[issue37605] CI should not depend on gmane response

[Terry J. Reedy]

Terry J. Reedy  added the comment:

nntp.rst, line 23, has already been updated to "s = NNTP('news.gmane.io').  But 
that also occasionally fails to respond.  I have not seen any test failures 
recently, so maybe something else has changes.  Should this be closed pending 
current failures?

--
versions: +Python 3.11 -Python 2.7, Python 3.7, Python 3.8, Python 3.9

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue37605] CI should not depend on gmane response

[robin]

robin  added the comment:

In other news, news.gmane.org is now news.gmane.io; see also 
https://bugs.python.org/issue39343

--
nosy: +eriker

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue37605] CI should not depend on gmane response

[Zachary Ware]

Zachary Ware  added the comment:

For doctests, I would prefer to just disable some or all of the doctests in 
Doc/library/nntplib.rst rather than go to the effort of trying to mock things.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue37605] CI should not depend on gmane response

[Ngalim Siregar]

Ngalim Siregar  added the comment:

is it allowed to mock news.gmane.org response?

if it is okay i would like to help

--
nosy: +nsiregar

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue37605] CI should not depend on gmane response

[Terry J. Reedy]

Terry J. Reedy  added the comment:

gmane was down awhile yesterday, so another merge (temporarily) blocked because 
of this.  If either of you knows how to fix this, please do so.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue37605] CI should not depend on gmane response

[Zachary Ware]

Zachary Ware  added the comment:

Victor, note that this issue is about the nntplib doctests in 
Doc/library/nntplib.rst.  I agree that any examples in that file that actually 
hit an external server should be marked as not doctests.

--
assignee:  -> docs@python
components: +Documentation
nosy: +docs@python, zach.ware
versions: +Python 2.7, Python 3.7, Python 3.8

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue37605] CI should not depend on gmane response

[STINNER Victor]

STINNER Victor  added the comment:

> On Travis, 'Documentation tests' passing is apparently required.  Line 23 of 
> nntplib.rst is ">>> s = NNTP('news.gmane.org')"  I use gmane and occasional 
> failures to respond are routine.  Being able to merge should not depend on 
> this external site.

Python test suite uses many external services:
https://pythondev.readthedocs.io/infra.html#services-used-by-unit-tests

You should all be skipped if the "network" resource is not used, which is the 
default.

Many tests were modified to catch network failures and ignore the test in this 
case. test_nntplib is known to be very fragile: it doesn't handle network 
issues.

I would suggest to remove functional tests which depend on external servers, 
since these tests fail often and nobody seems to care of test_nntplib.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue37605] CI should not depend on gmane response

[Terry J. Reedy]

New submission from Terry J. Reedy :

On Travis, 'Documentation tests' passing is apparently required.  Line 23 of 
nntplib.rst is ">>> s = NNTP('news.gmane.org')"  I use gmane and occasional 
failures to respond are routine.  Being able to merge should not depend on this 
external site.

In addition to a real failure, the following happen on PR 14675 earlier today.

https://travis-ci.org/python/cpython/jobs/559289634

Warning, treated as error:
**
File "library/nntplib.rst", line ?, in default
Failed example:
s = NNTP('news.gmane.org')
Exception raised:
Traceback (most recent call last):
  File "/home/travis/build/python/cpython/Lib/doctest.py", line 1328, in 
__run
exec(compile(example.source, filename, "single",
  File "", line 1, in 
s = NNTP('news.gmane.org')
  File "/home/travis/build/python/cpython/Lib/nntplib.py", line 1050, in 
__init__
_NNTPBase.__init__(self, file, host,
  File "/home/travis/build/python/cpython/Lib/nntplib.py", line 331, in 
__init__
self.welcome = self._getresp()
  File "/home/travis/build/python/cpython/Lib/nntplib.py", line 456, in 
_getresp
raise NNTPTemporaryError(resp)
nntplib.NNTPTemporaryError: 400 load at 16.59, try later

Try later is what I do.  This exception should somehow be skipped.  I don't 
know anything about Travis and doctest tests and how flakey doctests are 
supposed to be handled. If nothing else, put the whole example is a try-except 
block?

>>> try:  # News sites can fail to respond.
... s = NNTP('news.gmane.org')
... ...
... except: # If site failure, try later.
... pass

Since output varies, I presume output checking is suppressed.

--
components: Tests
messages: 348023
nosy: pablogsal, terry.reedy, vstinner
priority: normal
severity: normal
stage: needs patch
status: open
title: CI should not depend on gmane response
type: behavior
versions: Python 3.9

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com