[gdal-dev] [PSC] Integration of GSoC code : RFC or not ?

Jukka Rahkonen jukka.rahkonen at mmmtike.fi
Wed Jul 23 06:09:59 PDT 2014


Wolf Bergenheim <wolf+grass <at> bergenheim.net> writes:

> 
> 
> I'm not PSC but, since I have some experience in GSoC I thought I should
speak up.
> The fact that GDAL accepted Mikhail as a GSoC student sort of covers the
RFC part. Because there he proposed his project and the PSC should have
voted for or against this project if not wanting it in. In any case the
Project proposal can be copy-pasted into an RFC in case you do want to have
an RFC for the record.
> 
> 
> Also it's sort of implied in the GSoC program that the aim is to add code
to the trunk of a project, unless stated otherwise. And to me this seems to
have been the goal of this project too.
> So that being said, as long as the code meets the expected standards and
it should as long as the mentor is doing his job, I don't see any obstacles,
RFC is already technically written and so has the code, and the assigned
mentor should maybe give a vote of confidence, and that should sort of be
it. Personally I don't see a reason to vote if the Mentor is for merging
Mikhail's code.

I think that with larger projects like GDAL or Geoserver (which I follow as
a member of PSC) it is not reasonable to think that the code from GSoC
projects could generally go directly into trunk. Students are supposed to
program something useful in a relatively short time and taking care of all
the extra stuff that is needed before code can be accepted into trunk can be
out of the scope of GSoC and students may not be able to do it even with the
help from their mentors. The code in trunk must not break anything and  it
should compile not only now but also in the future which means that someone
must take the responsibility of maintaining the code. Geoserver is a bit
different project with lots of pluggable modules but the Community module
policy is worth reading
http://docs.geoserver.org/stable/en/developer/policies/community-modules.html.

I would say that editing the GSoC project plan into RFC and making the PSC
to vote if the code meets the standards by quality, license and other IP
stuff, and future maintainability would be a good idea and re-usable in the
future. 

-Jukka Rahkonen-








More information about the gdal-dev mailing list