[OpenLayers-Dev] Motion: 2.5 Final Release

Christopher Schmidt crschmidt at metacarta.com
Sun Oct 7 22:24:36 EDT 2007

There have been no new regressions reported in the OpenLayers 2.5
release in the time since the release of RC5. 

There is currently one outstanding issue that Tim had marked for 2.5 --
an improvement to GeometryCollection handling in GeoJSON parsing -- but
it isn't a regression, and the use case where it actually affects people
is very small. I'm of the opinion that since the GeoJSON spec is not yet
'done' -- there still could be more changes to it -- we shouldn't hold
the release for another edge case lack of support: instead, if we really
do want this into something we call 2.5, I'd be in favor of pulling it
back and doing a 2.5.1 when the spec is complete. (See
http://trac.openlayers.org/ticket/1067 -- this lack of funtionality only
affects the case where a feature is passed with a GeometryCollection as
the geometry.)

With that being the only outstanding issue, I'd like to motion that we
release OpenLayers 2.5 on Tuesday afternoon eastern time -- in about 40
hours -- unless we hear anything new in terms of bug reports in the
meantime. This release would not include the fix for #1067. (If the PSC
hasn't voted by that time, I propose that the release be made as soon as
the PSC has voted.)

I'm +1 on doing this, and will do the release engineering work if there
are votes against releasing.

Additionally, I'd like to  put #1067 in trunk, and after the GeoJSON
specification has been finalized, we pull any changes to the GeoJSON
format back into a 2.5.1 after sufficient testing. The change that we're
looking at is, imho, too risky to put into a 2.5 release this late in
the game -- I screwed that up once already ;) (Hence the RC5 instead of
RC4.) I'm interested in what the developer community thinks about doing 
this specifically to support full GeoJSON. 

The biggest reason for supporting the GeoJSON spec so strongly is that
it is a format that we can round trip effectively, unlike many other
formats. The simple feature model and limited geometry model allow us to
fully support GeoJSON input/output in OpenLayers, and that's important
for people to be able to have/understand.

Looking forward to feedback on either of these two items, and would like
to get the PSC to vote on the 2.5 release when they get a chance. 

Christopher Schmidt

More information about the Dev mailing list