Device layering work patch #1

Patrick Mauritz oxygene at studentenbude.ath.cx
Tue May 18 06:26:26 PDT 2004


esmith wrote:
> People might not mind if these risky patches are directly committed and
> let the chips fall where they may, since there hasn't been an official
> release yet. The pre-release state of this project should implicitly let
> everyone know this software is always potentially unstable and we should
> use it at our own risk. If someone wants to use software that is safe
> and reliable they can use something that has an official, stable
> release. A warning message on the kernel list could be sufficient notice
> of particularly risky updates.
that way you get the "big" things sorted out faster, but with less people
willing to adopt such a system all the small issues won't be handled 'til
after the first "stable" release - which isn't that stable then.

in longterm, better tools than cvs might help with using _many_ branches,
merging back and forth between feature branches (eg. "device layering"), an
integration branch and a stable branch, without them getting too much out
of sync. (notice that I said "longterm" - I'd say it's up to the commit bit
people to decide when/if the revision control (flame) wars should start
here)


patrick mauritz





More information about the Kernel mailing list