[quagga-dev 15869] Re: Someone rebased volatile/patch-tracking/8/proposed/ff branch

Lou Berger lberger at labn.net
Thu Jul 7 03:22:16 BST 2016


On July 6, 2016 10:53:19 AM Paul Jakma <paul at jakma.org> wrote:

> On Wed, 6 Jul 2016, Lou Berger wrote:
>> - bgpd now starts, but exit is no longer supported in config files
>> (which breaks existing configs) for example
>>    router bgp 123
>>    ...
>>    exit
> Are you testing at 'ff' (which will need a forced update, if you've
> previously fetched it - note; as will all these heads)? Or at
> 'pushback/b (which is sort of what 'ff' used to be at)?
It was a new clone:
 git clone git://git.savannah.nongnu.org/quagga.git 8ff-v3
 cd 8ff-v3
 git checkout -t remotes/origin/volatile/patch-tracking/8/proposed/ff

> I'm guessing that if the problme is that 'exit' that it is due to:
> 'lib, vtysh: Add support for marking a file with appropriate end of
> context'

nope, is due to changes in lib/command.c in 8/ff

> Which I have in the 'pushback/b' subset, pending a comment to Dinesh
> about compatibility. If you test from:
> proposed/ff - commit 4adca54da37e3d9360bcb4ab6d69ff20821ff31e
no joy

> does it work? If that works), a test from:
> proposed/nits/b - commit 77a73bee9b636024bbc9e8ba7803b8f08d675c7d
> Would be useful. Does that work?

>> - I did a simple test propagating a VPN safi update (in a RR config) and
>> the route didn't propagate.
>>  I have to run to meetings, but something in update processing is
>> definitely broken.
> Ok, if you can isolate to where, that'd be good.
ci, ci, ci, ci...

I don't have a good automated way to go backwards at the moment.  Maybe
Martin does - I probably won't have time to try to isolate until the end
of the week (and ID cutoff).  I'll see if Paul Z can...

> Also, it could be my editing. I had to do a lot of editing to fix merge
> issues, I might have fat-fingered stuff (no doubt did somewhere).
> Apologies in advance.

Understood - know you're working hard on this...
> regards,
> --
> Paul Jakma | paul at jakma.org | @pjakma | Key ID: 0xD86BF79464A2FF6A
> Fortune:
> All the existing 2.0.x kernels are to buggy for 2.1.x to be the
> main goal.
>  	-- Alan Cox

More information about the Quagga-dev mailing list