[no subject]

[DLW@MIT-AI]

SI:FILE-WRITE-FINISH just got an error in the call to SUBSTRING
that is made when the first character of STR is not a colon.
The error was that the third arg to SUBSTRING was a -1, indicating
that STR was a null string.  Now STR seems to be a reply code, and
the third arg is one less than the array-active-length of the string,
so I conclude that the string was zero length and so the check
of the first character natuarally did not produce a colon.  Unfortunately,
when I went in to look at the string I was the string
":GO-AHEAD
", which as far as I can see should not have caused that error.  Is
STR a string that can be "changed out from under you" by asynchronous
action?  (There was an interesting bug once in which such a string
got interned...)