NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/comp/lang/c/19255

Text File | 1993-01-05 | 4.3 KB | 150 lines

Newsgroups: comp.lang.c
Path: sparky!uunet!uunet.ca!canrem!telly!druid!darcy
From: [e-mail] (D'Arcy J.M. Cain)
Subject: Re: strtok
Message-ID: <[e-mail]>
Date: Tue, 5 Jan 1993 12:34:30 GMT
Distribution: usa
References: <[e-mail]> <[e-mail]> <[e-mail]>
Organization: D'Arcy Cain Consulting
Lines: 138

[e-mail] (Paul Sander) writes:
>Robert Osborne posted a strtok replacement to comp.lang.c on September 12,
>1991 that returns strtok's global state in a parameter to the caller, making
>it reentrant.  The message-ID is [e-mail], and I have copies of it
>if anyone is interested.

Here's one I wrote based on Henry Spencer's strtok code.  I had to change
the name since the semantics are completely different than the standard.

/*
 * xstrtok.c
 * Written by D'Arcy J.M. Cain
 * based on code by Henry Spencer
 * modified to assume an ANSI compiler by D'Arcy J.M. Cain
 * because strtok is a little brain dead about multiple calls
 * this is a modified version which carries the pertinent information
 * in a structure.  This allows more than one string to be parsed
 * at the same time (the code is re-entrant.)
 *
 * Get next token from string x->s (NULL on 2nd, 3rd, etc. calls),
 * where tokens are nonempty strings separated by runs of
 * chars from delim.  Writes NULs into s to end tokens.  delim need not
 * remain constant from call to call.   If quote is set then quotes (single
 * or double) are respected.
 *
 * structure in string.h should look like this:
 * typedef struct {
 *    char        *scanpoint;        / * filled in by xstrtok * /
 *    char        *str2parse;        / * string to parse - set for first call * /
 *    const char    *delim;            / * string of delimiters * /
 *    int            quote;            / * respect quoting if set * /
 * } XSTRTOK;
 */

#include    <string.h>

char    *xstrtok(XSTRTOK *xinfo)
{
    char        *scan;
    char        *tok;
    char        qchar = 0;
    const char    *dscan;

    /* figure out what stage we are at */
    if (xinfo->str2parse == NULL && xinfo->scanpoint == NULL)
        return(NULL);            /* last one went last time */

    if (xinfo->str2parse != NULL)
    {
        scan = xinfo->str2parse;    /* this is the first time */
        xinfo->str2parse = NULL;
    }
    else
        scan = xinfo->scanpoint;    /* we have already started */

    /* Skip leading delimiters.  */
    for (; *scan; scan++)
    {
        for (dscan = xinfo->delim; *dscan; dscan++)
            if (*scan == *dscan)    /* found a delimiter */
                break;                /* from inner loop */

        if (*dscan == '\0')            /* this character not a delimiter */
            break;                    /* from outer loop */
    }

    /* are we finished with the line? */
    if (*scan == '\0')
    {
        xinfo->scanpoint = NULL;
        return(NULL);
    }

    /* otherwise we now point to the next token */
    tok = scan;

    /* find the end of the token */
    /* three versions for speedup rather than testing within loop */

    /* first find empty string such as "" */
    if (xinfo->quote && (*tok == '\"' || *tok == '\'') && tok[0] == tok[1])
    {
        *tok = 0;
        xinfo->scanpoint = tok + 2;
        return(tok);
    }

    /* non-empty string respecting quotes */
    if (xinfo->quote) for (; *scan != '\0'; scan++)
    {
        for (dscan = xinfo->delim; *dscan != '\0';)    /* increment is in body */
        {
            /* have we found a delimiter? */
            if (*scan == *dscan++)
            {
                xinfo->scanpoint = scan + 1;    /* point to next character */
                *scan = '\0';                /* terminate the token */
                return(tok);                /* and return it */
            }

            /* is it a quote character */
            if (*scan == '\'' || *scan == '\"')
            {
                qchar = *scan;                /* search for same close quote */
                strcpy(scan, scan + 1);        /* don't return quotes */

                while (*scan && *scan != qchar)
                    scan++;

                strcpy(scan, scan + 1);        /* don't return quotes */
            }
        }
    }
    /* normal strtok semantics */
    else for (; *scan != '\0'; scan++)
    {
        for (dscan = xinfo->delim; *dscan != '\0';)    /* increment is in body */
        {
            /* have we found a delimiter? */
            if (*scan == *dscan++)
            {
                xinfo->scanpoint = scan + 1;    /* point to next character */
                *scan = '\0';                    /* terminate the token */
                return(tok);                    /* and return it */
            }
        }
    }

    /* Reached end of string.  */
    xinfo->scanpoint = NULL;
    return(tok);
}

--
D'Arcy J.M. Cain ([e-mail])  |
D'Arcy Cain Consulting              |   There's no government
Toronto, Ontario, Canada            |   like no government!
[phone]          DoD#0082   |