On 03/25/2011 03:33 PM, Paolo Bonzini wrote: > On 03/25/2011 03:28 PM, Holger Hans Peter Freyther wrote: >>> > >>> > I'll apply the Buffers.st part since it's good anyway. >> hmm. does this mean I can't read data out of a remotely closed socket? Or >> only >> when I have called close myself. Sorry for the stupid question. > > No, only when you have called close yourself. In fact, I'm not sure what > would call #close on the buffers; nothing, probably. But the patch would > definitely tell us if that's happening. > > Now that I think of it, while my patch to the buffer classes is good for GST, > you may want to replace the two implementations of #close with just "self > halt".
Okay, I can reproduce the issue using tcpkill of dsniff on the connection. I will debug tomorrow morning. While testing I think I stumbled across mem corruption of DoIt's (Eval [] in a file) and doing CTRL+C (e.g. when waiting for a socket...). It is causing a segfault when clearing the obstack. :) _______________________________________________ help-smalltalk mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-smalltalk
