|Lex surrogates lfcs.edinburgh.ac.uk!db@NSS.CS.UCL.AC.UK (Dave Berry) (1989-02-05)|
|Re: Lex surrogates schmidt@ORION.CF.UCI.EDU (Douglas C. Schmidt) (1989-02-05)|
|Re: Lex surrogates email@example.com (Vern Paxson) (1989-02-06)|
|Re: Lex surrogates rsalz@BBN.COM (Rich Salz) (1989-02-07)|
|Re: Lex surrogates wpl@PRC.Unisys.COM (1989-02-06)|
|Re: Lex surrogates firstname.lastname@example.org (Ken Yap) (1989-02-09)|
|Re: Lex surrogates email@example.com (1989-02-09)|
|Re: Lex surrogates tower@bu-cs.BU.EDU (1989-02-10)|
|[5 later articles]|
|Date:||Sun, 05 Feb 89 23:31:52 -0800|
|From:||"Douglas C. Schmidt" <schmidt@ORION.CF.UCI.EDU>|
In article <firstname.lastname@example.org> Dave Berry <lfcs.edinburgh.ac.uk!db writes:
++If Lex is as bad as these articles (and my experience) suggest, I'm surprised
++that GNU are using it for an optimising compiler.
Arrrrrggghh. I can't take the spread of misinformation any more! GNU
C and GNU C++ *don't* use LEX or FLEX or GNULEX (no, it doesn't
exist). The lexical analyzers are hand coded (I know, because I wrote
the perfect hash function that recognizes GNU C and C++ reserved words
in O(1) time). Please get the facts straight on this (which shouldn't
be too hard, since the source is freely available ;-)).
PS: They do use an LALR BISON (YACC) grammar, however.
[From "Douglas C. Schmidt" <schmidt@ORION.CF.UCI.EDU>]
Return to the
Search the comp.compilers archives again.