Sebb created COMDEV-164:
---------------------------

             Summary: parsepmcs.py does not handle new groups correctly
                 Key: COMDEV-164
                 URL: https://issues.apache.org/jira/browse/COMDEV-164
             Project: Community Development
          Issue Type: Bug
          Components: Reporter Tool
            Reporter: Sebb


The parsepmcs.py script currently parses 
http://people.apache.org/committer-index.html and generates lists of the auth 
groups that committers belong to, as well as the date when the person was first 
added to the new group. When a new group is seen, all the joining dates are 0, 
so the JSON files can be used to show subsequent changes to the auth groups 
without needing to know when the group itself was first seen.

However the process is broken, because it keeps track of new groups using the 
same array for both committee (e.g. abdera-pmc) and committer (e.g. abdera) 
groups. This would be fine except the "-pmc" suffix is stripped from the group 
name before it is stored in the array. Thus once the committer group is seen, 
the committee group won't be regarded as new if it is added in a future run. 
And vice versa. The process only works if the two groups are added in the same 
run, and even then there is an issue because the check for the new committer 
group is only done if it has not just been added. (the check for new committee 
group is OK).

It looks like the only effect of this is to record some ids as having a joining 
date when in fact they should not because they were part of the initial group.

This can be fixed by manually editting the json files, though determining which 
entries need to be fixed may not be trivial.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to