sed: code shrink

text    data     bss     dec     hex filename
 876354     493    7584  884431   d7ecf busybox_old
 876323     493    7584  884400   d7eb0 busybox_unstripped

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

editors/sed.c

@@ -14,49 +14,47 @@
  */
 
 /* Code overview.
+ *
+ * Files are laid out to avoid unnecessary function declarations.  So for
+ * example, every function add_cmd calls occurs before add_cmd in this file.
+ *
+ * add_cmd() is called on each line of sed command text (from a file or from
+ * the command line).  It calls get_address() and parse_cmd_args().  The
+ * resulting sed_cmd_t structures are appended to a linked list
+ * (G.sed_cmd_head/G.sed_cmd_tail).
+ *
+ * add_input_file() adds a FILE* to the list of input files.  We need to
+ * know all input sources ahead of time to find the last line for the $ match.
+ *
+ * process_files() does actual sedding, reading data lines from each input FILE *
+ * (which could be stdin) and applying the sed command list (sed_cmd_head) to
+ * each of the resulting lines.
+ *
+ * sed_main() is where external code calls into this, with a command line.
+ */
 
-  Files are laid out to avoid unnecessary function declarations.  So for
+/* Supported features and commands in this version of sed:
-  example, every function add_cmd calls occurs before add_cmd in this file.
+ *
-
+ * - comments ('#')
-  add_cmd() is called on each line of sed command text (from a file or from
+ * - address matching: num|/matchstr/[,num|/matchstr/|$]command
-  the command line).  It calls get_address() and parse_cmd_args().  The
+ * - commands: (p)rint, (d)elete, (s)ubstitue (with g & I flags)
-  resulting sed_cmd_t structures are appended to a linked list
+ * - edit commands: (a)ppend, (i)nsert, (c)hange
-  (G.sed_cmd_head/G.sed_cmd_tail).
+ * - file commands: (r)ead
-
+ * - backreferences in substitution expressions (\0, \1, \2...\9)
-  add_input_file() adds a FILE* to the list of input files.  We need to
+ * - grouped commands: {cmd1;cmd2}
-  know all input sources ahead of time to find the last line for the $ match.
+ * - transliteration (y/source-chars/dest-chars/)
-
+ * - pattern space hold space storing / swapping (g, h, x)
-  process_files() does actual sedding, reading data lines from each input FILE *
+ * - labels / branching (: label, b, t, T)
-  (which could be stdin) and applying the sed command list (sed_cmd_head) to
+ *
-  each of the resulting lines.
+ * (Note: Specifying an address (range) to match is *optional*; commands
-
+ * default to the whole pattern space if no specific address match was
-  sed_main() is where external code calls into this, with a command line.
+ * requested.)
-*/
+ *
-
+ * Todo:
-
+ * - Create a wrapper around regex to make libc's regex conform with sed
-/*
+ *
-	Supported features and commands in this version of sed:
+ * Reference http://www.opengroup.org/onlinepubs/007904975/utilities/sed.html
-
+ */
-	 - comments ('#')
-	 - address matching: num|/matchstr/[,num|/matchstr/|$]command
-	 - commands: (p)rint, (d)elete, (s)ubstitue (with g & I flags)
-	 - edit commands: (a)ppend, (i)nsert, (c)hange
-	 - file commands: (r)ead
-	 - backreferences in substitution expressions (\0, \1, \2...\9)
-	 - grouped commands: {cmd1;cmd2}
-	 - transliteration (y/source-chars/dest-chars/)
-	 - pattern space hold space storing / swapping (g, h, x)
-	 - labels / branching (: label, b, t, T)
-
-	 (Note: Specifying an address (range) to match is *optional*; commands
-	 default to the whole pattern space if no specific address match was
-	 requested.)
-
-	Todo:
-	 - Create a wrapper around regex to make libc's regex conform with sed
-
-	Reference http://www.opengroup.org/onlinepubs/007904975/utilities/sed.html
-*/
 
 //usage:#define sed_trivial_usage
 //usage:       "[-efinr] SED_CMD [FILE]..."
@@ -244,11 +242,13 @@ static int index_of_next_unescaped_regexp_delim(int delimiter, const char *str)
 		delimiter = -delimiter;
 	}
 
-	for (; (ch = str[idx]); idx++) {
+	for (; (ch = str[idx]) != '\0'; idx++) {
 		if (bracket >= 0) {
-			if (ch == ']' && !(bracket == idx - 1 || (bracket == idx - 2
+			if (ch == ']'
-					&& str[idx - 1] == '^')))
+			 && !(bracket == idx - 1 || (bracket == idx - 2 && str[idx - 1] == '^'))
+			) {
 				bracket = -1;
+			}
 		} else if (escaped)
 			escaped = 0;
 		else if (ch == '\\')
@@ -434,11 +434,47 @@ static int parse_subst_cmd(sed_cmd_t *sed_cmd, const char *substr)
  */
 static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
 {
+	static const char cmd_letters[] = "saicrw:btTydDgGhHlnNpPqx={}";
+	enum {
+		IDX_s = 0,
+		IDX_a,
+		IDX_i,
+		IDX_c,
+		IDX_r,
+		IDX_w,
+		IDX_colon,
+		IDX_b,
+		IDX_t,
+		IDX_T,
+		IDX_y,
+		IDX_d,
+		IDX_D,
+		IDX_g,
+		IDX_G,
+		IDX_h,
+		IDX_H,
+		IDX_l,
+		IDX_n,
+		IDX_N,
+		IDX_p,
+		IDX_P,
+		IDX_q,
+		IDX_x,
+		IDX_equal,
+		IDX_lbrace,
+		IDX_rbrace,
+		IDX_nul
+	};
+	struct chk { char chk[sizeof(cmd_letters)-1 == IDX_nul ? 1 : -1]; };
+
+	unsigned idx = strchrnul(cmd_letters, sed_cmd->cmd) - cmd_letters;
+
 	/* handle (s)ubstitution command */
-	if (sed_cmd->cmd == 's')
+	if (idx == IDX_s) {
 		cmdstr += parse_subst_cmd(sed_cmd, cmdstr);
+	}
 	/* handle edit cmds: (a)ppend, (i)nsert, and (c)hange */
-	else if (strchr("aic", sed_cmd->cmd)) {
+	else if (idx <= IDX_c) { /* a,i,c */
 		if ((sed_cmd->end_line || sed_cmd->end_match) && sed_cmd->cmd != 'c')
 			bb_error_msg_and_die("only a beginning address can be specified for edit commands");
 		for (;;) {
@@ -454,8 +490,9 @@ static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
 		/* "\anychar" -> "anychar" */
 		parse_escapes(sed_cmd->string, sed_cmd->string, strlen(cmdstr), '\0', '\0');
 		cmdstr += strlen(cmdstr);
+	}
 	/* handle file cmds: (r)ead */
-	} else if (strchr("rw", sed_cmd->cmd)) {
+	else if (idx <= IDX_w) { /* r,w */
 		if (sed_cmd->end_line || sed_cmd->end_match)
 			bb_error_msg_and_die("command only uses one address");
 		cmdstr += parse_file_cmd(/*sed_cmd,*/ cmdstr, &sed_cmd->string);
@@ -463,8 +500,9 @@ static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
 			sed_cmd->sw_file = xfopen_for_write(sed_cmd->string);
 			sed_cmd->sw_last_char = '\n';
 		}
+	}
 	/* handle branch commands */
-	} else if (strchr(":btT", sed_cmd->cmd)) {
+	else if (idx <= IDX_T) { /* :,b,t,T */
 		int length;
 
 		cmdstr = skip_whitespace(cmdstr);
@@ -475,7 +513,7 @@ static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
 		}
 	}
 	/* translation command */
-	else if (sed_cmd->cmd == 'y') {
+	else if (idx == IDX_y) {
 		char *match, *replace;
 		int i = cmdstr[0];
 
@@ -495,7 +533,7 @@ static const char *parse_cmd_args(sed_cmd_t *sed_cmd, const char *cmdstr)
 	/* if it wasnt a single-letter command that takes no arguments
 	 * then it must be an invalid command.
 	 */
-	else if (strchr("dDgGhHlnNpPqx={}", sed_cmd->cmd) == 0) {
+	else if (idx >= IDX_nul) { /* not d,D,g,G,h,H,l,n,N,p,P,q,x,=,{,} */
 		bb_error_msg_and_die("unsupported command %c", sed_cmd->cmd);
 	}
 
@@ -966,9 +1004,9 @@ static void process_files(void)
 			}
 			sed_cmd->in_match = !(
 				/* has the ending line come, or is this a single address command? */
-				(sed_cmd->end_line ?
+				(sed_cmd->end_line
-					sed_cmd->end_line == -1 ?
+					? sed_cmd->end_line == -1
-						!next_line
+						? !next_line
 						: (sed_cmd->end_line <= linenum)
 					: !sed_cmd->end_match
 				)