Latest Contributions by JoelBennett

Ok, so in the case where it doesn't zoom properly, you're having the view go to an extent whose xmin is -180 and xmax is 180.  This is impossible to display on a globe, so it zooms out to show as much of the globe as it can, and centers at the center of the extent, which happens to be the Prime Meridian (where x = 0).  The solution is to detect cases that wrap around the International Dateline, and build an extent whose xmin is on the left of it, and whose xmax is on the right of it.  This means the xmax will be greater than 180, but that's ok, because the API is capable of handling wrapping. To start with, you can remove our function from previously called "fixExtent".  Then you can add this function in its place: function adjustExtent(extent) { if (extent.xmin < 0) { extent = extent.clone(); extent.xmax += 360; extent.xmin += 360; } return extent; }   Then, we change the main code to this: var isSplit = false; var extents = []; var extent; for (var i = 0; i < objCoordinates.length; i++) { //etc... extent = polygon.extent; if (i === 0) extents.push(extent.clone()); else { extents.push(extent); if (Math.abs(extent.xmin - extents[i-1].xmax) > 90) isSplit = true; } } if (isSplit) { extent = adjustExtent(extents[0]); for (var x = 1; x < extents.length; x++) extent.union(adjustExtent(extents[x])); } else { extent = extents[0]; for (var x = 1; x < extents.length; x++) extent.union(extents[x]); } this.view.goTo(extent).catch(this.catchAbortError);   In the first (main) loop, we capture all the extents, and see if they're split across the dateline.  Afterwards, we loop through them again, and process them according to whether or not they're split across the dateline.  If they are, the ones the right side of the dateline are wrapped around the world once so their x values are greater than 180.  This will enable us to produce the smallest extent that crosses the dateline. ... View more

Actually, it wasn't terribly difficult to find out which ones aren't supported.  The following CSS cursors don't work with view.cursor, but do with view.container.style.cursor: none wait cell text vertical-text alias no-drop not-allowed all-scroll col-resize row-resize zoom-in zoom-out url-based (custom) cursors, e.g. "url(https://myServer/myCursors/cursor1.cur),auto" Most of those nobody would probably ever use, but four of those (one being the url-based) are found in my applications.   ... View more

If the 'x' values of your polygons range from -180 to 180, then it's WGS 84...if they're much larger, then it's almost certainly Web Mercator.  I have a workaround in mind, but need to know which coordinate system it is. ... View more

What coordinate system are these geometries using?  Is it WGS 84, Web Mercator, or something else? ... View more

As stated previously, "We should clone the extent because we plan to modify the values of that extent object".  In the first pass of the loop, we clone the first polygon's extent, because in each subsequent pass, that object will be modified in place by the union function. ... View more

Thanks for going through the trouble of creating and posting that.  It turns out it works just fine for me, so I went back to the example I tried before my first post, and found it working, so I was a bit dumbfounded.  But then I realized I used a different cursor on this try (progress) than previously (wait).  So then I tried it with the "wait" cursor, and it doesn't work in my example or yours. So in your example, if you change it to: view.cursor = "wait"; it doesn't work, but if you change it to: view.container.style.cursor = "wait"; it does work.  It makes me wonder how many cursors behave like this?  I don't really have the bandwidth to figure it out though, so I'll probably just stick with the latter approach... ... View more

Although there is no documentation I know of regarding this, you can "opt out" of this behavior.  The short of it is that you would add this in a script tag prior to loading the API: var dojoConfig = { has: { "featurelayer-snapshot-enabled": 0 // disable snapshot } };   This is derived from this thread, at the end of the solution post.  For reference, these settings are found in the esri/core/has.js file. ... View more

Hmm...I'm writing directly in JavaScript with AMD; are you using a different approach? ... View more

Although the bulk of that is correct, there is no documented "cursor" property for the objects that inherit View, and trying to use it makes no difference for me.  I've always had to use view.container.style.cursor, e.g.: view.container.style.cursor = "progress";   ... View more

Ok, this makes it somewhat difficult, because the full width of the calculated extent is the full width of the coordinate system itself.  The map therefore centers at the Prime Meridian. Does your dataset include data throughout the world, or just the US and the Pacific? ... View more

We should clone the extent because we plan to modify the values of that extent object, and we eliminate the potential of those changes causing problems for the owning polygon object.  (In this particular case, it probably wouldn't matter, but it's just good practice.) In the line: extent.union(polygon.extent); only "extent" is modified, not "polygon.extent", so there is no need to clone "polygon.extent". ... View more