Re: lex/flex and source correspondence

Thomas Dickey <dickey@saltmine.radix.net>
Thu, 29 Nov 2007 14:52:10 -0000

          From comp.compilers

Related articles
lex/flex and source correspondence parthaspanda22@gmail.com (2007-11-06)
Re: lex/flex and source correspondence gene.ressler@gmail.com (Gene) (2007-11-08)
Re: lex/flex and source correspondence dickey@saltmine.radix.net (Thomas Dickey) (2007-11-29)
| List of all articles for this month |

From: Thomas Dickey <dickey@saltmine.radix.net>
Newsgroups: comp.compilers
Date: Thu, 29 Nov 2007 14:52:10 -0000
Organization: RadixNet Internet Services
References: 07-11-023
Keywords: lex, performance
Posted-Date: 30 Nov 2007 00:00:16 EST

parthaspanda22@gmail.com wrote:
> A flex generated scanner could decide to store the entire input file
> in memory. That way, a single buffer can be used in several places in
> the tool chain that uses a scanner.


> Is this an idea that has been explored and thought of as unworthy?


> I know that multiple input files could increase the size of the
> buffer(s) but hey, we are living in the age of 100s of MB of main
> memory.


> Sincerely.
> [This is a very old idea. Turbo Pascal buffered full source files in
> RAM in the 1980s. -John]


Either lex or flex can do this by redefining the input/outputs.
However, flex can be used to support multiple parsers (by restricting
the exported symbols), while for lex it's (ahem) doable but not
simple.


Regarding buffering full source files - even as main memory's
increased, there are still text files that people want to edit which
are larger. So it's not a universal solution...
--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.