[quagga-dev 12340] Re: [PATCH] babeld: Remove babeld from Quagga

Martin Winter mwinter at opensourcerouting.org
Mon May 18 22:50:57 BST 2015


On 18 May 2015, at 8:00, Paul Jakma wrote:
> On Sun, 17 May 2015, Martin Winter wrote:
>
>> Why was babel removed?
>
> Suggested by several people on list, and none of the maintainers 
> disagree.

Or most people (like me) expected this to get resolved some other way. 
This is mainly based on that previous message of yours. Honestly, I 
expected more public discussion on the list based on your commitment of 
changing the commit messages if this would be acceptable to Juliusz and 
the pointer to the exact details on the additional GPL headers.
I really expected for others to comment, but it got silent on the list 
after this.

But assuming the case that a patch gets sometimes NACK’ed by a 
maintainer and there was no proposed patch for this
on the list, how would you know if none of the maintainers disagreed? 
Was this in some private discussion?

>> I don’t see any (public) request by Juliusz and your last message 
>> in this regard said:
>> (Message with Subject “[quagga-dev 12262] babeld sync-up”
>
> As far as I know, Juliusz does not want us to distribute his code with 
> GPL notices. I understand he doesn't intend to discuss that further. 
> However, legal advice suggests GPL notices are advisable on at least 
> some of the quagga-babeld files, if we are to distribute them.

For the record: ONE legal advice (from Software Freedom Law Center) 
which was (as far as I know) all over email
suggested this. I do not know how far you went into the details on the 
conversation with them. I had some (informal conversation with lawyer) 
which contradicted this. I have no clue on who is right or wrong, but 
it’s not that clear
without further (more detailed) discussions.

>> I did not see any message after this and I saw advanced notice of 
>> this removal (I would have expected the usual 14 day advance notice).
>
> It was proposed early last week amongst the maintainers to remove it 
> at the end of the week, and there were no objections to that as far as 
> I know. Only a discussion on the commit message.

I would *REALLY* like any master changes to go with a diff through the 
mailing list - with the usual 14 days
waiting period. Just for the case to avoid breakages and have more eyes 
on a commit. This commit was clearly not
well tested (I’m still not sure how you run the configure without 
finding this issues)
(And yes, for the record, I think the commit message is misleading)

>> The advanced notice would also have helped to someone catch possible 
>> failures. The build is now broken because of this. (I would have 
>> expected a proper build to just ignore the “—-enable-babeld” 
>> and my build still succeed even if I have this enabled. But the build 
>> now fails. See https://ci1.netdef.org/browse/QUAGGA-QMASTER-58 )
>
> Is it just cause of the --enable-babeld? Can you just remove that?

No. See message by Donald Sharp “[quagga-dev 12339] vtysh fails to 
build after babeld removal commit”.
Let’s continue to discuss the technical parts off the failure under 
that thread - easier for others to find and
follow.
Would appreciate a quick resolution on fixing the build.

> I'm loathe to make the configure scripts be a binding API. ;)

Mistakes in the code can be forgiven, but not running through the normal 
(referring to the 14days notice) procedure and committing something 
which seems to break most builds points at a serious lack of testing and 
no planning.

- Martin Winter




More information about the Quagga-dev mailing list