"Version not numeric," and debugger suggestion.

[David A. Moon <Moon at SCRC-STONY-BROOK>]

This is not a reply to your bug with zwei (probably really a bug with Release 5
pathnames).

    Date: Friday, 8 March 1985, 22:28-EST
    From: David Vinayak Wallace <Gumby at MC>

    In Symbolics 3600 Zwei in System 242.356, Hardcopy 20.10, FEP 18, on Lisp Machine Jimi Hendrix:

    I did system-e, c-X c-F.  The default was mc:gumby;foo >.
    I typed h:l;format<return>.  H: is a logical host pointing to a lispm.
    I got the error "Version not numeric."  Huh?  More likely it meant I
    should have entered a file type too, no?  

    In any case, there was no way to see what actually promped this error.
    In EMACS, you could always type "?" at an error message to enter the
    TECO debugger.  Perhaps when zwei intercepts an error and echos it in
    the modeline pane it should keep the stack context, and enter the
    debugger if, say, the escape key is typed?

The way we usually do this is to evaluate (setq trace-conditions t) and
then reproduce the problem.  Once you're done, (setq trace-conditions nil)
so you don't get constant trips to the Debugger afterwards.  This is
documented.