AWK - Writing a parser using AWK

Author

Writing a parser using AWK

Cesar A. K. Grossmann

2004-07-05, 8:55 pm

Hi

Is it possible to write a parser using AWK, the same way you write a
parser for C, usting lex/yacc (flex/bison, if you use GNU software)?

I wrote a small parser using C this way, and I was wondering if it is
possible to do the same using an AWK script. I googled the internet, but
the most common references to parser and awk I get where about the
process of compiling gawk (that uses flex/bison to generate part of the
code), so I'm turning to the USENET.

The file I have to parse have only two kind of records:

IDENTIFIER = NUMBER STRING STRING;

IDENTIFIER = STRING { NUMBER IDENTIFIER [, NUMBER IDENTIFIER ] };

IDENTIFIER is a sequence of letters and numbers and '_', starting with a
letter, STRING is a sequence of chars between "", and NUMBER is a
number, with or without floating point. That's it. The records can be
spread in several lines, or in one line by themselves, e.g.:

IDENTIFIER =
NUMBER
STRING
STRING
;

IDENTIFIER =
STRING {
NUMBER IDENTIFIER,
NUMBER IDENTIFIER,
...
}
;

The parser I wrote in C works well. I just have a "programmer's itch"...
If someone knows about a page that deals with such things, please, tell
me about that.

TIA
P.S.: Sorry for the "engrish".
--
..O. Cesar A. K. Grossmann ICQ UIN: 35659423
...O http://www.LinuxByGrossmann.cjb.net/
OOO Quidquid Latine dictum sit, altum viditur

Ulrich M. Schwarz

2004-07-06, 3:55 am

"Cesar A. K. Grossmann" <cakgguard-usenet2004@yahoo.com.br> writes:

> Hi
>
> Is it possible to write a parser using AWK, the same way you write a
> parser for C, usting lex/yacc (flex/bison, if you use GNU software)?
>
> I wrote a small parser using C this way, and I was wondering if it is
> possible to do the same using an AWK script. I googled the internet,
> but the most common references to parser and awk I get where about the
> process of compiling gawk (that uses flex/bison to generate part of
> the code), so I'm turning to the USENET.
>
> The file I have to parse have only two kind of records:
>
> IDENTIFIER =3D NUMBER STRING STRING;
>
> IDENTIFIER =3D STRING { NUMBER IDENTIFIER [, NUMBER IDENTIFIER ] };

You'll be running into problems if your strings may contain ";",
otherwise, I'd set RS to ";" and go from there.

If you need a more complex RS: gawk has an auto-set variable that has
the part that matched RS in it. (In my case, fields were marked by
starting in column 0. Urgh!)

Ulrich
--=20
"P=3DNP. Zumindest f=FCr N=3D1."
-- Florian Diedrich

Cesar Rabak

2004-07-06, 3:56 pm

Cesar A. K. Grossmann escreveu:
> Hi
>
> Is it possible to write a parser using AWK, the same way you write a
> parser for C, usting lex/yacc (flex/bison, if you use GNU software)?

Doing it by hand (w/o flex/bison) should be possible. IIRC, the Henry
Spencer "Amazing awk Assembler"
http://www.math.utah.edu/docs/info/gawk_22.html does it that way.

--
Cesar Rabak

Stephan Titard

2004-07-13, 3:55 am

Hi
something lighter but quite mind-opening
is the preprocessor (written in awk) that Axel-Tobias Schreiner uses in his

Object-oriented programming in C.
google with ooc-94.2.11.tar.gz

enjoy
HTH

"Cesar A. K. Grossmann" wrote:

> Hi
>
> Is it possible to write a parser using AWK, the same way you write a
> parser for C, usting lex/yacc (flex/bison, if you use GNU software)?
>
> I wrote a small parser using C this way, and I was wondering if it is
> possible to do the same using an AWK script. I googled the internet, but
> the most common references to parser and awk I get where about the
> process of compiling gawk (that uses flex/bison to generate part of the
> code), so I'm turning to the USENET.
>
> The file I have to parse have only two kind of records:
>
> IDENTIFIER = NUMBER STRING STRING;
>
> IDENTIFIER = STRING { NUMBER IDENTIFIER [, NUMBER IDENTIFIER ] };
>
> IDENTIFIER is a sequence of letters and numbers and '_', starting with a
> letter, STRING is a sequence of chars between "", and NUMBER is a
> number, with or without floating point. That's it. The records can be
> spread in several lines, or in one line by themselves, e.g.:
>
> IDENTIFIER =
> NUMBER
> STRING
> STRING
> ;
>
> IDENTIFIER =
> STRING {
> NUMBER IDENTIFIER,
> NUMBER IDENTIFIER,
> ...
> }
> ;
>
> The parser I wrote in C works well. I just have a "programmer's itch"...
> If someone knows about a page that deals with such things, please, tell
> me about that.
>
> TIA
> P.S.: Sorry for the "engrish".
> --
> .O. Cesar A. K. Grossmann ICQ UIN: 35659423
> ..O http://www.LinuxByGrossmann.cjb.net/
> OOO Quidquid Latine dictum sit, altum viditur

Stephan Titard

2004-07-13, 8:55 am

I forgot to mention that there is no tool like lex/yacc
*you* have to write the parser by hand - but awk is quite capable for writing

top-down parsers.
another good starting point is m1 written by John Bentley

HTH
stephan

Stephan Titard wrote:
[color=darkred]
> Hi
> something lighter but quite mind-opening
> is the preprocessor (written in awk) that Axel-Tobias Schreiner uses in his
>
> Object-oriented programming in C.
> google with ooc-94.2.11.tar.gz
>
> enjoy
> HTH
>
> "Cesar A. K. Grossmann" wrote:
>

Stepan Kasal

2004-07-13, 8:55 am

Hello,

In article <40F39019.27C79A90@rest.is.fake>, Stephan Titard wrote:
> Object-oriented programming in C.
> google with ooc-94.2.11.tar.gz

google for ooc-02.01.04.tar.gz to get an updated rerelease.

Stepan Kasal

Stepan Kasal

2004-07-13, 8:55 am

Hello,

> another good starting point is m1 written by John Bentley

as a convenience service to others: The m1 macro processor was written by
*Jon* Bentley (slightly different spelling) and was published in the last
chapter of sed&awk by Dale Dougherty and Arnold Robbins (O'Reilly publ.).

When you go to http://oreilly.com/catalog/sed2 and download the examples,
you get not only m1.awk---the code---but also m1.ps---an article about m1.

Happy reading,
Stepan Kasal

Sponsored Links