|redefinable input() unput(): want old lex/flex de-POSIXed email@example.com (Douglas Ray) (2008-12-16)|
|Re: redefinable input() unput(): want old lex/flex de-POSIXed firstname.lastname@example.org (George Neuner) (2008-12-16)|
|From:||Douglas Ray <email@example.com>|
|Date:||Tue, 16 Dec 2008 00:11:04 +1100|
|Organization:||Melbourne PC User Group Inc|
|Posted-Date:||15 Dec 2008 12:43:57 EST|
I'm revisiting a parser from 1991-1993, last compiled on SunOS (BSD, pre Solaris).
Where are the patches to Flex to enable redefining input() and
unput()? Surely someone's done this? Or is there some other LEX
implementation that escaped POSIX, and will compile on NetBSD
The project is a shell, has > 1000 lines related to
completion/expansion mechanisms of input tokens which I do not want to
relearn/rewrite just for POSIX disabilities.
Sorry if this is a FAQ. Thanks for any pointers,
[I've just been reviewing the history of lex and flex for my upcoming flex&bison,
and you have a problem. The internal structure of flex is utterly unlike that of
lex, and it's hard to imagine a patch to let you redefine input and unput(). The
ancient v7 sources are available, although I suspect that making them compile will
be hard. What version of lex does opensolaris have? -John]
Return to the
Search the comp.compilers archives again.