Latest Contributions by mgeorge

@FC_Basson, we're currently trying to get ahold of a device with similar specs.  ... View more

For those who are experiencing the flickering, do you still see here? with: https://bsvensson.github.io/various-tests/compare-webmaps/webmap.html?webmap=87e7f204fd1f4595ada62603359ec40c&api=https://jsdev.arcgis.com/next/ ... View more

Also we do have a workaround for the driver bug on dev, and we are hoping to get it on /next tomorrow or Monday. ... View more

@MillionMap that appears to be the same driver version as posted in https://community.esri.com/t5/arcgis-online-questions/labels-are-flashing-blinking-in-map-viewer-and/m-p/1396114/highlight/true#M58110 Maybe the computer needs a restart? Or it's possible that the manufacturer for your device hasn't updated the intel drivers if IT is only installing those provided by the manufacturer instead o intel directly.  The latest intel drivers for your machine be found here: https://www.intel.com/content/www/us/en/products/sku/203682/intel-core-i910885h-processor-16m-cache-up-to-5-30-ghz/downloads.html The latest version is 31.0.101.2127 which came out 2/1/2024   ... View more

@FC_Basson in addition to providing the cpu model, can you also provide the report under chrome://gpu?  ... View more

Hey all, I think we've narrowed this down. What appears to be happening is that for intel driver versions ~30.0.101.1191, there appears to be a driver bug where read-writes to textures are not always synchronized. For some context, labels are allocated some texture space that denotes the state of the label. Before we actually render the labels, we first do a GPGPU pass that modifies the label state. We then use the result of this, (a texture) in another render pass which actually renders the labels. What I'm seeing is that sometimes this texture is not updated before it is used, which is fixed if I explicitly force a synchronization, e.g., by calling readPixels to flush it. This should happen implicitly however.  In looking at the differences between 4.28 and 4.29, there appears to be a slight ordering difference resulting in the labeling state modification pass happening slightly later. Rearranging these calls appears to circumvent the issue, & we're currently looking to see if we can install a workaround for this.  I still would highly recommend updating drivers if possible as this synchronization issue could affect other things in ways that are hard to really guess upfront.  ... View more

@MillionMap yes we were able to reproduce this only after downgrading the intel display drivers to 30.0.101.1191. It seems like there is a bug in the intel graphics drivers that affects a cluster of versions around that. We're currently seeing if there's anything we can do on our end to workaround the issue, but as @RussRoberts, upgrading your intel display drivers to the latest version should fix the issue.  ... View more

Ah ok, makes sense @duvoisie 👍 ... View more

@duvoisie when you open the link @RussRoberts posted, does it look this this?  Or instead are you noticing that the labels very slowly fade in over many frames? I ask because your machine is using software emulation & might be hitting our VM codepath. If you are able to enable hardware acceleration, does the problem go away?  ... View more

Hi all! Sorry for the issue, we are working on coming up with a fix, but unfortunately we have yet to find a device internally that we can use to reproduce the issue. Likely it's specific to particular graphics & driver versions.  For those who are able to reproduce this, are you able to provide us the chrome GPU report? Go to chrome://gpu & click download report to file. This will give us more information about your GPU (including driver version), and whether chrome is applying any workarounds. It would also be helpful if you could go to https://webglreport.com/?v=2 and see what it says under "unmasked renderer." This will help us to know whether or not the issue is specific to the integrated or discrete graphics. Oftentimes chrome will prefer the integrated graphics.  ... View more

@LeonAus in 4.22 we added support for repeatLabel, repeatLabelDistance, allowOverrun, and labelPosition (curve / uncurved). We still don't yet have support for all the different labelPlacement options for polylines yet, but we did get some new people on the team recently, so I'm hoping we'll be able to make more progress in labeling soon! Not specific to lines, but better label centroid placement is something we're working on for next release (4.28). ... View more

Sorry this popped up & I'm just now realizing this was from a while ago. Our working theory is that this might be specific just to MapServers (trying to get access to a large one internally ATM) & if that's the case, we should be able to modify the querying logic for these services specifically.  ... View more

Hmm, I'm not sure we would make this opt-in by default, as 3.x also switches to snapshot mode, just at a lower count. Or potentially we could expose different strategies depending on the server. @NealFerguson are you seeing nothing render until 40+ seconds? If you open devtools, are you waiting on a really long count query? What server & version are you using?  From the testing we did internally, snapshot can significantly improve performance both for AGOL and enterprise (up to so certain point), but if there are some servers & versions in the wild where this is not true, it would be good to add exceptions internally for them. ... View more

Hi @Koob,  We added the custom websocket input for StreamLayers really just for simple cases, with GeoEventSource for more fully featured services. Note that as part of mmgeorge/node-stream-service: Example for connection a custom WebSocket stream service to the 4.x ArcGIS JS API (github.com), there's no logic for setting geometryDefinition or definitionExpression specified. We probably need to add to the known limits that this is unsupported for custom websockets (basically undefined behavior currently -- FYI, in some cases filter can actually be a string which is probably why you are seeing that code block not make any sense).  What is your use case here? Are you planning on changing the definitionExpression / geometryDefinition at runtime, and your custom websocket server can handle this?  ... View more

Great! Glad to hear it's working out for you.  ... View more