ext Alexander Kanevskiy <[email protected]> writes:

> On Thu, Sep 23, 2010 at 18:16, Roger WANG <[email protected]> wrote:
>> "Nashif, Anas" <[email protected]> writes:
>>
>>> Hi,
>>> We have noticed recently that bug fixes are being submitted but
>>> without a bug number in the change log (.changes file) but with
>>> detailed bug information in the commit message, which is used for
>>> communication, but is never part of the resulting package.
>>
>>> Please add the bmc numbers or features you are working on in the
>>> .changes file also, this is critical and otherwise changes will be
>>> declined.
>>
>> What should I do if I find a bug is fixed by the new version later,
>> after the package is accepted?
>
> if bug is not anymore reproducible with newly accepted version and
> this version doesn't link to that bug, then close bug with
> "WORKSFORME" resolution.

There is, however, a difference between a bug that can not be reproduced
and we might thus have a bug that we haven't fixed, and a bug that is
known and fixed already, and we can forget about it.  We might want to
distinguish these in Bugzilla.

I agree that recording all fixed bugs in .changes is good, but if we
enforce it strictly, we need a way to announce fixes in past versions as
well, and accept that people might not want to make a new release just
to add something to .changes.


I see two ways: We can allow people to edit already released .changes
entries (and make sure the tools will not be fooled by that), or we can
simply make a convention for listing historical fixes in new .changes
entries, maybe like so:

  * Frob the doodles.  Fixes: BMC#xxxx.

  * Already fixed in version 0.1:
  - Fixes: BMC#Bxxxx, BMC#xxxx.

This should be good enough, I'd say.
_______________________________________________
MeeGo-packaging mailing list
[email protected]
http://lists.meego.com/listinfo/meego-packaging

Reply via email to