[OpenLayers-Users] IE strategies

[Yves Moisan]

> My bad. It should read
> new OpenLayers.Strategy.BBOX({invalidBounds: function(){return true;}})

Well, I was dumb enough to stick it in the Cluster() strategy of my
2-strategy layer on my side of the story :-(.  At any rate, I can
confirm that an XHR is done on zoom in now.

I still have issues though.  One peculiar thing is that for some reason
the BBOX issued by OL to the server is consistenly larger than the one
specified in map.zoomToExtent.  Here are numbers (HTML =
map.zoomToExtent(new OpenLayers.Bounds()); and "Output BBOX" is sent out
to the server) :

Case 1 : specified CSS for map <div> : w X h = 512 X 256

HTML : 26.1925,-10.5710,26.1935,-10.5705

Output BBOX :
26.191626708984373,-10.571436645507813,26.194373291015623,-10.570063354492188

HTML = previous output BBOX =
26.191626708984373,-10.571436645507813,26.194373291015623,-10.570063354492188

Output BBOX :
26.190253417968748,-10.572123291015625,26.195746582031248,-10.569376708984375

HTML = previous output BBOX =
26.190253417968748,-10.572123291015625,26.195746582031248,-10.569376708984375

Output BBOX :
26.187506835937498,-10.57349658203125,26.198493164062498,-10.56800341796875

Case 2 : Map <div> CSS unspecified (defaults to screen width and some
height)

HTML = 26.1925,-10.5710,26.1935,-10.5705
Output BBOX :
26.18567756938934,-10.572552385597229,26.200322430610655,-10.56850761440277

HTML = previous output BBOX =
26.18567756938934,-10.572552385597229,26.200322430610655,-10.56850761440277
Output BBOX :
26.178355138778684,-10.574574771194458,26.20764486122131,-10.566485228805542

At first, I thought this had to do with me specifying BBOX values that
didn't quite match the width/height of my div but taking the output BBOX
and shoving it back into the HTML (which I thought to myself must be
"OL-compliant" as it generated it) still shows a zooming out behaviour.

I then figured there might be an interference between the two strategies
so I tested with BBOX + Cluster and BBOX separately.  Same behaviour.

There are also issues with the masfeatures parameter : a small number of
maxfeatures might bring in less features than maxfeatures even though
the number of features definitely is more than maxfeatures.  To
illustrate results comparing BBOX alone and Cluster, see :

http://www.flickr.com/photos/30818159@N05/3182451399/ shows BBOX +
Cluster (maxFeatures = 200) on the left and BBOX alone on the right.
Both return 101 features.  Good

http://www.flickr.com/photos/30818159@N05/3182451587/ shows BBOX +
Cluster maxFeatures = 200 on the left (results in 101 features) and
maxFeatures = 10 on the right (results in 9 features ???).  There
clearly should be 10 features returned.

Same for BBOX alone (the 9 features are apparent on the right) :
http://www.flickr.com/photos/30818159@N05/3182451511/

Thanx to you and Chris,

Yves