Audit nl(1)

1) Refactor the manpage to use the num-syntax and concise wording.
2) Build format instead of having a list of static strings.
3) BUGFIX: if (!buf[0] || buf[0] == '\n') Process last-read-line
           properly.
4) BUGFIX: In case we hit a formatting line, print a newline instead
           of just dropping it.
5) Use a switch instead of having spaghetti-cases.
6) Don't use printf-magic but explicitly do a putchar(' ')-loop.
7) Update usage(), indent properly.
8) BUGFIX: strchr is not NULL when type[0] is \0. Check for \0
           separately beforehand.
9) Reorder arg.h-cases for better readability.

README

@@ -49,7 +49,7 @@ The following tools are implemented ('*' == finished, '#' == UTF-8 support,
 =*| mktemp          non-posix                    none
 =*| mv              yes                          none (-i)
 =*| nice            yes                          none
-#*  nl              yes                          none
+#*| nl              yes                          none
 =*| nohup           yes                          none
 #*| paste           yes                          none
 =*| printenv        non-posix                    none

nl.1

@@ -1,4 +1,4 @@
-.Dd March 18, 2015
+.Dd March 22, 2015
 .Dt NL 1
 .Os sbase
 .Sh NAME
@@ -11,18 +11,19 @@
 .Op Fl d Ar delim
 .Op Fl f Ar type
 .Op Fl h Ar type
-.Op Fl i Ar incr
+.Op Fl i Ar num
 .Op Fl l Ar num
 .Op Fl n Ar format
 .Op Fl s Ar sep
-.Op Fl v Ar startnum
+.Op Fl v Ar num
-.Op Fl w Ar width
+.Op Fl w Ar num
 .Op Ar file
 .Sh DESCRIPTION
 .Nm
-reads lines from the named
+reads lines from
 .Ar file
-and writes them to stdout with non-empty lines numbered. If no
+and writes them to stdout, numbering non-empty lines.
+If no
 .Ar file
 is given
 .Nm
@@ -30,64 +31,61 @@ reads from stdin.
 .Sh OPTIONS
 .Bl -tag -width Ds
 .It Fl p
-Do not reset number for logical pages
+Do not reset line number for logical pages.
-.It Fl b Ar type
+.It Fl h Ar type | Fl b Ar type | Fl f Ar type
-Defines which lines will be numbered for body sections:
+Define which lines to number in the head | body | footer section:
 .Bl -tag -width pstringXX
 .It a
 All lines.
 .It n
 No lines.
 .It t
-Only non-empty lines (default).
+Only non-empty lines. This is the default.
 .It p Ns Ar expr
-Only lines which match
+Only lines matching
-.Ar expr ,
+.Ar expr
-a regular expression as defined in
+according to
 .Xr regex 7 .
 .El
 .It Fl d Ar delim
-Specify the delimiter (default is "\\:"). If only one character is specified, the second remains ':'.
+Set
-.It Fl f Ar type
+.Ar delim
-Same as
+as the delimiter for logical pages. If
-.Fl b
+.Ar delim
-except for footer sections.
+is only on character,
-.It Fl h Ar type
+.Nm
-Same as
+appends ":" to it. The default is "\e:".
-.Fl b
+.It Fl i Ar num
-except for header sections.
+Set the increment between numbered lines to
-.It Fl i Ar incr
+.Ar num .
-Defines the increment between numbered lines.
 .It Fl l Ar num
-Specify the number of adjacent blank lines to be considered as one. Default is 1.
+Set the number of adjacent blank lines to be considered as one to
+.Ar num .
+The default is 1.
 .It Fl n Ar format
-Specify the line number output format.
+Set the line number output
-The
 .Ar format
-can be any of the following:
+to one of:
 .Bl -tag -width pstringXX
 .It ln
 Left justified.
 .It rn
-Right justified.
+Right justified. This is the default.
 .It rz
 Right justified with leading zeroes.
 .El
 .Pp
-The default
-.Ar format
-is rn.
 .It Fl s Ar sep
-Defines the string used to separate line numbers and lines. By default this is
+Use
-a tab.
+.Ar sep
-.It Fl v Ar startnum
+to separate line numbers and lines. The default is "\et".
-Start counting from
+.It Fl v Ar num
-.Ar startnum
+Start counting lines from
-instead of the default 1.
+.Ar num .
-.It Fl w Ar width
+The default is 1.
-The number of characters to be occupied by the line number
+.It Fl w Ar num
-will be set to
+Set the width of the line number to
-.Ar width .
+.Ar num .
 The default is 6.
 .El
 .Sh SEE ALSO

nl.c

@@ -9,40 +9,33 @@
 #include "utf.h"
 #include "util.h"
 
-/* formats here specify line number and separator (not line content) */
+static size_t   startnum = 1;
-#define FORMAT_LN "%-*ld%s"
+static size_t   incr = 1;
-#define FORMAT_RN "%*ld%s"
+static size_t   blines = 1;
-#define FORMAT_RZ "%0*ld%s"
+static size_t   delimlen = 2;
-
+static int      width = 6;
-static char        type[] = { 'n', 't', 'n' }; /* footer, body, header */
+static int      pflag = 0;
-static char       *delim = "\\:";
+static char     type[] = { 'n', 't', 'n' }; /* footer, body, header */
-static const char *format = FORMAT_RN;
+static char    *delim = "\\:";
-static const char *sep = "\t";
+static char     format[8] = "%*ld%s";
-static int         width = 6;
+static char    *sep = "\t";
-static int         pflag = 0;
+static regex_t  preg[3];
-static size_t      startnum = 1;
-static size_t      incr = 1;
-static size_t      blines = 1;
-static size_t      delimlen = 2;
-static regex_t     preg[3];
 
 static int
 getsection(char *buf, int *section)
 {
-	int sectionchanged = 0;
+	int sectionchanged = 0, newsection = *section;
-	int newsection = *section;
 
 	for (; !strncmp(buf, delim, delimlen); buf += delimlen) {
 		if (!sectionchanged) {
 			sectionchanged = 1;
 			newsection = 0;
 		} else {
-			++newsection;
+			newsection = (newsection + 1) % 3;
-			newsection %= 3;
 		}
 	}
 
-	if (buf && buf[0] == '\n')
+	if (!buf[0] || buf[0] == '\n')
 		*section = newsection;
 	else
 		sectionchanged = 0;
@@ -51,27 +44,33 @@ getsection(char *buf, int *section)
 }
 
 static void
-nl(const char *name, FILE *fp)
+nl(const char *fname, FILE *fp)
 {
-	char *buf = NULL;
+	size_t i, number = startnum, size = 0;
 	int donumber, oldsection, section = 1, bl = 1;
-	size_t number = startnum, size = 0;
+	char *buf = NULL;
 
-	while (getline(&buf, &size, fp) != -1) {
+	while (getline(&buf, &size, fp) >= 0) {
 		donumber = 0;
 		oldsection = section;
 
 		if (getsection(buf, &section)) {
 			if ((section >= oldsection) && !pflag)
 				number = startnum;
+			putchar('\n');
 			continue;
 		}
 
-		if ((type[section] == 't' && buf[0] != '\n')
+		switch (type[section]) {
-		    || (type[section] == 'p' &&
+		case 't':
-		        !regexec(&preg[section], buf, 0, NULL, 0))) {
+			if (buf[0] != '\n')
-			donumber = 1;
+				donumber = 1;
-		} else if (type[section] == 'a') {
+			break;
+		case 'p':
+		     	if (!regexec(preg + section, buf, 0, NULL, 0))
+				donumber = 1;
+			break;
+		case 'a':
 			if (buf[0] == '\n' && bl < blines) {
 				++bl;
 			} else {
@@ -84,28 +83,30 @@ nl(const char *name, FILE *fp)
 			printf(format, width, number, sep);
 			number += incr;
 		} else {
-			printf("%*s", width, "");
+			for (i = 0; i < width; i++)
+				putchar(' ');
 		}
-		printf("%s", buf);
+		fputs(buf, stdout);
 	}
 	free(buf);
 	if (ferror(fp))
-		eprintf("%s: read error:", name);
+		eprintf("getline %s:", fname);
 }
 
 static void
 usage(void)
 {
-	eprintf("usage: %s [-p] [-b type] [-d delim] [-f type] "
+	eprintf("usage: %s [-p] [-b type] [-d delim] [-f type]\n"
-	    "[-h type] [-i incr] [-l num]\n[-n format] [-s sep] "
+	        "       [-h type] [-i num] [-l num] [-n format]\n"
-	    "[-v startnum] [-w width] [file]\n", argv0);
+		"       [-s sep] [-v num] [-w num] [file]\n", argv0);
 }
+
 static char
 getlinetype(char *type, regex_t *preg)
 {
 	if (type[0] == 'p')
-		eregcomp(preg, &type[1], REG_NOSUB);
+		eregcomp(preg, type + 1, REG_NOSUB);
-	else if (!strchr("ant", type[0]))
+	else if (!type[0] || !strchr("ant", type[0]))
 		usage();
 
 	return type[0];
@@ -115,13 +116,10 @@ int
 main(int argc, char *argv[])
 {
 	FILE *fp;
-	char *d;
 	size_t l, s;
+	char *d, *formattype, *formatblit;
 
 	ARGBEGIN {
-	case 'b':
-		type[1] = getlinetype(EARGF(usage()), &preg[1]);
-		break;
 	case 'd':
 		d = EARGF(usage());
 		l = utflen(d);
@@ -131,9 +129,9 @@ main(int argc, char *argv[])
 			break;
 		case 1:
 			s = strlen(d);
-			delim = emalloc(s + 2);
+			delim = emalloc(s + 1 + 1);
-			estrlcpy(delim, d, s + 2);
+			estrlcpy(delim, d, s + 1 + 1);
-			estrlcat(delim, ":", s + 2);
+			estrlcat(delim, ":", s + 1 + 1);
 			delimlen = s + 1;
 			break;
 		default:
@@ -143,27 +141,36 @@ main(int argc, char *argv[])
 		}
 		break;
 	case 'f':
-		type[0] = getlinetype(EARGF(usage()), &preg[0]);
+		type[0] = getlinetype(EARGF(usage()), preg);
+		break;
+	case 'b':
+		type[1] = getlinetype(EARGF(usage()), preg + 1);
 		break;
 	case 'h':
-		type[2] = getlinetype(EARGF(usage()), &preg[2]);
+		type[2] = getlinetype(EARGF(usage()), preg + 2);
 		break;
 	case 'i':
 		incr = estrtonum(EARGF(usage()), 0, MIN(LLONG_MAX, SIZE_MAX));
 		break;
 	case 'l':
-		blines = estrtonum(EARGF(usage()), 0, UINT_MAX);
+		blines = estrtonum(EARGF(usage()), 0, MIN(LLONG_MAX, SIZE_MAX));
 		break;
 	case 'n':
-		format = EARGF(usage());
+		formattype = EARGF(usage());
-		if (!strcmp(format, "ln"))
+		estrlcpy(format, "%", sizeof(format));
-			format = FORMAT_LN;
+
-		else if (!strcmp(format, "rn"))
+		if (!strcmp(formattype, "ln")) {
-			format = FORMAT_RN;
+			formatblit = "-";
-		else if (!strcmp(format, "rz"))
+		} else if (!strcmp(formattype, "rn")) {
-			format = FORMAT_RZ;
+			formatblit = "";
-		else
+		} else if (!strcmp(formattype, "rz")) {
-			eprintf("%s: bad format\n", format);
+			formatblit = "0";
+		} else {
+			eprintf("%s: bad format\n", formattype);
+		}
+
+		estrlcat(format, formatblit, sizeof(format));
+		estrlcat(format, "*ld%s", sizeof(format));
 		break;
 	case 'p':
 		pflag = 1;
@@ -184,7 +191,7 @@ main(int argc, char *argv[])
 	if (argc > 1)
 		usage();
 
-	if (argc == 0) {
+	if (!argc) {
 		nl("<stdin>", stdin);
 	} else {
 		if (!(fp = fopen(argv[0], "r")))
@@ -192,5 +199,6 @@ main(int argc, char *argv[])
 		nl(argv[0], fp);
 		fclose(fp);
 	}
+
 	return 0;
 }