[nanomsg] Re: status on pull request 369 - websocket changes
- From: Garrett D'Amore <garrett@xxxxxxxxxx>
- To: nanomsg@xxxxxxxxxxxxx
- Date: Thu, 19 Feb 2015 11:41:31 -0800
> On Feb 19, 2015, at 11:26 AM, Dirkjan Ochtman <dirkjan@xxxxxxxxxx> wrote:
>
> On Thu, Feb 19, 2015 at 3:25 AM, Garrett D'Amore <garrett@xxxxxxxxxx> wrote:
>> I’ve got an outstanding pull request — #369 — that fixes some issues with
>> websocket and as a result mangos websocket would be compatible with nanomsg.
>> These include changes to the RFC, as well as fixes in the code.
>>
>> I would *really* like to see that integrated, rather than stagnating. The
>> original author of the websocket transport has already given it his thumbs
>> up. What else is required to see this move forward?
>
> The reason I didn't merge it was (but I should probably have made this
> explicit) because I was a little concerned about the handshake
> discussion that happened on the mailing list. IIRC, Drew Crawford
> expressed a dissenting opinion on how it should work, but I've been
> too busy to go back and dig it up.
Drew has acknowledged that his objections were based on his misunderstanding,
and has rescinded his objection.
>
>> I’m concerned that the process for figuring out what gets accepted, and what
>> doesn’t, is too opaque. In particular, I’m going to be making another round
>> of changes — I need to fix the Surveyor protocol to use a backtrace just as
>> Martin had discussed earlier — this will include RFC updates and code
>> changes.
>
> For things that Martin already agreed to, I'm happy to merge them.
> I'll also merge stuff that is an obvious bug fix or that other
> community members agree on (as expressed before, I think). Sorry, I do
> not feel confident enough to make sort of high-level decisions on
> Martin's behalf.
I can understand.
>
>> Yet I’m concerned — Martin seems to be busy as we’ve hardly heard from him
>> of late, and the project seems to be running somewhat rudderless. That
>> makes it a somewhat scary proposition for companies like mine that are
>> planning on using this stuff in production. My fear is that nanomsg itself
>> will stagnant due to lack of developer involvement. It would be really
>> really unfortunate for libnanomsg to become abandonware.
>
> I very much agree. I wrote to Martin a few weeks ago to ask him when
> he would be able to spend more time on nanomsg. He said that he was
> moving to a different country for a new job, where he wasn't
> comfortable using work infra to do nanomsg stuff. He said he would be
> able to spend some time on nanomsg from this week on, but obviously
> that hasn't happened yet. I do hope he will find more time to review
> stuff once he's a bit more settled in. I also asked him if we should
> appoint more people who can actually merge pull requests -- I think
> you would make a good candidate for this.
Ok, it sounds like there is hope then. Its not just a matter of being appointed
btw, but also having some way to figure out what the standards for quality,
etc. should be. The CI is one helpful bit here, but I don’t think it is
sufficient.
It looks like martin has also responded to my surveyor RFC, so this bodes well.
- Garrett
>
> Cheers,
>
> Dirkjan
>
Other related posts:
