Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Weird artifacts with arcs in some paths

See original GitHub issue

Hi, I’ve run into an issue, where an arc segment seems to go missing when I convert a parsed path back to a string with d(). Here’s a visual example of what is happening:

The left side is the original path and the right side is what I got by running the path through parse_path("...").d(). It looks like an arc segment is just left out.

To understand better what’s going on, here are the paths that create the two hearts above:

<!-- Original -->
<path
	d="
	m 1672.2372,-54.8161
	a 14.5445,14.5445 0 0 0 -11.3152,23.6652
	l 27.2573,27.2572 27.2572,-27.2572
	a 14.5445,14.5445 0 0 0 -11.3012,-23.634
	a 14.5445,14.5445 0 0 0 -11.414,5.4625
	l -4.542,4.5420
	l -4.5437,-4.5420
	a 14.5445,14.5445 0 0 0 -11.3984,-5.4937
	z"
	fill="none" stroke="#FF423F" stroke-width="2" stroke-linecap="square"
/>
<!-- Parsed -->
<path
	d="
	M 1772.24,-54.8161
	A 14.5445,14.5445 0 0,0 1760.92,-31.1509
	L 1788.18,-3.8937 L 1815.44,-31.1509
	A 14.5445,14.5445 0 0,0 1804.14,-54.7849
	A 14.5445,14.5445 0 0,0 1792.72,-49.3224
	L 1788.18,-44.7804
	L 1783.64,-49.3224
	Z"
	fill="none" stroke="#FF423F" stroke-width="2" stroke-linecap="square"
/>

The final arc before Z is clearly missing for some reason.

Also here’s what I ran on the python prompt to get the result above:

>>> p = parse_path("m 1772.2372,-54.8161 a 14.5445,14.5445 0 0 0 -11.3152,23.6652 l 27.2573,27.2572 27.2572,-27.2572 a 14.5445,14.5445 0 0 0 -11.3012,-23.634 14.5445,14.5445 0 0 0 -11.414,5.4625 l -4.542,4.5420 -4.5437,-4.5420 a 14.5445,14.5445 0 0 0 -11.3984,-5.4937 z")
>>> p
Path(Move(to=(1772.2372-54.8161j)), Arc(start=(1772.2372-54.8161j), radius=(14.5445+14.5445j), rotation=0.0, arc=False, sweep=False, end=(1760.922-31.1509j)), Line(start=(1760.922-31.1509j), end=(1788.1793-3.893699999999999j)), Line(start=(1788.1793-3.893699999999999j), end=(1815.4365-31.1509j)), Arc(start=(1815.4365-31.1509j), radius=(14.5445+14.5445j), rotation=0.0, arc=False, sweep=False, end=(1804.1353-54.7849j)), Arc(start=(1804.1353-54.7849j), radius=(14.5445+14.5445j), rotation=0.0, arc=False, sweep=False, end=(1792.7213-49.3224j)), Line(start=(1792.7213-49.3224j), end=(1788.1793-44.7804j)), Line(start=(1788.1793-44.7804j), end=(1783.6356-49.3224j)), Arc(start=(1783.6356-49.3224j), radius=(14.5445+14.5445j), rotation=0.0, arc=False, sweep=False, end=(1772.2372-54.8161j)), closed=True)
>>> p.d()
'M 1772.24,-54.8161 A 14.5445,14.5445 0 0,0 1760.92,-31.1509 L 1788.18,-3.8937 L 1815.44,-31.1509 A 14.5445,14.5445 0 0,0 1804.14,-54.7849 A 14.5445,14.5445 0 0,0 1792.72,-49.3224 L 1788.18,-44.7804 L 1783.64,-49.3224 Z'

EDIT: I did some testing and it seems like this is what ommits the last arc: https://github.com/regebro/svg.path/blob/master/src/svg/path/path.py#L434

Not sure what a real fix should be like, but I can probably work around it at least in some cases, by just doing p.closed = False before p.d().