[Cerowrt-devel] Rebasing, rescheduling and rethinking cerowrt
dave.taht at gmail.com
Sun Dec 4 03:37:50 EST 2011
I am in the process of rebasing cerowrt on openwrt head,
reworking and cleaning up the patches that were good
and submitting those to the maintainers, and doing
a rethink as to how to proceed forward for rc8 given
all the (positive!) bufferbloat related churn in the
linux 3.2 and 3.3 development processes.
Too much changed out from under Cerowrt-rc7 to
deliver a stable release.
It IS serving it's original
purpose in being a R&D platform for inquiring
into bufferbloat and the difficulties home gateways
face in the coming decade admirably, but, a thorough
review of the roadmap for rc8 shows
several man years of development is required
to deliver everything on that list, and that list
is incomplete at best.
This is a major problem. Something has to give.
Yet there's some really good stuff in there. In
particular the new AQMs are looking promising,
as is BQL.
The cerowrt repository on github has been renamed to
Cerowrt-rc7. There will not be a 'final' rc7 release.
There will not be a public cerowrt-rc8 repo
for a while, either, while we get the above sorted out.
If you have any out of tree updates to cerowrt-rc7 and
would like them considered for rc8, please submit
them to the cerowrt-devel list.
This will be the last message to the bloat-devel list
regarding the cerowrt development process. Future
cross postings will stay focused on bloat related
R&D. I will also be adding some infrastructure for
postings to cerowrt-commits and more automated
builds in the hope of reducing the administrative
Any suggestions on how to better maintain and
interoperate with source trees kept in svn and git,
incorporating hundreds of packages, and still find
a decent way to hit one button and build an OS
are AND still move constructively forward
with innovative R&D are *welcomed*.
US Tel: 1-239-829-5608
FR Tel: 0638645374
More information about the Cerowrt-devel