Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [rt-pmc] Re: Virgo

Glyn Normington wrote:
> Hi Greg
> 
> On 14 Jan 2010, at 00:14, Greg Wilkins wrote:
> 
>> Glyn Normington wrote:
>>
>>> That said, there is no technical reason why an alternative Jetty-based server should not
>>> also be offered by Virgo. Like the proposal says this should be a fairly straightforward
>>> exercise. I am very happy to provide the necessary pointers to get the Jetty committers started.
>> Glyn,
>>
>> Surely this effort should be undertaken by the Virgo (and/or Gemini) teams.
> 
> For Gemini and Virgo to support Jetty, two thin adapters need to be written which will 
> require understanding of Gemini/Virgo and Jetty, so I would expect there to be collaboration
> between the teams, although who does what remains to be seen. The adapters will reside in
> Gemini and Virgo, so will be maintained by Virgo committers.


Glyn,

thanks for that statement!  That was what was missing from the proposal and was
giving us visions of having to maintain the work ourselves.

Now if only you can say the same in the virgo proposal under the jetty section,
then my concerns would be substantially addressed.


> The servlet container, OTOH, has significant externals, particularly configuration, 
> which are visible as externals of dm Server. So Virgo needs to support Tomcat,

I'm certainly not suggesting that tomcat support be dropped, or that even Jetty is
made the default.    I just want the drivers for jetty integration to come from
within the gemini/virgo projects - so they will pull our help rather than have
have us push it later (I guess we are pushing it now :)


> and 
> continue to do so, in order to serve the needs of the existing dm Server user community.
> Regardless of who does the work, adding the Jetty adapters should IMO be a follow-on
> activity rather than a pre-req to donation.

I don't believe the work needs to be done pre donation, but I do believe a commitment
to have such a module pre first release should be made as part of the project proposal.

Post donation, I think there will be lots of work for you to do, and I don't want
to see integration with other RT projects de prioritized and the need for non-virgo
committers to convince virgo committers that such a module is needed.

If the adaptors are truly thin and easy, then I don't see it as very onerous to
commit to providing them pre first release.

cheers







Back to the top