bodhi editing - QA hazard

[Michael Cronenworth]

Hi all,

It seems bodhi is abused more than I thought. Over the past few releases 
of being a part-QAer for stable Fedora releases, I've noticed a trend of 
packagers editing updates rather drastically. Take a recent example with 
perl[1]. I left positive karma on perl-5.10.1-115.fc13, but then the 
packager edited the update with a *new* copy of the package: 
perl-5.10.1-116.fc13. The problem? My karma is still giving this *new* 
update a positive point. What if the patches added totally destroyed 
perl's ability to work correctly? I would look like an idiot for giving 
it positive karma -- except only I know that this package's release 
changed. The same thing happened with the wine-1.2[2] update as each 
Release Candidate of wine was released under the same bodhi update.

Why does bodhi allow editing of the package version? It seems that field 
should be locked from editing and the only course of action would be to 
delete the update and create a new one. Another alternative would be to 
have bodhi modified to reset karma points if the package field changes.

I realize everyone has their work cut out for them, but this seems like 
a serious issue. What are your thoughts?

[1] http://admin.fedoraproject.org/updates/perl-5.10.1-116.fc13
[2] https://admin.fedoraproject.org/updates/wine-1.2.0-1.fc13