,
> : "Theodore Y. Ts'o" <tytso@MIT.EDU> wrote:
> : >Grumble.... *when* will where be a stable kernel/modutils combination
> : >that doesn't require 200+ kilbytes to the kernel before you have
> : >something which might (or might not) work? I understand that things
> : >might be unstable for a little while during the 2.1 development series,
> : >but if things remain too broken for too long, developers won't be able
> : >to work against the latest versions, either. This is Bad (tm).
> : Perhaps it's time for a bug fix freeze. Get the existing code to
> : compile clean and run, together with clean modutils and pcmcia before
> : rewriting more sections of the kernel.
> Agreed. I won't move to the latest 2.1.X kernels until module support
> works, period. As in, no extra patches, and a stable set of module
> tools. I'm totally dependent on having working modules.

and on Mon, 27 Jan 1997 23:49:40 GMT Alan Cox wrote:

> Not a good idea for the net code as their are some chunks ahead that
> will cause small changes but large scale noise.

Perhaps we can have a compromise. What about regulary 'calm down phases'
instead of 'freeze phases', e.g. every 5 releases, including anouncements
when these phases start and stop.
Durings these 'calm down phases' there could be bug fixes, minor code
changes and 'face lifting', but no addition of 'big' chunks of new stuff.