Problem with SqueakMap

goran.krampe at bluefish.se goran.krampe at bluefish.se
Tue Nov 25 14:42:11 UTC 2003


"Frank Shearar" <frank.shearar at rnid.org.uk> wrote:
> > goran.krampe at bluefish.se writes:
> >  > Should be no problem as far as SM is concerned. The new 
> > Kom is different
> >  > from old Comanche though, new Kom is not backwards 
> > compatible - even
> >  > though it is easy to port stuff over to it. So what did you use it
> >  > for?
> > 
> > I think that I originally loaded it in this image so I could study how
> > it starts up and shutdowns on image startup/shutdown.  The only
> > service that I have running allows me to remotely force a Monticello
> > commit of my development packages.  I often forget to publish my MC
> > packages before I leave for the day.  I haven't tested it yet but it's
> > such a simple service that I should be able to get it working even in
> > the event of large differences.
> > 
> > Thanks to everybody for the help on this...
> 
> I've been getting the same error, but I haven't installed Comanche or Kom
> in my image.

Eh, are we talking about the same image as before? That *had* Comanche
installed?

> After some hacking and swearing and lots of plain blind luck, I tried
> 
>     SMSqueakMap default loadFull
> 
> which did the trick for me. In retrospect this seems quite strange. Why
> should I have been getting the "cards at: k..." error when hitting the
> "upgrade map from net" menu item, but not from the above snippet?

Oj, you got it when trying to update the map? Ok.

Well, those two mechanisms are different. The "upgrade map from net"
tries to only load the transactions since you updated last. #loadFull
simply throws away the whole map and reloads it in full from the server.

If the former didn't work - it might be a bug in SM1.x regarding deleted
packages. I need more info to resolve that.
But I am not very keen to do it, since we hopefully are moving to SM2
which doesn't do incremental updates. :)

> frank

regards, Göran



More information about the Squeak-dev mailing list