Since we're nearing 3.6 release, I think any of the solutions is
reasonable - the problem will start shrinking soon.

Daniel

Doug Way <dway at riskmetrics.com> wrote:
> Jimmie Houchin wrote:
> 
> > I have avoided updating hoping things would be fixed before I 
> > needed/wanted to update.
> >
> > I decided to try out the Jabber client.
> > I needed to update first.
> >
> > From a 5248 image I can't get past 5256 due to a MNU: 
> > initializeNetworkIfFail: in the NetNameResolver class.
> 
> 
> This update still brings up an exception for some people, depending on 
> how fast your updates are coming in.  However, the problem is actually 
> harmless, and you can just close the notifier and ask to load the rest 
> of the updates, and you'll get them all.
> 
> For more details, see:
> http://lists.squeakfoundation.org/pipermail/squeak-dev/2003-June/060209.html
> http://lists.squeakfoundation.org/pipermail/squeak-dev/2003-June/060282.html
> 
> I suppose people are going to keep running into this, though, and 
> legitimately asking questions on the list about it.  There are a few 
> things I could do about it:
> 
> 1. Do nothing.  The number of people updating through this changeset 
> will gradually decrease.  Or,
> 2. Replace the 9 network rewrite changesets with 1 combined changeset 
> and 8 dummy changesets as Joshua suggested.  This would fix the 
> problem.  (Assuming there are no order dependencies in the changesets, 
> which I think is the case.)  However, this has the slightly unsettling 
> effect that some 3.6alpha users will have different changeset contents 
> than others.  Although the resulting source code in the image should be 
> the same, so it's not really a big deal.  Or,
> 3. Put a warning at the beginning of the first network-rewrite update, 
> warning that on some machines, an exception may occur, which can be 
> safely closed and one can continue loading more updates.  I sort of like 
> this option, it's simple, but would hopefully stop the confusion.
> 
> - Doug Way