WIP send command parser

4 files changed, 116 insertions, 19 deletions

diff --git a/client/cmd.cpp b/client/cmd.cpp
index 4a2c8a3..1ec2cb8 100644
--- a/client/cmd.cpp
+++ b/client/cmd.cpp
@@ -48,9 +48,10 @@ void cmd_send(char* addr_str) {
 
 	char* data;
 	size_t data_size;
-	if (strtodata(data_str, &data, &data_size)) {
+	int err = strtodata(data_str, &data, &data_size);
+	if (err <= 0) {
 		printf("data format error at index %d:\n%s\n%*s^\n",
-				(int) data_size, data_str, (int) data_size, "");
+				-err, data_str, -err, "");
 		return;
 	}
 
diff --git a/client/parse.cpp b/client/parse.cpp
index d15207c..223dc5d 100644
--- a/client/parse.cpp
+++ b/client/parse.cpp
@@ -1,31 +1,116 @@
+#include <math.h>
 #include <stdlib.h>
 #include <string.h>
+#include <stdio.h>
 
 #include "parse.h"
 
-static unsigned ifsrun(const char* input, const char* ift) {
-	unsigned i;
-	for (i = 0; input[i] != '\0' && strchr(ift, input[i]); i++);
-	return i;
+static int parse_str(const char* str, char* data, size_t* size) {
+	char closing = str[0];
+	char escape = false;
+	bool scan = data == NULL;
+	int i = 0;
+	size_t len = strlen(str);
+
+	switch (str[i]) {
+		case '\'':
+			escape = false;
+			break;
+		case '\"':
+			escape = true;
+			break;
+		default:
+			return -i;
+	}
+
+	for (i = 1; i < len && str[i] != '\0'; i++) {
+		char c = str[i];
+
+		if (c == closing) {
+			if (scan) printf("string%s of length %d\n", escape ? " (w/ escape)" : "", i - 1);
+			return i + 1; // +1 for closing quote
+		}
+
+		if (scan) *size += 1;
+	}
+
+	return -i;
+}
+
+static int parse_num(const char* str, char* data, size_t* size) {
+	const char* ifs = IFS;
+	size_t len = strcspn(str, ifs);
+	bool scan = data == NULL;
+	int i = 0;
+	int base = 10;
+	bool bytestring = false;
+
+	const char* colon = strchr(str, ':');
+	if (colon != NULL && colon < str + len) { // byte string
+		base = 16;
+		bytestring = true;
+	} else if (len > 2 && strncmp(str, "0x", 2) == 0) { // hexadecimal prefix
+		base = 16;
+		i += 2;
+	}/* else if (len > 1 && strncmp(str, "0", 1) == 0) { // octal prefix
+		base = 8;
+		i += 1;
+	}*/
+
+	const char* set;
+	// if (base == 8) set = SET_OCT;
+	if (base == 10) set = SET_DEC;
+	if (base == 16) {
+		if (bytestring) set = SET_HEX_STR;
+		else set = SET_HEX;
+	}
+	
+	size_t len_ok = strspn(str + i, set) + i;
+	if (len != len_ok) return -len_ok;
+
+	if (scan) {
+		if (base == 10) *size += 1;
+		else if (base == 16) {
+			if (!bytestring) {
+				*size += (len - i + 1) / 2;
+			} else {
+				for (; colon != NULL && colon < str + len; colon = strchr(str, ':')) {
+					*size += 1;
+				}
+			}
+		}
+	}
+
+	if (scan) printf("number (base %d%s) of length %lu\n", base, bytestring ? " as bytestring" : "", len - i);
+	return len;
 }
 
 int strtodata(const char* str, char** data, size_t* size) {
 	const char* ifs = IFS;
 	*size = 0;
 	size_t i;
-	size_t str_len = strlen(str);
+	size_t len = strlen(str);
 
-	// TODO: finish this parser
-	// for (i = 0; i < str_len; i++) {
-	// 	unsigned ifs_run = ifsrun(&str[i], ifs);
-	// 	
-	// }
+	for (i = 0; i < len;) {
+		// skip whitespace
+		int run;
+		run = strspn(&str[i], ifs);
+		if (run > 0) printf("skipping whitespace for %d bytes...\n", run);
+		i += run;
+		// end of string
+		if (str[i] == '\0') break;
 
-	*size = str_len;
+		if ((run = parse_str(str + i, NULL, size)) > 0) { i += run; continue; }
+		if ((run = parse_num(str + i, NULL, size)) > 0) { i += run; continue; }
 
-	*data = (char*) malloc(*size);
+		// no format detected
+		return -i + run;
+	}
+	printf("end of string w/o parse errors\n");
+	printf("buffer size is now %lu\n", *size);
+	exit(0);
 
-	memcpy(*data, str, *size);
+	*data = (char*) malloc(*size);
 
 	return 0;
 }
diff --git a/client/parse.h b/client/parse.h
index ac06446..10274e7 100644
--- a/client/parse.h
+++ b/client/parse.h
@@ -4,6 +4,11 @@
 
 #define IFS " \t\n"
 
+#define SET_OCT "01234567"
+#define SET_DEC "0123456789"
+#define SET_HEX SET_DEC"abcdefABCDEF"
+#define SET_HEX_STR SET_HEX":"
+
 /**
  * \brief modify \p token to point to the first token when broken up on \p ifs
  * and return the remaining data
@@ -26,13 +31,12 @@ char* consume_token(char* token, const char* ifs);
  * \param data  pointer to \c char* that will store the resulting data
  * \param size  size of \p data
  *
- * \return 0 if the string was parsed succesfully, or 1 if the string could not
- * be parsed succesfully
+ * \return 0 or a negative integer representing the index where there is a
+ * syntax error if there was an error, or a positive integer representing the
+ * amount of bytes parsed from \p str
  *
  * \note The pointer that \p data refers to will not be initialized by this
  * function if parsing fails
- * \note \p size will contain the index of \p str where the first invalid data
- * was found if parsing fails
  */
 int strtodata(const char* str, char** data, size_t* size);
 
diff --git a/client/readme.md b/client/readme.md
index 9d755aa..04471d2 100644
--- a/client/readme.md
+++ b/client/readme.md
@@ -12,3 +12,10 @@ goal (in order of implementation):
     ls                  list connected puzzle modules
 ```
 
+
+```
+send 0x39 "Hello world!" de:ad:be:ef 0xff     5        0a       0750
+          ^~~~~~~~~~~~~~ ^~~~~~~~~~~ ~^~~     ~^       ~^       ~~~~^
+          STR_INTP       BYTE_ARR    UNSIGNED UNSIGNED UNSIGNED UNSIGNED
+                                     (hex+0x) (dec)    (hex)    (oct)
+```