Open Stack

Websockets is an API for bi-directional client-server communication over TCP, not an implementation. Thereby it's not "an alternative" per se; alternatives to websockets would be long connections with chunked transfer like comet, etc. (which also have issues). There are various implementations of websockets in various languages; the Python implementations have some very serious problems (see my earlier posts in this thread, or the blueprint whiteboard linked below).

For now though, for *anyone* who's interested in that discussion, please direct any further feedback on that future feature to the blueprint (https://blueprints.launchpad.net/horizon/+spec/realtime-communication). I've linked a number of resources from the blueprint whiteboard. I recommend doing your homework and/or testing your hypotheses for Python-based solutions as there's a lot of prior work in this area that's better-off not being rehashed to death.

As far as this thread is concerned, I think we've achieved sufficient consensus on the current topic (LESS) and we'll move forward with that blueprint and code review as such.

All the best,

    - Gabriel

> -----Original Message-----
> From: openstack-bounces+gabriel.hurley=nebula.com at lists.launchpad.net
> [mailto:openstack-
> bounces+gabriel.hurley=nebula.com at lists.launchpad.net] On Behalf Of
> Adam Young
> Sent: Wednesday, May 30, 2012 2:14 PM
> To: Devin Carlen
> Cc: openstack at lists.launchpad.net
> Subject: Re: [Openstack] Nodejs in horizon
> 
> On 05/30/2012 04:37 PM, Devin Carlen wrote:
> > Long story short - we will work to make node.js an optional build time
> component and leave it as an distro packaging issue.  node.js was being
> evaluated as a potential solution to
> https://blueprints.launchpad.net/horizon/+spec/realtime-communication,
> but that blueprint isn't targeted for Folsom, so it's very future.  We'll have a
> lot of time to evaluate python based alternatives.
> 
> Sounds good.  Does websockets qualify as an alternative?  If so,  I'll update
> the blueprint
> 
> 
> 
> >
> >
> > Devin
> >
> >
> > On May 29, 2012, at 10:26 AM, Adam Young wrote:
> >
> >> On 05/29/2012 12:29 PM, Gabriel Hurley wrote:
> >>> Without rehashing backstory which is available in public archives of this
> thread, while node is currently on the table for LESS it also may play a role
> >>> in future needs as well.
> >>>
> >>> As for your link, yes there are LESS compilers in other languages (there's
> even a nascent one in Python that's very much not ready for prime time) but
> requiring PHP or any other language besides Python opens a similar can of
> worms, and PHP gets us nowhere on any other fronts.
> >>>
> >>> /me refrains from any other comments about PHP.
> >>>
> >>> Moreover, the canonical LESS compiler which is the best supported and
> most stable is the main one written for node.js. Unless there were a 100%
> working Python version I see no reason not to favor the "real" LESS compiler.
> >>>
> >>> All the best,
> >> For LESS,  it think it is fine to use even server side Javascript.  The CSS
> should be compiled at RPM/DEB build time, and not at run time for live
> deployments,  so that is a bit of a non-issue.  I'd also be fine with using the
> client side LESS implementation, especially if we want to use the same
> implementation at development and live deployment time,  but I can
> understand if there are issues with doing this.
> >>
> >>
> >> Node.js is a whole different server side technology,  and that should not
> be implemented at this time.
> >>
> >>
> >>>      - Gabriel
> >>>
> >>> On May 29, 2012, at 1:00 AM, "Matthias Runge"<mrunge at matthias-
> runge.de>   wrote:
> >>>
> >>>> On 28/05/12 16:21, Thierry Carrez wrote:
> >>>>> John Postlethwait wrote:
> >>>>>> Sorry if I've missed anything below, this thread has become rather
> >>>>>> fragmented and messy (at least in my email clients) but I will try to
> >>>>>> address the main points I have seen so far:
> >>>> Sorry, if I jump in late in this thread, I may have skipped some basics.
> >>>> If I get it right, nodejs is just required to compile LESS to css, right?
> >>>>
> >>>> There is at least one alternative without requiring nodejs:
> >>>>
> >>>> https://github.com/leafo/lessphp
> >>>> --
> >>>> Matthias Runge<mrunge at matthias-runge.de>
> >>>>
> >>>> _______________________________________________
> >>>> Mailing list: https://launchpad.net/~openstack
> >>>> Post to     : openstack at lists.launchpad.net
> >>>> Unsubscribe : https://launchpad.net/~openstack
> >>>> More help   : https://help.launchpad.net/ListHelp
> >>>>
> >>> _______________________________________________
> >>> Mailing list: https://launchpad.net/~openstack
> >>> Post to     : openstack at lists.launchpad.net
> >>> Unsubscribe : https://launchpad.net/~openstack
> >>> More help   : https://help.launchpad.net/ListHelp
> >>
> >> _______________________________________________
> >> Mailing list: https://launchpad.net/~openstack
> >> Post to     : openstack at lists.launchpad.net
> >> Unsubscribe : https://launchpad.net/~openstack
> >> More help   : https://help.launchpad.net/ListHelp
> 
> 
> _______________________________________________
> Mailing list: https://launchpad.net/~openstack
> Post to     : openstack at lists.launchpad.net
> Unsubscribe : https://launchpad.net/~openstack
> More help   : https://help.launchpad.net/ListHelp