Lex and initial start conditions

[Mike Bryant]

In article <1990Jun3.024659.4122 at holos0.uucp> lbr at holos0.uucp (Len Reed) writes:
>In article <254 at samna.UUCP> jeff at samna.UUCP (Jeff Barber) writes:
>
>For real control of start conditions, use flex, the Berkeley fast lexical
>analyzer. It runs *far* faster than lex, both at lex time and run time, doesn't
>require a bunch of table scaling (% this and % that),  has enhancements
>to the start conditions, allows you to more control over the lexer program,
>is available in source form, and runs on DOS as well as Unix.  (That last
>item is unforunately important to me.)  Flex has been posted to the net.
>
>I'd never go back to lex....
>-- 
>Len Reed
>Holos Software, Inc.
>Voice: (404) 496-1358
>UUCP: ...!gatech!holos0!lbr

Does anyone know if the scanner generator Gamma-GLA has ever been
posted   to  the  net?  A paper on this tool was presented in the
Proceedings of the Summer 1988 USENIX  Conference  by  Robert  W.
Gray   from   the   University   of  Colorado. The paper showed a
benchmark of 10,000 lines of Pascal code run through several lex-
ical analyzers on a Sun 3/260 with the following results:

				Time
	Program		user	sys	total
	------------	-----	-----	-----
	gla		1.4	0.2	1.6
	flex (-cf)	3.18	0.24	3.42
	flex (-cem)	7.48	0.36	7.84
	lex		11.92	0.18	12.1

Also, when and where was Flex last posted to the net?  Thanks!
-- 
Mike Bryant      bryant at Summation.WA.COM
Summation Inc.   11335 NE 122nd Way  Kirkland, WA 98034