This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

classification
Title: Inspect library ignore comments at the end of a function (inspect.getsource)
Type: enhancement Stage: resolved
Components: Library (Lib) Versions: Python 3.10, Python 3.9, Python 3.8
process
Status: closed Resolution: fixed
Dependencies: Superseder:
Assigned To: Nosy List: BTaskaya, iritkatriel, miss-islington, noureddine.hamid, taleinat, yselivanov
Priority: normal Keywords: patch

Created on 2020-10-22 10:48 by noureddine.hamid, last changed 2022-04-11 14:59 by admin. This issue is now closed.

Pull Requests
URL Status Linked Edit
PR 23630 merged iritkatriel, 2020-12-03 14:59
PR 23643 merged miss-islington, 2020-12-04 16:45
PR 23644 merged miss-islington, 2020-12-04 16:46
Messages (9)
msg379289 - (view) Author: nhamid (noureddine.hamid) Date: 2020-10-22 10:48
inspect.getsource ignore comments at the end of the function:

for example this function:

def matmul_single(A, x, out):
  from numpy import matmul
  out[:] = matmul(A, x)
  # Some comment here...

using the inspect library:
>>> inspect.getsource(matmul_single)                                                                                                                                     
>>> "def omp_matmul_single(A, x, out):\n  from numpy import matmul\n out[:] = matmul(A, x)\n"

the result does not contain the comments at the end of the function.
msg382416 - (view) Author: Irit Katriel (iritkatriel) * (Python committer) Date: 2020-12-03 15:12
1. For a comment line, the tokenizer emits a COMMENT token followed by an NL token for the newline. The inspect.BlockFinder.tokeneater increments its "last" field to the last line it identified as belonging to the code block. Currently it increments it when it sees a NEWLINE token, but not for an NL token.

2. For a comment line, the tokenizer does not emit an INDENT/DEDENT token, so the indentation level when it is processes is assumed to be equal to that of the previous line.

PR 23630 aims to include comment lines in the block if their start column is after the start column of the opening line of the block:

   def f():
      return 42

     # this is a part of f
   # this is not a part of f
msg382417 - (view) Author: Irit Katriel (iritkatriel) * (Python committer) Date: 2020-12-03 15:16
For reference - this script:
---------------------------------------------
import inspect
import tokenize
from pprint import pprint as pp

src=[
 'def f():\n',
 '    return 1\n',
 '    #that was fun',
 '\n',
 '#Now comes g\n',
 'def g():\n',
 '    return 2\n']

pp(list(tokenize.generate_tokens(iter(src).__next__)))
---------------------------------------------

Outputs:
---------------------------------------------
[TokenInfo(type=1 (NAME), string='def', start=(1, 0), end=(1, 3), line='def f():\n'),
 TokenInfo(type=1 (NAME), string='f', start=(1, 4), end=(1, 5), line='def f():\n'),
 TokenInfo(type=54 (OP), string='(', start=(1, 5), end=(1, 6), line='def f():\n'),
 TokenInfo(type=54 (OP), string=')', start=(1, 6), end=(1, 7), line='def f():\n'),
 TokenInfo(type=54 (OP), string=':', start=(1, 7), end=(1, 8), line='def f():\n'),
 TokenInfo(type=4 (NEWLINE), string='\n', start=(1, 8), end=(1, 9), line='def f():\n'),
 TokenInfo(type=5 (INDENT), string='    ', start=(2, 0), end=(2, 4), line='    return 1\n'),
 TokenInfo(type=1 (NAME), string='return', start=(2, 4), end=(2, 10), line='    return 1\n'),
 TokenInfo(type=2 (NUMBER), string='1', start=(2, 11), end=(2, 12), line='    return 1\n'),
 TokenInfo(type=4 (NEWLINE), string='\n', start=(2, 12), end=(2, 13), line='    return 1\n'),
 TokenInfo(type=60 (COMMENT), string='#that was fun', start=(3, 4), end=(3, 17), line='    #that was fun'),
 TokenInfo(type=61 (NL), string='', start=(3, 17), end=(3, 17), line='    #that was fun'),
 TokenInfo(type=61 (NL), string='\n', start=(4, 0), end=(4, 1), line='\n'),
 TokenInfo(type=60 (COMMENT), string='#Now comes g', start=(5, 0), end=(5, 12), line='#Now comes g\n'),
 TokenInfo(type=61 (NL), string='\n', start=(5, 12), end=(5, 13), line='#Now comes g\n'),
 TokenInfo(type=6 (DEDENT), string='', start=(6, 0), end=(6, 0), line='def g():\n'),
 TokenInfo(type=1 (NAME), string='def', start=(6, 0), end=(6, 3), line='def g():\n'),
 TokenInfo(type=1 (NAME), string='g', start=(6, 4), end=(6, 5), line='def g():\n'),
 TokenInfo(type=54 (OP), string='(', start=(6, 5), end=(6, 6), line='def g():\n'),
 TokenInfo(type=54 (OP), string=')', start=(6, 6), end=(6, 7), line='def g():\n'),
 TokenInfo(type=54 (OP), string=':', start=(6, 7), end=(6, 8), line='def g():\n'),
 TokenInfo(type=4 (NEWLINE), string='\n', start=(6, 8), end=(6, 9), line='def g():\n'),
 TokenInfo(type=5 (INDENT), string='    ', start=(7, 0), end=(7, 4), line='    return 2\n'),
 TokenInfo(type=1 (NAME), string='return', start=(7, 4), end=(7, 10), line='    return 2\n'),
 TokenInfo(type=2 (NUMBER), string='2', start=(7, 11), end=(7, 12), line='    return 2\n'),
 TokenInfo(type=4 (NEWLINE), string='\n', start=(7, 12), end=(7, 13), line='    return 2\n'),
 TokenInfo(type=6 (DEDENT), string='', start=(8, 0), end=(8, 0), line=''),
 TokenInfo(type=0 (ENDMARKER), string='', start=(8, 0), end=(8, 0), line='')]
msg382505 - (view) Author: Tal Einat (taleinat) * (Python committer) Date: 2020-12-04 16:45
New changeset 6e1eec71f59c344fb23c7977061dc2c97b77d51b by Irit Katriel in branch 'master':
bpo-42116: Fix inspect.getsource handling of trailing comments (GH-23630)
https://github.com/python/cpython/commit/6e1eec71f59c344fb23c7977061dc2c97b77d51b
msg382526 - (view) Author: miss-islington (miss-islington) Date: 2020-12-04 20:20
New changeset 81ac030d03bdaedd724603af6f89f9248a5f2700 by Miss Islington (bot) in branch '3.9':
bpo-42116: Fix inspect.getsource handling of trailing comments (GH-23630)
https://github.com/python/cpython/commit/81ac030d03bdaedd724603af6f89f9248a5f2700
msg382527 - (view) Author: miss-islington (miss-islington) Date: 2020-12-04 20:20
New changeset 3b14f18205b17d1634e21bd7bc48152247590d9f by Miss Islington (bot) in branch '3.8':
bpo-42116: Fix inspect.getsource handling of trailing comments (GH-23630)
https://github.com/python/cpython/commit/3b14f18205b17d1634e21bd7bc48152247590d9f
msg382528 - (view) Author: Tal Einat (taleinat) * (Python committer) Date: 2020-12-04 20:21
Thank you for reporting this, Noureddine Hamid!

Thanks for the PR, Irit!
msg383047 - (view) Author: nhamid (noureddine.hamid) Date: 2020-12-15 12:51
thank you for the fix, I forgot to mention that python 3.6 and python 3.7 have this issue too.
msg383054 - (view) Author: Irit Katriel (iritkatriel) * (Python committer) Date: 2020-12-15 13:21
Thanks for the report. 3.6 is no longer maintained and 3.7 is getting security fixes only. So this won't be backported to those versions.
History
Date User Action Args
2022-04-11 14:59:37adminsetgithub: 86282
2020-12-15 13:21:15iritkatrielsetmessages: + msg383054
versions: - Python 3.6, Python 3.7
2020-12-15 12:51:27noureddine.hamidsetmessages: + msg383047
versions: + Python 3.6, Python 3.7
2020-12-04 20:21:43taleinatsetstatus: open -> closed
resolution: fixed
messages: + msg382528

stage: patch review -> resolved
2020-12-04 20:20:13miss-islingtonsetmessages: + msg382527
2020-12-04 20:20:12miss-islingtonsetmessages: + msg382526
2020-12-04 16:46:04miss-islingtonsetpull_requests: + pull_request22512
2020-12-04 16:45:54miss-islingtonsetnosy: + miss-islington
pull_requests: + pull_request22511
2020-12-04 16:45:46taleinatsetmessages: + msg382505
2020-12-03 18:59:33BTaskayasetnosy: + BTaskaya
2020-12-03 15:16:05iritkatrielsetmessages: + msg382417
2020-12-03 15:12:06iritkatrielsetnosy: + taleinat

messages: + msg382416
versions: + Python 3.9, Python 3.10
2020-12-03 14:59:31iritkatrielsetkeywords: + patch
nosy: + iritkatriel

pull_requests: + pull_request22498
stage: patch review
2020-10-24 00:26:43terry.reedysetnosy: + yselivanov
2020-10-22 10:50:14noureddine.hamidsettype: enhancement
2020-10-22 10:48:43noureddine.hamidcreate