[Bug: 21.5-b18] ECB 2.27 makes XEmacs spike CPU usage and hang onCygWin

[Ben Wing]

> Warning: Error in 'pre-gc-hook' (stack-overflow (Lisp nesting exceeds
> 'max-lisp-eval-depth'))
> 
> Warning: Error in `post-gc-hook': Attempt to throw outside of 
> function to catch '((#<INTERNAL OBJECT (XEmacs bug?) 
> (symbol-value-forward type
> 13) 0x75db58 (error (byte-code ...
> 
> The rest I cannot retrieve since it hangs and cannot access 
> lossage. However the mini-buffer is cycling between those two 
> messages.
> 
> If I (setq max-lisp-eval-depth 4000) XEmacs crashes (3000 
> still exhibits the hang).
> 
> Unfortunately the crash does not provide any output and the 
> cygwin stackdump has 0 bytes.
> 
> Let me know what additional information I can provide or how 
> better could I reproduce.

Under cygwin you have to run under gdb to get results.

Try putting a breakpoint on signal_1() and calling debug_backtrace() so we
can see the Lisp backtrace.