Re: CLOS hash tables & threads - crash/bug

[Sam S. <sd...@gn...>]

Don Cohen wrote:
>  > >> i.e., we will avoid crashes, but the semantics (and, thus,
>  > usefulness) would be lost anyway.
> It's a lot more useful to find yourself in an unexpected state in
> lisp than in the shell looking at a segfault message.

for a long-running process with a valuable state - yes.
otherwise - not sure.
what is preferable in C - segfault or memory leak?
in production code - memory leak (less dangerous consequences).
while debugging - segfault (pinpoints the error).

BTW, Don, did you try my tiny patch?