Re: [Gnu-arch-users] [ANNOUCEMENT] gnuarch-1.2.2rc2 release

[James Blackwell]

jblack wrote:
>> I have just cut and released 1.2.2rc2. This release candidate fixes a
>> 1.2.1 regression that caused get-changeset to refuse get-changeset.
>> Thanks goes to Aaron Bentley who stayed up extra late to fix it.

-t wrote:
> Could you please be a little bit more proactive about alerting me to 
> the impact of such things on making GNU releases?

I intended to. Real life stepped in just after I had mailed the list,
but before I could mail you. You'll get a private email from me on the
same day an rc comes out. Promise. ;)

> For example, I was assuming we were counting down two weeks from
> 1.2.2rc1 before I was supposed to either make a release (expected
> case) or raise a red flag (unlikely but that's part of what I'm there
> for).

We were, but now that rc2 came out today, the clock restarts, and now
I'm hoping for a Sept 24 release. Though it may seem silly to throw away
three days of testing time just for one lousy patch, its a consistant
process so that everybody's on the same page, that makes sure that
anything that goes out as an official release had a fair shot at
testing.

> Now I'm a little confused where I should be in my part of the process,
> in your view.  I gather the clock is slightly reset by rc2 but I'm not
> sure precisely what the co-maintainers' expectations are regarding the
> status of their work in relation to GNU releases.....

You can review the rc at any time you like, but I suggest you give
everyone ~ a week to catch the obvious stuff. After a week (7 days +- 1
day), I'll send you an email with a subject along the lines of "RC 1
WEEK OLD". At that point, I recommend you figure out when during the
next week that you can audit the code.

As with always, if anybody finds an rc regression, we'll either A) pull
the regressing patch or B) fix the regression. Then, I'll cut a new rc
and the clock starts over. 

If we find a bug during the rc process, that goes straight into the
development line, but doesn't go into the rc process. Instead, it'll
wait for the next rc, which will be coming along in about 2-3 weeks. As
with anything, there *can* be exceptions. If a really horrid bug comes
along, then I can be convinced to pull it into the rc, and treat it as
if it were an rc regression.

Btw, you're eligible for the prize too. If you find an rc regression and 
spend 20 minutes hacking up a test case and fix, then you get to have
a night out at the movies on someone else's tab. 

The day before release (sometimes two if real life causes me to have to
postpone release by a day) I'll send you an email with a subject along
the lines of "IMMINENT FULL RELEASE for x.y.z"


-- 
James Blackwell          Try something fun: For the next 24 hours, give
Smile more!              each person you meet a compliment!

GnuPG (ID 06357400) AAE4 8C76 58DA 5902 761D  247A 8A55 DA73 0635 7400