[OpenLayers-Dev] 2.5 Tickets

[Christopher Schmidt]

Devs -

3 patches need review for 2.5:
 * Support W3C Geo in Format.GeoRSS. An oversight, but an important one 
   to maintain backwards compatibility for people wanting to switch from
   Layer.GeoRSS to Format.GeoRSS
 * GeoJSON Draft4 Support. Updating GeoJSON draft to revision 4, based
   on talks at FOSS4G.
 * KML should automatically determine NS. This bug has caused several
   reports of misbehavior already, and so although it's not a 'bug' per
   se, I'd like to get it fixed in 2.5.

2 patches are awaiting pullups: 
 #1023, add ServiceVersion option to TMS, is as close to 0 risk as 
   you can get, and will better support gdal2tiles.
 #1040 fixes a mistake in the spherical-mercator example.

1 patch awaiting commit:
 #1013 , OpenLayers.Layer.WMS.Untiled doesn't clone correctly.
         (Regression) 

2 patches needing work:
 #1034, OpenLayers.Util.pagePosition occasionally fails in IE.
 #1043, OpenLayers zooms in too far on an extent

1043 and 1034 are both blockers for the release, as they are regressions
from previous behavior. The three currently awaiting review would be
nice to have for 2.5, but I'm willing to let it go without it. I'm
willing to do the legwork to pull those in, and I'm convinced they are
low risk. all three have tests. If I could get a reviewer on those, it
would be great.

Bart, can you share the HTML which gave you the error for #1034?

Tim, can you comment on what you think is the best path for #1043? I
think that we need to go to a threshold-like setup, where we pull the
closest resolution only if it's within 5% of the total resolution or
something -- something which is 40% away should zoom out one more step,
and we've broken that behavior. (For more detailed information, grab me
on IRC or AIM, and I'll show you a test case or 4.)

If we can get those resolved, I'll do my best to get another 2.5 RC out
24 hours after we get all the bugs fixed in trunk.

If you have any negative feedback on the changes in #1024, #1028, and
#1044, let me know. The former are clearly relatively harmless and
likely helpful -- the last is a purely selfish move so that we don't
constantly deal with problems with KML files from different Namespaces
floating by. However, I'm willing to be convinced on any of them that an
RC is not the right time.

Best Regards,   
-- 
Christopher Schmidt
MetaCarta