Author terry.reedy
Recipients cheryl.sabella, terry.reedy
Date 2017-05-28.20:00:22
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1496001623.02.0.382710077392.issue30495@psf.upfronthosting.co.za>
In-reply-to
Content
What I have learned:
1. As with hg, the local master should be updated at least at the beginning of each work session, and python.exe rebuilt.
2. Patches to master should be developed with all 'other' files updated.  For completely new branches, this is automatic.  Branches created from PRs start out of sync and should be updated.

One can often, perhaps even 'usually', get away with not doing the above, but occasional mysterious glitches can occur.  Example: when I created a branch from your bdb docstring PR, 9 days later, and did not update, 'import re' failed in the new branch.  Problem disappeared after update.

By 'grep', I mean any multi-file search, starting in a specified directory, such as IDLE's 'Find in Files', not the specific unix program.  I develop code and test for x by repeatedly running just 'test_x', but grepping and running test_idle before commit is essential.  As we expand test files, test_idle will become more and more useful.
History
Date User Action Args
2017-05-28 20:00:23terry.reedysetrecipients: + terry.reedy, cheryl.sabella
2017-05-28 20:00:23terry.reedysetmessageid: <1496001623.02.0.382710077392.issue30495@psf.upfronthosting.co.za>
2017-05-28 20:00:23terry.reedylinkissue30495 messages
2017-05-28 20:00:22terry.reedycreate