pattern: fix captures, add split and global-match-str

core/Pattern.carp

@@ -39,6 +39,30 @@ If you want to replace all occurrences of the pattern, use `-1`.")
   (doc from-chars "creates a pattern that matches a group of characters from a list of those characters.")
   (defn from-chars [chars]
     (Pattern.init &(str* @"[" (String.from-chars chars) @"]")))
+
+  (defn global-match-str [p s]
+    (Array.copy-map &Array.unsafe-first &(global-match p s)))
+
+  (doc split "splits a string by a pattern.")
+  (defn split [p s]
+    (let-do [idx (find-all p s)
+             strs (global-match-str p s)
+             lidx (Array.length &idx)
+             result (Array.allocate (Int.inc lidx))]
+      (Array.aset-uninitialized! &result 0
+        (substring s 0 (if (> lidx 0) @(Array.unsafe-nth &idx 0) (length s))))
+      (for [i 0 (Int.dec (Array.length &idx))]
+        (let [plen (length (Array.unsafe-nth &strs i))]
+          (Array.aset-uninitialized! &result (Int.inc i)
+            (substring s (+ @(Array.unsafe-nth &idx i) plen)
+                         @(Array.unsafe-nth &idx (Int.inc i))))))
+      (when (> lidx 0)
+        (let [plen (length (Array.unsafe-nth &strs (Int.dec lidx)))]
+          (Array.aset-uninitialized! &result lidx
+            (suffix-string s (+ @(Array.unsafe-nth &idx (Int.dec lidx))
+                                plen)))))
+      result))
+
 )
 
 (defmodule String
@@ -84,7 +108,7 @@ If you want to replace all occurrences of the pattern, use `-1`.")
 
   (doc chomp "trims a newline from the end of a string.")
   (defn chomp [s]
-    (Pattern.substitute #"\n$" s "" 1))
+    (Pattern.substitute #"\r$" &(Pattern.substitute #"\n$" s "" 1) "" 1))
 
   (doc collapse-whitespace "collapses groups of whitespace into single spaces.")
   (defn collapse-whitespace [s]

core/carp_pattern.h

@@ -309,11 +309,13 @@ init:                     /* using goto's to optimize tail recursion */
                     case 't': { /* tab? */
                         char h = *(p + 1);
                         p += 2;
-                        if (  (*s == '\r' && h == 'r')
-                           || (*s == '\n' && h == 'n')
-                           || (*s == '\t' && h == 't')
-                           ) { s++; goto init; }
-                        else s = NULL;
+                        if ((*s == '\r' && h == 'r') ||
+                            (*s == '\n' && h == 'n') ||
+                            (*s == '\t' && h == 't')) {
+                            s++;
+                            goto init;
+                        } else
+                            s = NULL;
                         break;
                     }
                     case '0':
@@ -424,7 +426,7 @@ Array Pattern_internal_push_onecapture(PatternMatchState *ms, int i, String s,
     if (i >= ms->level) {
         if (!i)
             return Array_push_String(captures, s, i,
-                                     ms->capture[i].len); /* add whole match */
+                                     e - s); /* add whole match */
         else
             carp_regerror("invalid capture index %cd", C_ESC, i + 1);
     } else {

test/pattern.carp

@@ -61,8 +61,7 @@
                 "matches? works as exptected II")
   (assert-equal test
                 true
-                (and* (matches? #"\n" "\n") (matches? #"\n" "\r")
-                      (matches? #"\n" "\r\n"))
+                (and (matches? #"\n" "\n") (matches? #"\r" "\r"))
                 "matches? works as exptected on newlines special case")
   (assert-equal test
                 true
@@ -84,6 +83,10 @@
                 "sub 2-3 3-4"
                 &(substitute #"(\d)-(\d)" "1-2 2-3 3-4" "sub" 1)
                 "substitute works as expected")
+  (assert-equal test
+                &[@"" @"1" @"2" @"3" @""]
+                &(split #"\-\-" "--1--2--3--")
+                "split works as expected")
   (assert-equal test
                 "sub sub sub"
                 &(substitute #"(\d)-(\d)" "1-2 2-3 3-4" "sub" -1)