From 7142994367871a0271b98242e72323362898f7d8 Mon Sep 17 00:00:00 2001
From: DTB <trinity@trinity.moe>
Date: Fri, 26 Apr 2024 19:08:48 -0600
Subject: [PATCH] pg(1): start work on a quoted strtok(3) and internal
 environment

---
 src/pg.c | 88 +++++++++++++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 77 insertions(+), 11 deletions(-)

diff --git a/src/pg.c b/src/pg.c
index dfa21f9..5a912aa 100644
--- a/src/pg.c
+++ b/src/pg.c
@@ -6,7 +6,7 @@
  * the terms of the GNU Affero General Public License as published by the Free
  * Software Foundation, either version 3 of the License, or (at your option) any
  * later version.
- * 
+ *
  * This program is distributed in the hope that it will be useful, but WITHOUT
  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
  * FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more
@@ -42,6 +42,70 @@ static FILE *input;
 static char *prompt = ": ";
 static char *program_name = "pg";
 
+static char *permute_out(char *str, size_t index){
+	size_t j;
+
+	for(j = index; str[j - 1] != '\0'; ++j)
+		str[j - 1] = str[j];
+
+	return str;
+}
+
+/* strtok(3p), but supports double-quotes and escapes (but only for escaping
+ * quotes). UTF-8 is safe only in str. Unmatched quotes in str are considered
+ * literal. The behavior of strtok_quoted when '"' or '\\' are in sep is
+ * undefined. */
+/* TODO: Seems to only ever return NULL. */
+static char *strtok_quoted(char *str, char *sep){
+	static char *s;
+	size_t i;
+	size_t j;
+
+	if(str != NULL)
+		s = str;
+
+	while(strchr(sep, *s) == NULL)
+		if(*++s == '\0')
+			return NULL; /* no remaining characters except seperators */
+
+	{
+		char in_escape;
+		int in_quotes;
+
+		in_escape = 0;
+		in_quotes = -1;
+		for(i = 0; s[i] != '\0'; ++i)
+			switch(s[i]){
+			case '\\':
+				if(in_escape)
+					permute_out(s, i--);
+				in_escape = !in_escape;
+				break;
+			case '"':
+				if(in_escape){
+					s[i] = s[i - 1];
+					permute_out(s, i--);
+				}else if(in_quotes != -1){
+					permute_out(s, in_quotes);
+					--i;
+					permute_out(s, i--);
+					in_quotes = -1;
+				}else
+					in_quotes = i;
+				break;
+			case '\0':
+				return s;
+			default:
+				if(!in_escape && strchr(sep, s[i]) != NULL){
+					s[i] = '\0';
+					return s;
+				}
+			}
+	}
+
+	return s;
+}
+
 /* Page at most l bytes from f without consideration of a buffer (print them to
  * stdout). */
 static int pg_b_u(FILE *f, size_t l){
@@ -66,8 +130,8 @@ static int pg_l_u(FILE *f, size_t l, char nl){
 	return c;
 }
 
-static int cmd_quit(int argc, char **argv){ return EX_UNAVAILABLE; }
-static int cmd_default_page(int argc, char **argv){
+static int cmd_quit(int argc, char **argv, char **envp){return EX_UNAVAILABLE;}
+static int cmd_default_page(int argc, char **argv, char **envp){
 
 	if(argc > 1) /* This shouldn't be possible. */
 		return EX_USAGE;
@@ -82,11 +146,11 @@ static int cmd_default_page(int argc, char **argv){
 	else
 		return EX_SOFTWARE;
 }
-static int cmd_page_down_lines(int argc, char **argv){
+static int cmd_page_down_lines(int argc, char **argv, char **envp){
 
 	switch(argc){
 	case 1:
-		return cmd_default_page(argc, argv);
+		return cmd_default_page(argc, argv, envp);
 	case 2: /* not implemented */
 	default:
 		fprintf(stderr, "Usage: %s" /*" (lines)"*/ "\n", argv[0]);
@@ -95,13 +159,15 @@ static int cmd_page_down_lines(int argc, char **argv){
 }
 
 /* A CmdMap must be NULL-terminated. */
-static struct CmdMap{ char *name; int (*fn)(int, char **); } builtins[] = {
+static struct CmdMap{ char *name; int (*fn)(int, char **, char **); }
+builtins[] = {
 	{ "", cmd_default_page },
 	{ "+", cmd_page_down_lines },
 
 	/* don't make the user feel trapped */
-	{ "exit", cmd_quit }, { "q", cmd_quit }, { ":q", cmd_quit },
-	{ ":q!", cmd_quit }, { "quit", cmd_quit }, { "ZZ", cmd_quit },
+	{ ":q", cmd_quit }, { ":q!", cmd_quit },
+	{ "exit", cmd_quit }, { "q", cmd_quit }, { "Q", cmd_quit },
+	{ "quit", cmd_quit }, { "ZZ", cmd_quit },
 	
 	{ NULL, NULL }
 };
@@ -109,7 +175,7 @@ static struct CmdMap{ char *name; int (*fn)(int, char **); } builtins[] = {
 #define ARGV_MAX 10
 /* Find and execute the command in the command map, given a corresponding
  * command line. */
-static int cmdline_exec(struct CmdMap *map, char *cmdline){
+static int cmdline_exec(struct CmdMap *map, char *cmdline, char **envp){
 	/* Command line word splitting is naive and based on whitespace ONLY; no
 	 * fancy quoting or escaping here. Adding that would (ideally) entail
 	 * replacing strtok(3) with something specific to this task. */
@@ -128,7 +194,7 @@ static int cmdline_exec(struct CmdMap *map, char *cmdline){
 
 	for(; map->name != NULL; map = &map[1])
 		if(strcmp(map->name, argv[0]) == 0)
-			return map->fn(argc, argv);
+			return map->fn(argc, argv, envp);
 
 	fprintf(stderr, "%s: %s: not found\n", program_name, argv[0]);
 	return EX_USAGE;
@@ -182,7 +248,7 @@ int main(int argc, char *argv[]){
 			return EX_OK;
 
 		{	int r;
-			switch((r = cmdline_exec(builtins, (char *)cmd))){
+			switch((r = cmdline_exec(builtins, (char *)cmd, NULL))){
 			case EX_OK: case EX_USAGE: break;
 			case EX_UNAVAILABLE: return EX_OK;
 			default: return r;