On Mon, 2008-07-28 at 01:46 +0200, Dag Wieers wrote:
> Not that I disagree with you, but there are only 2 windows per year to 
> patch this if it is not a security-issue (or no security issue is found in 
> that package).
> 
> So the best that could be done is roughly 183 (6 months) if a problem is 
> found only just after a release and it is not a security issue :-/

Perhaps Redhat should reevaluate this policy, at least during the few
months following the initial release.  The problem with these very long
cycles to get bugs fixed, especially early in the release cycle, is that
many times the bugs in the initial releases are significant enough to be
show stoppers for deployment in some organizations.  Until the product
is widely deployed the smaller, more subtle bugs won't be found.  

For example, with RHEL4 we were unable to perform significant
deployments in our environment until U3.  When we finally started those
deployments we then found more bugs, perhaps less critical, but
certainly issues that needed to be addressed.  Most of these weren't
fixed until U5.  That means it took basically two years to get RHEL4
into a full deployment state in our organizations.  Certainly there's a
way to get these bug fixes into the updates following the initial
release a little faster.  Perhaps during the first year after a release
a 3-4 month update release cycle to "accelerate" these initial bug
fixes.

Maybe this just doesn't work because of the time to QA everything, but I
consider the entire QA excuse pretty weak based on their track record of
releasing horribly broken packages.

> I hate it too, but the only way to improve this with the given rules, is 
> to test Beta releases for existing bugs and escalate (somehow). I think 
> Beta's should be promoted/tested more to reduce regressions.

One problem with the betas is that Redhat seems to consider them
"unsupported previews" rather than a real beta.  They almost never
update the beta channel after a release with fixes, and just try to open
a support case against a beta.  It's generally met with a "it's a
unsupported beta" response.  I have filed bugs against the betas in
bugzilla that did get fixed, but others are completely ignored and end
up getting released as a regression and then I have to open a support
case, and hope it gets fixed in another update or two.  Very
frustrating.

Later,
Tom


_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list

Reply via email to