doctest test finder doesnt find line numbers of properties

[RonnyPfannschmidt]

BPO	17446
Nosy	@blueyed, @bitdancer, @RonnyPfannschmidt, @Vgr255, @mscuthbert, @timofurrer
PRs	bpo-17446: Get line numbers of properties in doctest #3419
Files	fix_issue_17446.patch

^{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 = None
closed_at = None
created_at = <Date 2013-03-17.16:08:27.046>
labels = ['type-bug', '3.7']
title = 'doctest test finder doesnt find line numbers of properties'
updated_at = <Date 2019-10-28.17:25:43.144>
user = 'https://github.com/RonnyPfannschmidt'

bugs.python.org fields:

activity = <Date 2019-10-28.17:25:43.144>
actor = 'blueyed'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = []
creation = <Date 2013-03-17.16:08:27.046>
creator = 'Ronny.Pfannschmidt'
dependencies = []
files = ['41769']
hgrepos = []
issue_num = 17446
keywords = ['patch']
message_count = 12.0
messages = ['184384', '228122', '259300', '259306', '259618', '259620', '259621', '261372', '262146', '293747', '303699', '355572']
nosy_count = 6.0
nosy_names = ['blueyed', 'r.david.murray', 'Ronny.Pfannschmidt', 'abarry', 'mscuthbert', 'tuxtimo']
pr_nums = ['3419']
priority = 'normal'
resolution = None
stage = 'patch review'
status = 'open'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue17446'
versions = ['Python 3.7']

Linked PRs

• gh-61648: Detect line numbers of properties in doctests #113161
• [3.12] gh-61648: Detect line numbers of properties in doctests (GH-113161) #113164
• [3.11] gh-61648: Detect line numbers of properties in doctests (GH-113161) #113165

[RonnyPfannschmidt]

examples that are found on a property dont detect the line number

class example(object):
  @property
  def me(self):
   """
   >>> 1/0
   """
   pass

[BreamoreBoy]

@ronny can you provide a patch for this?

[mscuthbert]

this is my first contribution to Python core so I really have no idea how to do this, but I have found a solution (works in Py3.4, 2.7):

in doctest.py after line 1087 ("lineno = getattr(obj, 'co_firstlineno', None)-1") add these lines:

    if lineno is None and isinstance(obj, property) and \
            obj.fget is not None:
        obj = obj.fget.__code__
        lineno = getattr(obj, 'co_firstlineno', None) 
        # no need for -1 because of decorator. 

I can try to make a patch file for this, but just want to be sure I'm on the right track for contributing first. I know how to do a Git pull, but not hg/patch.

(p.s., I think the current code "lineno = getattr(obj, 'co_firstlineno', None)-1" has an error; if the getattr does not find 'co_firstlineno', it will return None and then subtract 1 from None which is a TypeError).

[timofurrer]

I took the ideas from @Michael.Cuthbert and wrote a proper test. It's my first patch so I hope everything's fine with it. If not I'm happy for feedback :)

[mscuthbert]

The test looks great to me. Does anyone on nosy know the proper way to request a patch review?

[Vgr255]

Left a comment on Rietveld. I don't have time right now to check the test, but I suspect you tested it before submitting the patch, so it should probably be fine.

[timofurrer]

Yes, I've tested it.

[mscuthbert]

looks like we're stuck on a style change (backslash to parens; ironic: I chose backslash to match surrounding code; I never use them myself). tuxtimo - could you fix this? (or I'll do once the move to github is done). Thanks!

[mscuthbert]

Here's a rather obscure bug that I was able to catch before we put this into action: doctests inside the __doc__ for namedtuples (and perhaps all namedtuples?) are instances of property, have .fget, but do not have .fget.__code__. Thus one more check is needed:

    if (lineno is None and 
            isinstance(obj, property) and
            obj.fget is not None and 
            hasattr(obj.fget, '__code__')):
        obj = obj.fget.__code__
        lineno = getattr(obj, 'co_firstlineno', None)

[mscuthbert]

just poking to see if this patch is worth trying to get into 3.7

[mscuthbert]

A pull request has been in for about a month -- is it possible to review or merge or comment? Thanks!

[blueyed]

The PR appears to need a better test according to #3419 (comment).

[mscuthbert]

The code I contributed to 3.7 is still, I believe, the best path forward. Based on my experiences trying to contribute to cpython, I will not be continuing on this issue. It can be closed.

[AA-Turner]

The linked PR (#3419) seems to have been closed with no discussion, I'm reopening.

A