Re: [leafnode-list] Fetchnews Problem

[krasel@xxxxxxxxxxxxxxxxxxxxxxxxxxxx (Cornelius Krasel)]

Jeff Grossman wrote:

> No, 1.9.14 works fine for me.  For some reason with .16 and I think .15,
> Leafnode would not correctly put the last message retreived.  And, for some
> reason I was not getting all of the messages.  This only seemed to happen
> with the Supernews server.  I have since switched back to 1.9.14 and
> everything is working fine.

If this is true, it is probably the getaline() patch in 1.9.15 which
causes the trouble. If you feel adventurous (and are interested in
helping), you might try the following version of getaline (just remove
the old getaline() function from miscutil.c and put this one in):

/*
 * Lars Wirzenius: read a line into memory, with no max length
 * return a pointer to the line, or NULL in case of error
 *
 * strip \r at EOL
 * replace NUL bytes with spaces
 */
char *getaline(FILE *f) {
    static char *buf;       /* buffer for line */
    static size_t size;     /* size of buffer */
    size_t len;             /* # of chars stored into buf before '\0' */
    char * p;
    int i;

    len = 0;
    if (!size)
	size = 256;
    if (!buf)
	buf = critmalloc( size, "reading line" );

    while ((p=fgets(buf+len, size-len, f))) {
	/* replace NUL bytes (which shouldn't be there anyway)
	 * with spaces
	 */
	i = size - len;
	while ( *p++ != '\n' && i-- > 0 ) {
	    if ( *p == '\0' )
		*p = ' ';
	}
	len += strlen(buf+len);
	if ( buf[len-1] == '\n' )
	    break;		/* the whole line has been read */

	size += size;		/* exponential growth of buffer */
	buf = critrealloc(buf, size, "reading line" );
    }

    if ( len == 0 )
	return NULL;

    if (len && (buf[len-1] == '\n')) { /* go back on top of the newline */
	--len;
	if (len && (buf[len-1] == '\r')) /* also delete CR */
	    --len;
    }

    buf[len] = '\0';        /* unconditionally terminate string,
                               possibly overwriting newline */

    if ( debug )
        syslog( LOG_DEBUG, "<%s\n", buf );
    return buf;
}

I would be very interested whether your problems with Supernews
persist in that case.

--Cornelius.

-- 
/* Cornelius Krasel, U Wuerzburg, Dept. of Pharmacology, Versbacher Str. 9 */
/* D-97078 Wuerzburg, Germany   email: phak004@xxxxxxxxxxxxxxxxxxxxxx  SP4 */
/* "Science is the game we play with God to find out what His rules are."  */

-- 
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list