[SKiDL BUG] symbol ends with _0_0

[shanzez]

Describe the bug Did you mean to filter out '2'?

Symbols that end with _0_0 are being filtered out (conversion flag of '0') and causing execution to fail like:

Traceback (most recent call last):
  File "/temp/gen2/boards/main.py", line 139, in <module>
    usb_connectors['MT1', 'MT2', 'MT3', 'MT4'] += gnd # todo - connect to chassis?
TypeError: unsupported operand type(s) for +=: 'NoneType' and 'Net'

Conversion flag: _1 means use 1st _2 means use 2nd _0 means use either

Many KiCad symbols use this and cannot be imported because of this.

To Reproduce Steps to reproduce the behavior:

skidl.Part('CM4IO', 'USB_67298-4090', footprint='todo')

Expected behavior No errors during execution.

Desktop (please complete the following information):

• Debian Linux Buster
• Python version: 3.9
• SKiDL version: 1.0.0

Additional context Comment out these lines (854 and 855) in kicad.py and you can execute successfully:

 853         # Don't add this unit to the part if the conversion flag is 0.
 854         #if not conversion_flag:
 855         #    continue

[devbisme]

I'm not sure what this means:

Did you mean to filter out '2'?

I don't have any official info on V6 schematic libs except for the two references in my code. The V6 library format doc doesn't really mention these suffixes appended to the symbol-unit names. And the source code is nearly devoid of comments. So I made most of my decisions based on just looking through the V6 symbol libraries. If you have more solid or detailed information, I would like some pointers to it.

[shanzez]

Sorry about that.

Did you mean to filter out '2'?

Did you mean to filter out conversion flag 2?

[shanemmattner]

I ran into this issue as well when I imported a part to a custom library. Thanks @shanzez for figuring out what's going on. After changing AD8137YRZ_0_0 to AD8137YRZ_0_1 the new part worked fine. Here is the original and modified kicad_symbol file AD8137YRZ.txt AD8137YRZ_working.txt

[devbisme]

I'll look into it.

[dehekkendekrekker]

I ran into a problem of which this is the root cause. I think the last number in NAME_x_x should not be regarded as a conversion flag, but as a sub unit number. I've noticed that some part definitions in kicad symbol files don't store pin information in this section, but for instance 74xx.kicad_sym does.

See part 74LS08 for instance, this has section 74LS08_5_0 holding information about the power pins.

I've created a PR that should fix the issue, but please review.