[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: suite tags (Re: Doing some stable QA work)

Don Armstrong wrote:
On Fri, 15 Aug 2008, Frank Lichtenheld wrote:
On Thu, Aug 14, 2008 at 08:08:10PM -0700, Steve Langasek wrote:
On Thu, Aug 14, 2008 at 11:01:46PM -0400, Filipus Klutiero wrote:
Tagging lenny and sid does not imply that other suites are free
from the bug,
Yes, it does.

It only implies that for britney; the BTS itself ignores it save for

Thanks for the clarification.

It's quite likely that we should change this, but it's not the way it
works currently. [1]

How to interpret those tags has always been problematic in my mind,
but I suppose the following interpretation would resolve what the
release managers are looking for, and would also work for the bts.

*** proposal ***
For the purposes of determining buginess, setting a distribution tag
limits the bugginess of a package in distribtions to the intersection
of the distributions and the version-based bugginness of a package.

This means that a bug that would normally be buggy in etch and lenny
(but fixed in sid) due to the versions present of that package would
only be present in lenny if tagged lenny and tagged sid.

Does this also mean that the bug would not be present anywere when the intersection would be empty (tagged lenny with a version higher than the one in testing for instance)? We should probably look at all the possibilities and think about good defaults for them?

For the purposes of archival, we just use the buginess state above
instead, defaulting to (testing distribution), sid.

*** end proposal ***

Does that meet the needs of the RMs et al? Are there any objections?

I think it's a bit more complex than what you propose, but it seems to be a good starting point.



Reply to: