G1 E GCODE is getting generated without a whole number

[cjackson234]

Version

2.4 RC

Operating system type + version

Windows 10 19042.1348

3D printer brand / version + firmware version (if known)

Voron Switchwire - Klipper v0.10.0-184-gdd714fc7

Behavior

G1 E GCODE is getting generated without a whole number
For instance
G1 E0.8 F2100 outputs as G1 E.8 F2100

• Steps needed to reproduce the problem
  Slice any file
• Expected Results
  G1 E0.8
• Actual Results
  G1 E.8
  • Screenshots from PrusaSlicer preview are preferred
    

Is this a new feature request?
no

[neophyl]

#6996

[lukasmatena]

@cjackson234 Does it cause any issues? Does your Klipper+Voron combination not process it correctly?

[cjackson234]

Klipper immediately throws an error and the print fails

[bubnikv]

Klipper immediately throws an error and the print fails

If that is the case, would you please open a Klipper issue?

[cjackson234]

If that is the case, would you please open a Klipper issue?

Sure, but they will correctly point out that the problem doesn't exist in other slicers.

[bubnikv]

But they don't conform to G-code standard.

[cjackson234]

I will reach out to them now.

Is this something that could be patched in the meantime by adding an additional gcode flavor?

[bubnikv]

We are not going to fix it for PrusaSlicer 2.4.0. We will fix it after the release.

You can write yourself a simple post processing script, even a simple

sed 's/  / /g'

will likely work

Edit: Sorry that was meant to #7503

[cjackson234]

I will test that while I wait on their response. Thanks

[bubnikv]

We are not going to fix it for PrusaSlicer 2.4.0. We will fix it after the release.

Sorry guys, this response was meant to #7503

[bubnikv]

Klipper immediately throws an error and the print fails

Which klipper.

Frankly I am printing with PrusaSlicer on Prusa Mini with Klipper and I did not experience any issue.

[bubnikv]

@th33xitus

Why did you change that behavior in the first place? No criticism, just wondering.

We optimized G-code export, you will enjoy a significantly faster G-code export now. Along the way we optimized for the file size while conforming to the G-code specification. I am frankly pretty surprised if there was a number parser that cannot parse such a number.

[cjackson234]

Klipper v0.10.0-184-gdd714fc7

What version are you using?

[dw-0]

@bubnikv Sorry, yeah i saw that in the release notes of 2.4.1-alpha2, that's why i deleted my comment. Should have looked first.

Im just printing g-code generated by the new 2.4.1-rc without any issues. My current Klipper version is v0.10.0-169-g520273e5.

[cjackson234]

@bubnikv Sorry, yeah i saw that in the release notes of 2.4.1-alpha2, that's why i deleted my comment. Should have looked first.

Im just printing g-code generated by the new 2.4.1-rc without any issues. My current Klipper version is v0.10.0-169-g520273e5.

Very interesting. Would you mind sharing additional details? Are you using octoprint or fluidd/mainsail?

[dw-0]

Im using Mainsail.
I will update my Klipper to newest and try again.

[bubnikv]

I am using
https://github.com/raymondh2/MainsailOS
https://github.com/matthewlloyd/klipper.git
the klipper for Prusa Mini was synced with upstream around 5 months ago.

[dw-0]

I just updated to Klipper to v0.10.0-184-gdd714fc7.
Working fine... no issues on my side. Same g-code as before.

[cjackson234]

Closing the issue for now as other users were not able to reproduce. I'll post back here when I find a solution.

[cjackson234]

This is a configuration issue. How variables set has changed. In my print start gcode was parsing a value incorrectly. That caused all subsequent G1 E commands to fail. Oddly, manually changing the compiled gcode to start with a whole number fixed the issue for me.

That being said, thank you for looking into this.

[ledvinap]

You can add zeroes with

sed -e 's/\([eE]\)\.\([0-9]\)/\10.\2/g'

[kevinmichaeldudley]

My "day job" is developing and marketing CNC software for 35 years.
I have over 1000 users running "god knows how many different CNC machines".
FYI: The fact is, in the CNC world, not all machines will accept "no leading 0".
Is it "not to standards" possibly, but in the real world we make OUR software to WORK on everything, as I believe the slicer gcode output should also.
JMHO