comm(1) (was: Re: uniq truncates lines > 2048 bytes)

# Kris Kennaway:
> On Wed, Jan 26, 2005 at 09:10:47AM +1100, Tim Robbins wrote:
> >
> > I was planning on going through all text-processing utilities in the base
> > system some time and either fixing line length problems or documenting them,
> > similar to what I did with multibyte character support. I may make a start
> > at that today.
> 
> If someone could fix comm(1) that would be a big help for me, because
> I have a local hack I have to carry around in all of my local package
> source trees.

Does this patch help? 

I haven't made any contributions before, so I just hope I'm
not making a fool of myself with this...

HTH,
Mario


Index: comm.c
===================================================================
RCS file: /home/ncvs/src/usr.bin/comm/comm.c,v
retrieving revision 1.21
diff -u -r1.21 comm.c
--- comm.c	2 Jul 2004 22:48:29 -0000	1.21
+++ comm.c	29 Jan 2005 01:30:13 -0000
_at__at_ -66,6 +66,8 _at__at_
 FILE   *file(const char *);
 void	show(FILE *, const char *, const wchar_t *, wchar_t *);
 int     wcsicoll(const wchar_t *, const wchar_t *);
+wchar_t *fgetwla(FILE *);
+
 static void	usage(void);
 
 int
_at__at_ -75,7 +77,7 _at__at_
 	int ch, flag1, flag2, flag3, iflag;
 	FILE *fp1, *fp2;
 	const wchar_t *col1, *col2, *col3;
-	wchar_t line1[MAXLINELEN], line2[MAXLINELEN];
+	wchar_t *line1, *line2;
 	const wchar_t **p;
 
 	flag1 = flag2 = flag3 = 1;
_at__at_ -123,12 +125,14 _at__at_
 	for (read1 = read2 = 1;;) {
 		/* read next line, check for EOF */
 		if (read1) {
-			file1done = !fgetws(line1, MAXLINELEN, fp1);
+			line1 = fgetwla(fp1);
+			file1done = (line1 == NULL);
 			if (file1done && ferror(fp1))
 				err(1, "%s", argv[0]);
 		}
 		if (read2) {
-			file2done = !fgetws(line2, MAXLINELEN, fp2);
+			line2 = fgetwla(fp2);
+			file2done = (line2 == NULL);
 			if (file2done && ferror(fp2))
 				err(1, "%s", argv[1]);
 		}
_at__at_ -170,6 +174,10 _at__at_
 			if (col2)
 				(void)printf("%ls%ls", col2, line2);
 		}
+		if (read1)
+			free(line1);
+		if (read2)
+			free(line2);		
 	}
 	exit(0);
 }
_at__at_ -177,10 +185,15 _at__at_
 void
 show(FILE *fp, const char *fn, const wchar_t *offset, wchar_t *buf)
 {
+	wchar_t *line;
 
+	
+	line = buf;
 	do {
-		(void)printf("%ls%ls", offset, buf);
-	} while (fgetws(buf, MAXLINELEN, fp));
+		(void)printf("%ls%ls", offset, line);
+		free(line);
+		line = fgetwla(fp);
+	} while (line != NULL);
 	if (ferror(fp))
 		err(1, "%s", fn);
 }
_at__at_ -208,7 +221,13 _at__at_
 int
 wcsicoll(const wchar_t *s1, const wchar_t *s2)
 {
-	wchar_t *p, line1[MAXLINELEN], line2[MAXLINELEN];
+	wchar_t *p, *line1, *line2;
+	int comp;
+
+	if( (line1 = malloc(wcslen(s1) + 1)) == NULL)
+		err(1, "%s", __func__);
+	if( (line2 = malloc(wcslen(s2) + 1)) == NULL)
+		err(1, "%s", __func__);
 
 	for (p = line1; *s1; s1++)
 		*p++ = towlower(*s1);
_at__at_ -216,5 +235,28 _at__at_
 	for (p = line2; *s2; s2++)
 		*p++ = towlower(*s2);
 	*p = '\0';
-	return (wcscoll(line1, line2));
+	
+	comp = wcscoll(line1, line2);
+	free(line1);
+	free(line2);
+	
+	return comp;
+}
+
+wchar_t *
+fgetwla(FILE *fp)
+{
+	wchar_t *line, *wtmp;
+	size_t len;
+	
+	if ( (wtmp=fgetwln(fp, &len)) == NULL) 
+		return NULL;
+	
+	if ( (line=malloc((len + 1) * sizeof(wchar_t))) == NULL) 
+		err(1, "%s", __func__);
+	
+	wcsncpy(line, wtmp, len);
+	line[len] = L'\0';
+
+	return line;
 }

--