Hi Alejandro and Daniel (and all you other Squeakers of course)!

"Magistrello Alejandro (SFA)" <magistra at TELEFONICA.COM.AR> wrote:
> Hi=20
> >
> >[SNIP]
> >> There was allegedly a situtation where things misbehaved when SM
> >> couldn't find any servers. Goran, you blamed me. Bad, bad goran ;-)

Sorry, as you see below - it was my fault. :-)

> >:-) Well, I am not sure I blaed you - not finding the servers was =
> probably
> a
> >proxy/firewall problem in itself (I hope, haven't heard anymore from
> Alejandro)
> 
> it was not,=20
>
> i answered to a mail where you say this:
> >And you could also try with commenting out the first 2 lines in
> >#pingServer: if you are having trouble with the DNS or something:
> >
> >	"serverAddr _ NetNameResolver addressForName: (aServerName upTo: $:)
> >timeout: 5.
> >	serverAddr =3D nil ifTrue: [^false]."
> 
> and this made it work.

Hmmm. This is a little check that I added because the timeout in httpGet
was too long or simply seemed to hang when I tried to reach a
nonexistant machine (map.squeak.org for example). I am not sure how we
can make this more robust - any takers?

> don=B4t know how to reproduce the dnu,=20
> so attached is the debuglog from a fresh image, whitout SM and filein =
> RC1.
> 
> Best regards
> Alejandro Magistrello

Ah!

Looking at this log it seems that it was my fault! :-) It is my little
postscript in RC1/RC2 that tries to install SM Package Loader that
doesn't handle an empty SqueakMap. Ok, I will fix that. Sorry Daniel for
implying you had anything to do with it! :-)
And btw - the other "bug" I talked about - having SMLoader open up
"empty" was probably the "Smalltalk version" bug I reported in another
post - just so you know.

regards, Göran