PrusaSlicer "cancel object" bug?
I think I just encountered a bug in PrusaSlicer (latest) and/or the MK4 firmware (latest). I was printing two objects and about 20 minutes in had an issue with one of them so I successfully cancelled the failed object. The other object continued to print fine (oh happy day!) until that [remaining] object finished whereas the entire printer just stopped there (last spot on the remaining object) and did no more.
The details are that there were only two objects in the print job and the cancelled object was the taller of the two. So it looked like the printer did not know what to do once the shorter object completed printing since the taller object had been cancelled. The print job was at 82% completed when the shorter object finished.
Firmware: 6.2.3-8862
PrusaSlicer: 2.9.1
I am not sure if it is a slicer issue, a firmware issue or a combination of the two. As it could be that my cancelling the taller object also nuked the end-of-print gcode or something else entirely. I have used cancel-object before without issue but best I can recall the cancelled object(s) were never the "last" object (based on height).
Anyone else encounter this or have seen this scenario specifically work (cancel the tallest of two object before the other object completes) and the print job end normally?
I will see if I can replicate it when I have a chance.
Thanks.
Best Answer by John Doe:
OK, recently upgraded to 2.9.2 and have been unable to recreate this issue. So either it was a "fluke" or it was resolved in the update.
RE: PrusaSlicer "cancel object" bug?
If there is a bug it will be with the printer firmware. The slicer just generates gcode. The slicer doesn't know if any are going to be cancelled so it just generates the gcode exactly as it would to do the 'complete' print.
Its up to the printer what it does with it once it has it. The slicer obviously did its job of labelling the objects correctly otherwise you wouldn't have been able to cancel one at all.
RE: PrusaSlicer "cancel object" bug?
I agree with @neophyl that this is likely a firmware bug (though it could be incorrect/incomplete object tagging in PrusaSlicer).
Since you are seeing the bug, what I would do if I were you is:
- Create a simple project consisting of two cubes (both for quick printing and to protect any IP of the project where you first discovered this issue).
- Scale one of the cubes in z to be taller and name each one something that makes sense (like "tall" and "short").
- Save this as a 3mf.
- Slice and attempt to duplicate the issue.
- If you aren't able to duplicate, iterate until you can (maybe change the order of objects in the list, move to different sides of the plate, etc.) saving both the 3mf and gcode/bgcode for each iteration, or you run out of patience.
- Once you are able to duplicate the issue:
- Submit a bug report on the github repository for buddy firmware ( https://github.com/prusa3d/Prusa-Firmware-Buddy/issues , if you don't have a (free) github account you'll need to create one) by clicking the green "New Issue" button and selecting "bug report".
- You may want to include a link to this forum message thread in the description, but not to the exclusion of writing a complete description in the bug report. (Don't make the developer's job harder by making them follow links.)
- Include both the 3mf and the gcode/bgcode of the iteration where you were able to duplicate the issue in the bug report.
Here is a good essay on general bug reporting if you've never written one before (It is good reading even if you have written bug reports before just to be sure you aren't unintentionally writing poor bug reports): https://www.chiark.greenend.org.uk/~sgtatham/bugs.html
See my (limited) designs on:
Printables - https://www.printables.com/@Sembazuru
Thingiverse - https://www.thingiverse.com/Sembazuru/designs
RE:
I don't think this is as bug, it's normal behavior. The printer is going thru the gcode for the cancelled part and deleting it. If you watch after cancelling an object you'll notice that the print head will move over to where the cancelled object is and momentarily pause every layer before returning to the active part. So in your case since the cancelled object was taller it was running thru all the gcode for that part.
This is because there is no way for the printer to cancel all of the lines in the code that it hasn't gotten to yet.
If you had just waited it would have eventually finished.
RE:
Brian, I don't think that was it. I waited at least 5 minutes. The head stopped at the last position of the object it was printing and did not do anything else. It held there until I cancelled the print.
RE: PrusaSlicer "cancel object" bug?
OK, recently upgraded to 2.9.2 and have been unable to recreate this issue. So either it was a "fluke" or it was resolved in the update.