On Tuesday 10 May 2011 00:13:08 Greg KH wrote:
On Tue, May 10, 2011 at 12:04:10AM +0200, Juergen Weigert wrote:
Other options I could think of:
- implement the checks as rpmlint, so you get to see them earlier. - silently change spec-files during check-in. - accept the original submission, and send back another sr with a fix to
the requester.
- ignore the rubbish that we have in the specfiles, and educate
our customers, that there is no hope for improvement.
How about: - fix up obvious problems in the packages before requiring changes in the license-bot to be stopping others from checking stuff in?
It isn't difficult to fix up 170 different packages and submit the changes if your tool can flag them. It's a pain, yes, but to expect others to do the work for you when you changed the rules is a bit unfair, don't you think? 'Flagging' is (surely) the more gentle way of rejecting, but it's also saying 'nope, it can't enter yet'. According to my experience (doing manual review for Factory), nobody cares for anything except a reject. Instead of flagging stuff, I so far preferred to write personal e-mails. For 1/3 of them I get a response, so after a certain grace period (time for the packager to supersede it's submission), they're all rejected.
Or, if it's too much of a pain, post the problem packages here, and have others help you in fixing them. 170 packages could easily be divided up and fixed up in a day by a small handful of people. That's a word, but usually if your package is rejected because of it's license tag, you change it and resubmit, done in a matter of minutes. -- Mit freundlichen Grüßen, Sascha Peilicke http://saschpe.wordpress.com