Various fixes

README.md

@@ -7,6 +7,7 @@ Using _mpc_ might be of interest to you if you are...
 
 * Building a new programming language
 * Building a new data format
+* Parsing an existing programming languages
 * Parsing an existing data format
 * Embedding a Domain Specific Language
 * Implementing [Greenspun's Tenth Rule](http://en.wikipedia.org/wiki/Greenspun%27s_tenth_rule)
@@ -15,12 +16,12 @@ Using _mpc_ might be of interest to you if you are...
 Features
 --------
 
-* Type-Generic Parser Combinators
+* Type-Generic
-* Predictive Recursive Descent Parsers
+* Predictive, Recursive Descent
-* Error Message Support
-* Regular Expression Support
-* Grammar Support
 * Easy to Integrate (One Source File in ANSI C)
+* Error Messages
+* Regular Expression Parser Generator
+* Grammar Parser Generator
 
 
 Alternatives
@@ -28,7 +29,7 @@ Alternatives
 
 The current main alternative C based parser combinator is a branch of [Cesium3](https://github.com/wbhart/Cesium3/tree/combinators).
 
-This project has several downsides which _mpc_ overcomes:
+_mpc_ provides a number of features that this project does not offer, but it also overcomes a number of potential downsides:
 
 * _mpc_ Works for Generic Types
 * _mpc_ Doesn't rely on Boehm-Demers-Weiser Garbage Collection
@@ -73,7 +74,7 @@ mpc_ast_t* parse_maths(const char* input) {
 }
 ```
 
-If you were to input something like `"(4 * 2 * 11 + 2) - 5"` into this function the `mpc_ast_t` you get out would look something like this:
+If you were to input `"(4 * 2 * 11 + 2) - 5"` into this function, the `mpc_ast_t` output would look something like this:
 
 ```python
 >:

TODO.md

@@ -1,3 +1,8 @@
+- Escape/Crop String Literals
+- Escale/Crop Char Literals
+- Escape/Crop Regex Literals
+- Combinator that scans input then returns cursor.
+
 - Test All Regex Features
 - Test Regex Range Feature
 - Add proper tests for everything in general

mpc.c

@@ -269,10 +269,6 @@ static mpc_err_t* mpc_err_count(mpc_err_t* x, int n) {
   return y;
 }
 
-char* mpc_err_filename(mpc_err_t* x) {
-  return x->filename;
-}
-
 void mpc_err_expected(mpc_err_t* x, char** out, int* out_num, int out_max) {
   
   int i;
@@ -284,17 +280,10 @@ void mpc_err_expected(mpc_err_t* x, char** out, int* out_num, int out_max) {
   }
 }
 
-int mpc_err_line(mpc_err_t* x) {
+char* mpc_err_filename(mpc_err_t* x) { return x->filename; }
-  return x->state.row;
+int mpc_err_line(mpc_err_t* x) { return x->state.row; }
-}
+int mpc_err_column(mpc_err_t* x) { return x->state.col; }
-
+char mpc_err_unexpected(mpc_err_t* x) { return x->state.next; }
-int mpc_err_column(mpc_err_t* x) {
-  return x->state.col;
-}
-
-char mpc_err_unexpected(mpc_err_t* x) {
-  return x->state.next;
-}
 
 /*
 ** Input Type
@@ -878,20 +867,21 @@ static mpc_err_t* mpc_stack_merger_err(mpc_stack_t* s, int n) {
 
 /*
 ** This is rather pleasant. The core parsing routine
-** is written in about 300 lines of C.
+** is written in about 200 lines of C.
 **
 ** I also love the way in which each parsing type
 ** concisely matches some construct or pattern.
 **
-** Particularly nice are the `either` and `also`
+** Particularly nice are the `or` and `and`
 ** types which have a broken but mirrored structure
 ** with return value and error reflected.
 **
 ** When this function was written in recursive form
 ** it looked pretty nice. But I've since switched
 ** it around to an akward while loop. It was an
-** unfortunate change but if was a noble attempt
+** unfortunate change for code simplicity but it
-** in the name of performance (and not smashing the stack).
+** is noble in the name of performance (and 
+** not smashing the stack).
 **
 ** But it is now a pretty ugly beast...
 */
@@ -910,7 +900,6 @@ int mpc_parse_input(mpc_input_t* i, mpc_parser_t* init, mpc_result_t* final) {
   
   /* Variables */
   char* s;
-  mpc_val_t* t;
   mpc_result_t r;
 
   /* Go! */
@@ -1178,9 +1167,7 @@ static void mpc_undefine_unretained(mpc_parser_t* p, int force) {
   
   switch (p->type) {
     
-    case MPC_TYPE_FAIL:
+    case MPC_TYPE_FAIL: free(p->data.fail.m); break;
-      free(p->data.fail.m);
-      break;
     
     case MPC_TYPE_ONEOF: 
     case MPC_TYPE_NONEOF:
@@ -1188,17 +1175,9 @@ static void mpc_undefine_unretained(mpc_parser_t* p, int force) {
       free(p->data.string.x); 
       break;
     
-    case MPC_TYPE_APPLY:
+    case MPC_TYPE_APPLY:    mpc_undefine_unretained(p->data.apply.x, 0);    break;
-      mpc_undefine_unretained(p->data.apply.x, 0);
+    case MPC_TYPE_APPLY_TO: mpc_undefine_unretained(p->data.apply_to.x, 0); break;
-      break;
+    case MPC_TYPE_PREDICT:  mpc_undefine_unretained(p->data.predict.x, 0);  break;
-    
-    case MPC_TYPE_APPLY_TO:
-      mpc_undefine_unretained(p->data.apply_to.x, 0);
-      break;
-    
-    case MPC_TYPE_PREDICT:
-      mpc_undefine_unretained(p->data.predict.x, 0);
-      break;
     
     case MPC_TYPE_MAYBE:
     case MPC_TYPE_NOT:
@@ -1216,13 +1195,8 @@ static void mpc_undefine_unretained(mpc_parser_t* p, int force) {
       mpc_undefine_unretained(p->data.repeat.x, 0);
       break;
     
-    case MPC_TYPE_OR:
+    case MPC_TYPE_OR:  mpc_undefine_or(p);  break;
-      mpc_undefine_or(p);
+    case MPC_TYPE_AND: mpc_undefine_and(p); break;
-      break;
-    
-    case MPC_TYPE_AND:
-      mpc_undefine_and(p);
-      break;
     
     default: break;
   }
@@ -1889,6 +1863,10 @@ static mpc_val_t* mpcf_re_range(mpc_val_t* x) {
       if (tmp != NULL) {
         range = realloc(range, strlen(range) + strlen(tmp) + 1);
         strcat(range, tmp);
+      } else {
+        range = realloc(range, strlen(range) + 1 + 1);
+        range[strlen(range) + 1] = '\0';
+        range[strlen(range) + 0] = s[i+1];      
       }
       i++;
     }

mpc.h

@@ -15,6 +15,7 @@
 #include <stdarg.h>
 #include <string.h>
 #include <math.h>
+#include <errno.h>
 
 /*
 ** Error Type