xml.etree.ElementTree findall() fails when using custom TreeBuilder

[jackjansen]

BPO	34941
Nosy	@jackjansen, @scoder, @serhiy-storchaka, @miss-islington
PRs	bpo-34941: Fix searching Element subclasses. #9766 [3.7] bpo-34941: Fix searching Element subclasses. (GH-9766) #9867 [3.6] bpo-34941: Fix searching Element subclasses. (GH-9766) #9868
Files	testetns.py

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/serhiy-storchaka'
closed_at = <Date 2018-10-14.08:41:19.590>
created_at = <Date 2018-10-09.09:32:14.985>
labels = ['expert-XML', '3.8', 'type-bug', 'library', '3.7']
title = 'xml.etree.ElementTree findall() fails when using custom TreeBuilder'
updated_at = <Date 2018-10-14.08:41:19.589>
user = 'https://github.com/jackjansen'

bugs.python.org fields:

activity = <Date 2018-10-14.08:41:19.589>
actor = 'serhiy.storchaka'
assignee = 'serhiy.storchaka'
closed = True
closed_date = <Date 2018-10-14.08:41:19.590>
closer = 'serhiy.storchaka'
components = ['Library (Lib)', 'XML']
creation = <Date 2018-10-09.09:32:14.985>
creator = 'jackjansen'
dependencies = []
files = ['47859']
hgrepos = []
issue_num = 34941
keywords = ['patch']
message_count = 5.0
messages = ['327385', '327388', '327695', '327696', '327698']
nosy_count = 5.0
nosy_names = ['jackjansen', 'scoder', 'eli.bendersky', 'serhiy.storchaka', 'miss-islington']
pr_nums = ['9766', '9867', '9868']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue34941'
versions = ['Python 3.6', 'Python 3.7', 'Python 3.8']

[jackjansen]

When using a TreeBuilder with a custom factory (returning a subclass of Element with some extra functionality) there is an issue with findall().

XPaths using an indexed predicate no longer work: the findall() result is always empty. The attached test script shows the problem.

I have analysed this as far as finding that the issue is in ElementPath.prepare_predicate.<locals>.select(), the last one. Specifically, the statement
elems = list(parent.findall(elem.tag))
always returns an empty list (if the Element class is subclassed).

When using Python 2.7 everything works fine. When adding a findall() method to the Element subclass that simply calls Element.findall() everything also works fine. I suspect some issue with the _elementtree C implementation but I don't understand it.

[jackjansen]

Correction to the last line:
When using Python 2.7 everything works fine. When adding a findall() method to the Element subclass that simply calls ElementPath.findall() everything also works fine. I suspect some issue with the _elementtree C implementation but I don't understand it.

[serhiy-storchaka]

New changeset b11c566 by Serhiy Storchaka in branch 'master':
bpo-34941: Fix searching Element subclasses. (GH-9766)
b11c566

[miss-islington]

New changeset b1c8003 by Miss Islington (bot) in branch '3.7':
bpo-34941: Fix searching Element subclasses. (GH-9766)
b1c8003

[serhiy-storchaka]

New changeset 7c1c42b by Serhiy Storchaka in branch '3.6':
[3.6] bpo-34941: Fix searching Element subclasses. (GH-9766) (GH-9868)
7c1c42b