Has anyone else been experiencing wait/kill errors lately? It seems to have started right around the time I pulled a version that went with the server.js instead of the old server-smoothie.js
It pretty much locks up the tab LW3 is running on and I can’t even do real-time speed or power adjustments.
However Chrome itself is not locked up as I’m posting this on a second tab while the job is still running.
I’m on Chrome Version 54.0.2840.100 (64-bit)
Linux Ubuntu 16.04.1 LTS
What DPI is that raster image?
That one was 150 but I noticed it Sunday with a smaller 90dpi image as well.
Hm. I never got that on Windows. It sounds like a javascript loop or memory leak somewhere. This should be debuggable in the browser. I will try to find it. I someone could tell me how to replicate that problem on Windows, that would help.
I do not have any issues with Windows, and I have run some huge rasters through LW.
Am I the only one running LW on Linux? I guess I can install it on my windows laptop and see if I can replicate it.
I’ve had that issue with my Mac. It is usually preceded by LW being totally frozen (even though the job is running.)
Yeah, that’s what the wait/kill message is for. It recognizes that your session has essentially frozen and offers you the option to continue to wait it out or kill the running process.
The chrome tab actually locks up with the laser indicator showing in whatever spot it was on last update. Then it may at some point update with a snapshot of further into the job and then lock up again. Once the job is complete, it eventually frees up the tab again.
An interesting note, I actually closed the browser mid-job during one of these to try to get the job to stop and the job actually worked through to completion. Tested that again and as it continued the only way to actually stop the job was to go ctrl+c to break out of the node session in terminal.
The whole gcode of the job is sent to the node server on job start, the client is not needed for the rest and does only get the position messages to update the dro. I don’t see any heavy usage of the browser in that state.
Definitely happening on all jobs now. How can I pull an old revision to test? I’d like to go back and test just before the merge into the single server.js to compare.
I would check your browser plugins first. It could be one of those.
None installed that I’m aware of. This is a Linux laptop dedicated only to the laser. Worked fine until 2 pulls ago. There has to be a way for git to let me pull a prior revision.
If you want to load an older version of LW, you can check the commits on Gthub and copy the SHA-key of the wished commit. Then do a “git checkout SHA-key”.
Cool. Will give it a try tomorrow night and see if it clears the problem. Thanks!
