tag.previous_sibling and tag.find_previous_sibling() return different results

[Alex Krupp]

text = "<br>a<div>b</div>"
soup = BeautifulSoup(text, "lxml")

soup.div.previous_sibling
> 'a'

soup.div.find_previous_sibling()
> <br/>

[Leonard Richardson (leonardr)]

Thanks for filing this ticket. You've found a case where the defined behavior is confusing, but I don't think it's worth changing.

.previous_sibling is a fact about the parse tree. find_previous_sibling() is a method call that invokes all the logic of the find* methods. The default behavior of the find* methods is to look only for tags. That's why find_previous_sibling() with no arguments finds a tag.

It would also be reasonable if a find* method invoked with no arguments found _everything_, but that would break existing code, and if you want to find everything, you can use the generators, e.g. previous_siblings.