Back to building my 3d printer. Fixed some bugs with my 3dviewer on the pi. Found some new bugs too. No show stoppers, and I got a bunch of aluminum cut flawlessly.
Looking to roll your fixes out soon to everyone!
@jlauer there is a type of gcode for circles, or something of that nature that isn’t rendering correctly. I need to dive in and see what’s missing logic wise.
I’m pretty sure I know the block of code that it’s in (as there was a haxx I had to do with the new buffer geometry stuff)… gonna review tonight once the kids are in bed
G2 and G3 arcs would be my guess. It doesn’t render but it runs fine because CP doesn’t manipulate gCode from file to controller.
Guess how many times we’ve had this inquiry… you’re doing good work!
@raykholo welp. Gonna take a stab at it tonight I think
I should say: “it doesn’t render correctly”. I’m pretty sure something renders - imagine a corner fillet, you’d either get 2 lines forming a sharp corner, or a diagonal.
@raykholo good to know. I was going to see if it was a new bug I introduced or long time thing. Hopefully I can hack in a solution
Search this group for g2, G3, arc, render, and so on.
well… that was easier to find than expected:
G19 G3 Y35.018 Z0. J0.305 K0.
Changing planes then ignores the rest of the line. Im going to stop ignoring this and attempt to parse the rest of the line correctly
BOOYAH. ok… got that working with a very simple hack. I think i’d like to revisit that at some point. I now need to track down these little nuggets floating in the ether.
missing/deleted image from Google+
ahh damnit… in a live workspace G17s sometimes cause the call stack to recurse… my guess is the gcode widget is calling the reloading the gcode in the viewer… gonna have to do this a little more elegantly… fresh loading works most of the time, reloading the page causes it to poop out.
Before:
missing/deleted image from Google+
After:
missing/deleted image from Google+
My phone buzzed with all the github emails 
@jlauer 3d viewer and /chilipeppr is more your territory, will you be handling these PR’s?
Yes sir
well… simple fix just went up and it seems to work: https://github.com/dchote/widget-3dviewer/commit/0d836d70c3fa54e65301a2df39535fbe8cf1f755#diff-1458bc3eb855dd3b40a27bc171a5aed9
I think the issue may have been the gcode viewer messing with the case… or something of that nature. I’m going to review tomorrow. Also, I think I’ll merge in Mitch Bradley’s pull request #13 on mainline 3dviewer. I like some of the changes he made. We may want to grock these plane shift, extra gcode things up higher, where he made some changes, and then simplify the call stack as he did.
Ok. I just pushed up the Mitch Bradley merge in my branch. Took a lot to get it working right. Its live on my workspace: http://chilipeppr.com/dchote
Tweening now works, i guess they changed something in the latest tween lib that broke my usage.
Github repo is updated: https://github.com/dchote/widget-3dviewer
missing/deleted image from Google+
@Daniel_Chote just checked your workspace and it looks like the flyout menus don’t work in the Gcode widget because it’s getting bad/no data back from 3D viewer? That means likely something broke with the 3D viewer’s pubsub on retrieving Gcode data per line. The 3D viewer figures out XYZ values for each line so that you can move to that position when you mouseover in the Gcode widget. Each time you mouseover a line in the Gcode widget, it asks the 3D viewer for the meta data for that line. If you open your debug console you’ll see a Javascript error each time.