Date: Tue, 10 Nov 1998 10:26:33 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Sahab Yazdani <sahaby AT geocities DOT com>
cc: djgpp AT delorie DOT com
Subject: Re: Parsing...
In-Reply-To: <364778A4.86092F30@geocities.com>
Message-ID: <Pine.SUN.3.91.981110102604.26796G-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp AT delorie DOT com


On Mon, 9 Nov 1998, Sahab Yazdani wrote:

> #include <stdio.h>
> 
> to this string
> 
> stdio.h

If this is as simple as it sounds, just use `sscanf', like this:

   sscanf (buffer, " # include <%[^>]> \n", parsed_string);

(It's amazing how many people overlook the tremendous power of
formatted input!)  For bullet-proof code, test the value returned by
`sscanf': it should be 1 if all's well; if it's not, back up and
recover.

Note that the C Standard allows whitespace both before and after the
leading `#', and sscanf will eat up zero or more whitespace characters
when it sees a blank in the format string.  That is why the format
above looks a bit strange at first sight.

> BTW if you are wondering why I would want to start parsing C/C++ files,
> I'm doing it because I want to start developing a Source Browser for
> DJGPP

If that is your goal, I would suggest to get as much help as you can
from the available tools instead of reinventing the wheel.  For
example, the -M option to GCC will cause it to generate all the
information about header files included by a given source file, in a
format that's much easier to read.

In those cases where you do need to read the source, I'd suggest to
use Flex if you need to parse the file.  Flex will generate C code
which is much faster in the case where the source file must be parsed
for multiple alternatives.  (A port of Flex is avialable in
v2gnu/flx254b.zip from the usual DJGPP sites.)