On Wednesday 29 December 2004 2:18 pm, I wrote:
This is with the 3.8.5beta1 VM from John McIntosh's iDisk upon loading from SM. The sm directory was empty.
And with the 3.7.5Beta3 VM it worked fine.
Ned Konz ned@squeakland.org wrote:
On Wednesday 29 December 2004 2:18 pm, I wrote:
This is with the 3.8.5beta1 VM from John McIntosh's iDisk upon loading from SM. The sm directory was empty.
And with the 3.7.5Beta3 VM it worked fine.
The latest 'fix' to the problem of some dangerous becomes is happy (so far as my tests can tell) with the previously fatal case of 'a becomeForward: a' which is what your walkback looks like. Try it again and in the debugger see if the object being becomed is the same on both sides. If it is, then the _very_ latest VM ought to be ok with it. A not-quite so latest one will fail in what looks suspiciously like this manner.
I wrote a long and rather boring explanation of the problem(s) in this are in the 'becomeForward makes corrupted objects' mantis bug report.
tim -- Tim Rowledge, tim@sumeru.stanford.edu, http://sumeru.stanford.edu/tim Useful random insult:- One bit short of a word.
squeak-dev@lists.squeakfoundation.org