python-daemon

Clone
Source Code
GIT

#20 Proposed fix for issue #18 -- require "docutils".
Closed 5 years ago by bignose. Opened 5 years ago by ftobia.
ftobia/python-daemon master  into  master

file modified
+4 -8 
@@ -56,9 +56,6 @@ 
 

  

          # Setuptools metadata.
 

          zip_safe=False,
 

-         setup_requires=[
 

-             "docutils",
 

-             ],
 

          test_suite="setup.test_suite",
 

          tests_require=[
 

              "unittest2 >=0.5.1",
 
@@ -68,6 +65,10 @@ 
 

              "docutils",
 

              ],
 

          install_requires=[
 

+             "docutils",  # Technically only required by the build step,
 

+                          # but common install tools like Setuptools
 

+                          # and pip don't respect "setup_requires" yet.
 

+                          # See <URL:https://github.com/pypa/setuptools/issues/457>.
 

              "setuptools",
 

              "lockfile >=0.10",
 

              ],
 
@@ -92,11 +93,6 @@ 
 

              ],
 

          )
 

  

- # Docutils is only required for building, but Setuptools can't distinguish
 

- # dependencies properly.
 

- # See <URL:https://github.com/pypa/setuptools/issues/457>.
 

- setup_kwargs['install_requires'].append("docutils")

How does removing this help? The install_requires parameter had a "docutils" item, so why remove it from here but add it in another place?

 
- 

  

  if __name__ == '__main__':
 

      setup(**setup_kwargs)

no initial comment

How does removing this help? The install_requires parameter had a "docutils" item, so why remove it from here but add it in another place?

Thanks for suggesting this.

Does this change the behaviour? The install_requires parameter already had a "docutils" item, this change appears to have no net effect on the install_requires parameter.

That's a good point, this change shouldn't change behavior. Let me dig into this a bit more.

Given that "docutils" appears to have been added to "install_requires", I'm confused as to why issue #18 was even happening in the first place.

Please there is a chance that will be merge or fix this problem? The OpenStack project have an open critical ticket about this problem.

See #22 to an another approch to fix the problem, fully refactor of the packaging system.

https://pagure.io/python-daemon/pull-request/22

@hobbestigrou:

Please there is a chance that will be merge or fix this problem? The OpenStack project have an open critical ticket about this problem.

@ftobia, have you found anything more since you planned to try reproducing and investigating this?

I'm sorry, I haven't dug into this any deeper or come up with any other solutions. At this point I'm not planning on investigating further. Sorry again.
Edited 5 years ago by hberaud

On 16-Jan-2019, Natal Ng=C3=A9tal wrote:

Please there is a chance that will be merge or fix this problem?

This merge request has been closed with discussion from the proposer,
who has acknowledged they do not have the time to work on it.

The OpenStack project have an open critical ticket about this
problem.

That appears to be something that needs a different merge request or
bug report, then.

--=20
\ =E2=80=9CThe generation of random numbers is too important to b=
e left |
`\ to chance.=E2=80=9D =E2=80=94Rob=
ert R. Coveyou |
_o__) |=

Ben Finney ben@benfinney.id.au

Pull-Request has been closed by bignose
5 years ago
Metadata
None
No Tags
Changes Summary 1
+4 -8
file changed
setup.py
Powered by Pagure 5.13.3
Documentation • File an Issue • About • SSH Hostkey/Fingerprint
© Red Hat, Inc. and others.