Path: sranha!katsu
From: katsu@sra.co.jp (WATANABE Katsuhiro)
Message-ID: <KATSU.90Dec18220948@sran14.sra.co.jp>
Date: 18 Dec 90 22:09:48
Organization: Software Research Associates, Inc.,Japan
In-reply-to: rkc@xn.ll.mit.edu's message of 11 Dec 90 21:14:26 GMT
Newsgroups: comp.unix.questions
Subject: Re: sed: linewraps and / and * characters
Distribution: 
References: <1990Dec11.211426.11924@xn.ll.mit.edu>

# Sorry for my poor English.

In article <1990Dec11.211426.11924@xn.ll.mit.edu> rkc@xn.ll.mit.edu (Rob Cunningham) writes:

> convert input that looks like this:
> 
> /* 
> 	  Some Text Here
> */
> 8888.999
> 
> /****
> ...
> 
> To input that looks like this
> /* 
> 	  Some Text Here
> */
> 4.3
> 
> /****

> 	  I know sed should be able to do this, but I can't figure out

  I think it is a job for awk or perl.(Because sed have no arithmetic
operator/function, it is hard to convert 8888.999 to 4.3 .)
  How about this awk script?

--- cut --- cut --- cut ---
BEGIN {FS="."}
/^\/\*/, /^\*\// {print; next}
NF == 0 {print; next}
{
	i = 1;
	while ($1 >= 10) {i++; $1 /= 10}
	if (NF > 1) {
		j = 1;
		while ($2 > 10) {j++; $2 /= 10}
	} else {
		j = 0;
	}
	printf("%d.%d\n", i, j);
}
--- cut --- cut --- cut ---

  If I were certain that it did not contain TAB, SPC, and any other
stuffs in lines not commented out, I would use length().

BUGS:
    9.90000 will be transformed into 1.5, while 9.9 into 1.1 .
    Both 0.9 and .9 will be transformed into 1.1 .
    Of course, comment marks (/* */) must show at the head of the line.

--
----____----____
WATANABE Katsuhiro	Software Research Associates, Inc. Japan.
Not execute, but evaluate.