Element.itertext() works differently in C and Python implementations

[serhiy-storchaka]

In the Python implementation it returns empty iterator for Comment and ProcessingInstruction. In the C implementation it emits their content.

>>> import xml.etree.ElementTree as ET
>>> list(ET.Comment('comment').itertext())
['comment']

cc @scoder

Linked PRs

• gh-149571: Make Element.itertext consistent for Comment and ProcessingInstruction #149714
• gh-149571: Fix the C implementation of Element.itertext() #149929
• [3.15] gh-149571: Fix the C implementation of Element.itertext() (GH-149929) #150509
• [3.14] gh-149571: Fix the C implementation of Element.itertext() (GH-149929) #150510
• [3.13] gh-149571: Fix the C implementation of Element.itertext() (GH-149929) #150511

[serhiy-storchaka]

Let's compare with lxml.

>>> import lxml.etree as ET
>>> e = ET.Element('tag')
>>> e.text = 'text'
>>> e.append(ET.Comment('comment'))
>>> list(e.iter())
[<Element tag at 0x7769b3723540>, <!--comment-->]
>>> list(e.itertext())
['text']

>>> import xml.etree.ElementTree as ET
>>> e = ET.Element('tag')
>>> e.text = 'text'
>>> e.append(ET.Comment('comment'))
>>> list(e.iter())
[<Element 'tag' at 0x7769b368a610>, <Element <function Comment at 0x7769b371b7f0> at 0x7769b3746520>]
>>> list(e.itertext())
['tail', 'comment']

So I think the correct behavior is to skip comments and processing instructions.