Latest Contributions by mgeorge

@AndrewMurdoch1 - is this the same as the other app where you are seeing the graphical bug in 4.17? Performance-wise, if you are able to use hosted feature layers (rather than query geojson -> create feature-layer from feature collection), that should perform dramatically better. With lines, server-side generalization can improve performance immensely. With the client-side feature collection, it requires loading all the data upfront and building spatial indices/querying/quantizing/re-projecting all on the client.  1 larger feature layer will also usually perform better than many small ones because each feature layer includes some upfront memory costs. ... View more

Ok great! (Just realized I posted with the wrong account 😅 - this is my official one) ... View more

Hi Stacy. Yes sorry, we did a hotfix for this.  ... View more

Yes ... it took us a while but we were finally able get this in! Let me know if you run into any issues. ... View more

Hi Tom,  For Vector Tiles, we do indeed use the Mapbox Vector Tile spec. The f=pbf format of FeatureServices however is an esri Feature Tile serialization spec. Feature Tiles are more for dynamic visualizations, whereas Vector Tiles are better for things like basemaps.  The documentation on this has been held up, but I think more details should be forthcoming. I'll check in again.  We've also worked to improve the performance of pbf encoded Feature Tiles in our upcoming 4.17 release. Whereas before we were doing some quite a bit of conversion to deserialize features, we've reworked things so that we can work directly against the compressed binary data.  Matt ... View more

4.17 is releasing very soon. This seems to be a race condition, so slight differences in timings (from misc. code changes, etc) could all contribute to whether or not this happens in particular versions.   ... View more

Sorry Andrew Murdoch‌! I would like to say yes, but this isn't the kind of thing we usually backport (that's mainly reserved for browser breaking changes that prevent earlier versions of the API from working) ... View more

Andrew Murdoch‌ I just installed a fix into our 4.17 dev branch, which seems to at least fix my repro. I'll let you know when that get's pushed to /next.  ... View more

Thanks! I think I can repro this with https://codepen.io/matt9222/pen/rNePaLW?editors=1000. Seems to be a timing issue with removing a layer while it's still loading.  This is on ubuntu 20/chrome (using an outdated chrome due to the latest version having super borked devtools on linux). Need to have devtools open + cache disabled (can repro maybe 30% of the time).  I think these should be "ok" errors as they are basically just aborts - we should be catching them though (& rethrow only if not an AbortError). Taking a look.  ... View more

Ah ok so you are getting the loop.  ... View more

You can send it to me (mgeorge@esri.com), however I think the HAR capture likely won't help us much here. We've made a bunch of changes to the way feature layers load, as well as fixed a rare issue that could happen with the workers communication, so it might be best to wait for that and see if we can still see the problem (the changes are currently live on /next).  ... View more

Oh sorry, if what I think is happening is actually what's causing the issue, it would only theoretically happen if the layer is added and removed very quickly. Just adding a feature layer shouldn't cause the problem.  ... View more

Andrew Murdoch‌ do you see any adverse affects in your apps when you see the above error (i.e., the "infinite-loop" bug in the OP), or just the error?  Likely this is a race that has to do with a layer being added/removed very quickly (probably with 100 ms of being added). When you create a FeatureLayer, that spins up a connection to a worker which communicate over a message port. The error "Can't invoke, port is closed" coming from the AttributeStore means that while the worker was trying to setup the shared memory region (which happens immediately when the connection is established), the connection was closed. We can potentially move this error behind a flag, as this likely shouldn't have any adverse consequences.  Does this sound about right for your case? Does your app add/remove a layer very quickly?  ... View more

Thanks Andrew Murdoch‌, I'll be looking out for this as well.  ... View more