Author marco.buttu
Recipients Jelle Zijlstra, berker.peksag, brett.cannon, ezio.melotti, marco.buttu, python-dev, r.david.murray, terry.reedy, zach.ware
Date 2017-01-04.11:06:08
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <>
Thinking a little bit about it, I believe that the easiest way to proceed is to make one patch for every file or two, with the only purpose to make their doctests to pass. Otherwise if I make a patch with all the doctest directives, or just with the +SKIP options, we will have hundreds of failures, and it will be really hard to understand what is going wrong, also because the tests are not isolated, and we will have some random failures. Reducing the number of failures by making a patch with all the doctest/testsetup/testcleanup directives is not a solution, because we eventually fall again into a big patch, and the doctests are still not pass too. 

That is why I think having one patch for two files will reduce a lot the effort in making and reviewing the patch. We will end up in about 9 patches, and it will be really easy to review most of them (I think couple of minutes). It will be also easy to test them. For instance, if the patch involves the files Doc/library/copyreg.rst and Doc/library/ctypes.rst, we just have to to to the Doc/ directory and run: 

  $ sphinx-build -b doctest . build/doctest \
  > library/copyreg.rst \
  > library/ctypes.rst

expecting 0 failures. 

Maybe having one new issue for each patch will also simplify the process, because the purpose of the patch will be well explained by the issue description and title. Let me know what you think, and thank you very much for your time.
Date User Action Args
2017-01-04 11:06:08marco.buttusetrecipients: + marco.buttu, brett.cannon, terry.reedy, ezio.melotti, r.david.murray, python-dev, berker.peksag, zach.ware, Jelle Zijlstra
2017-01-04 11:06:08marco.buttusetmessageid: <>
2017-01-04 11:06:08marco.buttulinkissue27200 messages
2017-01-04 11:06:08marco.buttucreate