feat(regexp): support for more metacharacters

Water-Melon · Mar 1, 2024 · 3eedca3 · 3eedca3
1 parent 1a106d0
commit 3eedca3
Show file tree

Hide file tree

Showing 4 changed files with 76 additions and 27 deletions.
diff --git a/docs/Melon Developer Guide.txt b/docs/Melon Developer Guide.txt
@@ -1594,9 +1594,15 @@ Their definitions can be found in melon/include/mln_types.h.
         Match a nonnumeric character.
      m) \n
         Just like '\n' in C.
-     n) \a
+     n) \w
         Match a letter.
-     o) \\
+     o) \W
+        Match a non-letter character.
+     p) \s
+        Match a whitespace.
+     q) \S
+        Match a non-whitespace.
+     r) \\
         Just like '\\' in C.
 
   21) JSON

diff --git a/include/mln_regexp.h b/include/mln_regexp.h
@@ -16,25 +16,28 @@ typedef mln_array_t mln_reg_match_result_t;
 #define M_REGEXP_MASK_NEW     ((unsigned int)0x02000000)
 #define M_REGEXP_SPECIAL_MASK ((unsigned int)0x0000ffff)
 
-#define M_REGEXP_ALPHA        128
-#define M_REGEXP_NUM          129
-#define M_REGEXP_NOT_NUM      130
-#define M_REGEXP_SEPARATOR    160
-
-#define M_REGEXP_LBRACE       161
-#define M_REGEXP_RBRACE       162
-#define M_REGEXP_LPAR         163
-#define M_REGEXP_RPAR         164
-#define M_REGEXP_LSQUAR       165
-#define M_REGEXP_RSQUAR       166
-#define M_REGEXP_XOR          167
-#define M_REGEXP_STAR         168
-#define M_REGEXP_DOLL         169
-#define M_REGEXP_DOT          170
-#define M_REGEXP_QUES         171
-#define M_REGEXP_PLUS         172
-#define M_REGEXP_SUB          173
-#define M_REGEXP_OR           174
+#define M_REGEXP_ALPHA          128
+#define M_REGEXP_NOT_ALPHA      129
+#define M_REGEXP_NUM            130
+#define M_REGEXP_NOT_NUM        131
+#define M_REGEXP_WHITESPACE     132
+#define M_REGEXP_NOT_WHITESPACE 133
+#define M_REGEXP_SEPARATOR      160
+
+#define M_REGEXP_LBRACE         161
+#define M_REGEXP_RBRACE         162
+#define M_REGEXP_LPAR           163
+#define M_REGEXP_RPAR           164
+#define M_REGEXP_LSQUAR         165
+#define M_REGEXP_RSQUAR         166
+#define M_REGEXP_XOR            167
+#define M_REGEXP_STAR           168
+#define M_REGEXP_DOLL           169
+#define M_REGEXP_DOT            170
+#define M_REGEXP_QUES           171
+#define M_REGEXP_PLUS           172
+#define M_REGEXP_SUB            173
+#define M_REGEXP_OR             174
 
 
 #define mln_reg_match_result_new(prealloc) mln_array_new(NULL, sizeof(mln_string_t), (prealloc))

diff --git a/include/mln_utils.h b/include/mln_utils.h
@@ -150,10 +150,11 @@ extern int socketpair(int domain, int type, int protocol, int sv[2]);
 #define mln_socket_close close
 #endif
 
-#define mln_isalpha(x) (((x) >= 'a' && (x) <= 'z') || ((x) >= 'A' && (x) <= 'Z'))
-#define mln_isdigit(x) ((x) >= '0' && (x) <= '9')
-#define mln_isascii(x) ((x) >= 0 && (x) <= 127)
-#define mln_isprint(x) ((x) >= 32 && (x) <= 126)
+#define mln_isalpha(x)      (((x) >= 'a' && (x) <= 'z') || ((x) >= 'A' && (x) <= 'Z'))
+#define mln_isdigit(x)      ((x) >= '0' && (x) <= '9')
+#define mln_isascii(x)      ((x) >= 0 && (x) <= 127)
+#define mln_isprint(x)      ((x) >= 32 && (x) <= 126)
+#define mln_iswhitespace(x) ((x) == ' ' || (x) == '\t' || (x) == '\n' || (x) == '\f' || (x) == '\r' || (x) == '\v')
 
 #define MLN_AUTHOR "Niklaus F.Schen"
 

diff --git a/src/mln_regexp.c b/src/mln_regexp.c
@@ -228,6 +228,36 @@ MLN_FUNC(static, int, mln_match_here, \
             reglen -= len_0;
             goto again;
         }
+        if (c_0 == M_REGEXP_NOT_ALPHA && textlen > 0) {
+            if (mln_isalpha(*text)) {
+                return -1;
+            }
+            ++text;
+            --textlen;
+            regexp += len_0;
+            reglen -= len_0;
+            goto again;
+        }
+        if (c_0 == M_REGEXP_WHITESPACE && textlen > 0) {
+            if (!mln_iswhitespace(*text)) {
+                return -1;
+            }
+            ++text;
+            --textlen;
+            regexp += len_0;
+            reglen -= len_0;
+            goto again;
+        }
+        if (c_0 == M_REGEXP_NOT_WHITESPACE && textlen > 0) {
+            if (mln_iswhitespace(*text)) {
+                return -1;
+            }
+            ++text;
+            --textlen;
+            regexp += len_0;
+            reglen -= len_0;
+            goto again;
+        }
         if (c_0 == M_REGEXP_DOT && textlen > 0) {
             regexp += len_0;
             ++text;
@@ -773,12 +803,18 @@ MLN_FUNC(static inline, int, mln_get_char, (unsigned int flag, char *s, int len)
                 return '|';
             case '\\':
                 return '\\';
-            case 'a':
+            case 'w':
                 return M_REGEXP_ALPHA;
+            case 'W':
+                return M_REGEXP_NOT_ALPHA;
             case 'd':
                 return M_REGEXP_NUM;
             case 'D':
                 return M_REGEXP_NOT_NUM;
+            case 's':
+                return M_REGEXP_WHITESPACE;
+            case 'S':
+                return M_REGEXP_NOT_WHITESPACE;
             default:
                 return s[1];
         }
@@ -829,9 +865,12 @@ MLN_FUNC(static inline, int, mln_get_length, (char *s, int len), (s, len), {
             case 't':
             case '-':
             case '|':
-            case 'a':
+            case 'w':
+            case 'W':
             case 'd':
             case 'D':
+            case 's':
+            case 'S':
             case '\\':
             default:
                 return 2;