problem with -f flag in lex?

[Kris Kugel]

On the suggestion of a fellow netter, I tried recompiling some
lex code using the -f flag (no table compression) to see if I 
could speed it up some.

But when I tried to run it, it started dumping core.

I removed the -f flag, and the core dumps stopped.  I compared
the two output files from lex, and they seemed to differ only
in data.  I then tried tracing both versions, and the one that
didn't work failed to recognize the end of the input, and ran
off the end of the world.

the difference seems to be that the test in the statement
(near *yylastch = yych = input())
	if ( yyt == yycrank && !yyfirst )
fails on the -f version at the end of file, (yyt was == yycrank +1)
while the non-f version exits on a break inside the body of the if.

Does anybody know what would cause this problem?  is the -f flag
bad news?
	Kris A. Kugel
	Storage Tek:    ...{ uunet!nbires, ncar, ihnp4 }!stcvax!stc-auts!kak

(p.s this happended with ULTRIX and Sun 3.2, both BSD 4.2 variants)