I get these kinds of plotting errors in a 7.99 nightly from august 12 on fedora 38:
If I choose “print” instead, U6D prints correctly the same as U6C.
I get another issue when printing, though - the page frame is not centered on the page, such that the very right end of the drawing (frame on the right side) will end up outside of the paper…
I think I have seen this earlier, but thought it got fixed…
Does someone know if there was something done about it, or remembers where to find previous mentions of similar issues? I’m not quite sure what phrases or words to search for…
That was about zero width / opacity lines, it seesm. Here it looks like there is an issue with the 180 degree arc / half circle going the wrong way sometimes, but only on the U6D and not on units A-C.
There were numerous issues regarding arcs and plotting/printing in the past. (search strings: arc, reverse, backwards). Closed gitlab-issues for instance: 11135, 12465, 12502,
As far as I know there are no issues open regarding this topic.
If you attach the project I could check with todays nightly.
I can see the reversed arc also.
It seems restricted to unit D, but this is a false conclusion.
The false arc is affected by the mirror command, every unit which is mirrored around the y-axis “flipped from top to bottom” is falsely plotted.
I was not able to reproduce with the current v7 version, so it seems this is a re-introduced regression.
To me the “mirror around Y-axis” seems counterintuitive, but the icon below is what I have done. It has the tooltip “Flips selected item from top to bottom”.
What is your KiCad version? Would you mind copying it here for me to include in a bug report, as I don’t have the latest one?
(If it is a lot of work for you / you don’t have the computer at hand at the moment, I can later install on another computer and take it from there, but in case it is easy for you it speeds up the process… )
The commit above solved the specific case of plotting the flipped NAND gate, but introduced new regressions in plotting of existing schematics as described in a comment to the gitlab issue linked by me a few posts above.
A second commit to this issue seems to have fixed the new regressions too, as far as I can see from testing my own example as well as those given in #11135, #12465, #12502 and #15469:
Hopefully both displaying the schematic on screen as well as plotting and printing now works properly.
Testing and reporting any potential errors still present is encouraged.