Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: (ASCEND) Are Patches Rolled Into Subsequent Patch Releases?

On Mon, 24 Nov 1997, Gerald Honma wrote:

> It's not clear to me. Starting with a general release, say 5.0A for 
> Pipeline. There are currently seven (7) patch releases identified at 
> http://www.ascend.com/769.html. Specifically, 5.0Ap1, 5.0Ap3, 5.0Ap4, 
> 5.0Ap5, 5.0Ap7, 5.0Ap8, and 5.0Ap10.
> 
> Does 5.0Ap10 INCLUDE all the fixes mentioned in the previous 5.0Ap# 
> patches? The documentation for each patch does not seem to mention/repeat 
> the "fixes" identified in previous patches.

Yes.  That doesn't mean they don't reintroduce an old problem though.

> If the answer to the above is "yes", is it correct to say that 5.0Ap10 is 
> a STABLE release of the features identified in the 5.0A Release Notes?

No.  

1. p releases are often not stable; they don't appear any more stable than
i releases to me. 

2. "new features" and "bug fixes" can have very different meanings
depending on what Ascend wants.  I would suggest that the k56 modem code
is a hell of a new feature.  I don't call it a bug fix.  Yet it had to go
into p, otherwise they would only have it in i and wouldn't be able to
tell people to go to the latest buggy p release instead of the latest
buggy i release.  In general, they do try to keep the p releases somewhat
consistent with the feature set in the base revision.  

> Also, is it true that if additional features beyond those identified in 
> the 5.0A Release Notes are not required for my situation that I do not 
> need to upgrade to the 5.1 line? That is, are all "bugs" associated with 
> 5.0A features going to be "fixed" in a 5.0A patch? Or, are some 5.0A 
> "bugs" ONLY fixed by upgrading to 5.1? If this is the case, this seems 

Sometimes you have to upgrade to the next revision and not just a new p
release.  You will sometimes have new p releases for an old revision even
after the new one is out; they will fix some things, but seldom all bugs
that have been fixed in the new revision.

Of course, you have to consider that around 80% of the Ascend bugs we
experience seem to be one long set of bugs that are never fixed (it is
possible someone just keeps reintroducing them, but they look the same
from outside).  

> like a never ending cycle where one is forced to upgrade.

You catch on quick.  It is even better when there is no release for you to
upgrade to, you can't downgrade past a certain version because your modem
cards don't support it, and the current version barely scrapes by. 

> BTW, the above mentioned Web page states, "Whenever possible, we recommend 
> using a Patch Release because it is FULLY (my emphasis) 
> regression-tested." I've seen some comments recently indicating that Patch 
> releases are not "FULLY" regression-tested. ???

There have been some x.x revisions (ie. no p or i) that have been
absolutely unusable.  Don't put too much stock in Ascend's regression
testing.

++ Ascend Users Mailing List ++
To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd:	<http://www.nealis.net/ascend/faq>
References: