From 5d6fa135b1eadbceac04e5456fabb7e53feedc10 Mon Sep 17 00:00:00 2001
From: Bruce Hill <bruce@bruce-hill.com>
Date: Mon, 2 Sep 2024 23:26:55 -0400
Subject: Add Text:find_all()

---
 builtins/text.c | 20 ++++++++++++++++++++
 builtins/text.h |  6 +-----
 environment.c   |  3 ++-
 test/text.tm    | 18 ++++++++++++++++++
 4 files changed, 41 insertions(+), 6 deletions(-)

diff --git a/builtins/text.c b/builtins/text.c
index 4ec3d835..2f3fbb46 100644
--- a/builtins/text.c
+++ b/builtins/text.c
@@ -1402,6 +1402,26 @@ public Text_t Text$quoted(Text_t text, bool colorize)
 #undef add_escaped
 }
 
+public array_t Text$find_all(Text_t text, Text_t pattern)
+{
+    if (pattern.length == 0) // special case
+        return (array_t){.length=0};
+
+    array_t matches = {};
+
+    Int_t i = I_small(1);
+    for (;;) {
+        int64_t len;
+        Int_t found = Text$find(text, pattern, i, &len);
+        if (I_is_zero(found)) break;
+        Text_t match = Text$slice(text, found, Int$plus(found, Int64_to_Int(len-1)));
+        Array$insert(&matches, &match, I_small(0), sizeof(Text_t));
+        i = Int$plus(found, Int64_to_Int(len));
+    }
+
+    return matches;
+}
+
 public Text_t Text$replace(Text_t text, Text_t pattern, Text_t replacement)
 {
     Text_t ret = {.length=0};
diff --git a/builtins/text.h b/builtins/text.h
index 9f187a37..20ea9a25 100644
--- a/builtins/text.h
+++ b/builtins/text.h
@@ -12,11 +12,6 @@
 #include "types.h"
 #include "where.h"
 
-typedef struct {
-    enum { FIND_FAILURE, FIND_SUCCESS } status;
-    int32_t index;
-} find_result_t;
-
 int printf_text(FILE *stream, const struct printf_info *info, const void *const args[]);
 int printf_text_size(const struct printf_info *info, size_t n, int argtypes[n], int sizes[n]);
 
@@ -38,6 +33,7 @@ Text_t Text$quoted(Text_t str, bool colorize);
 Text_t Text$replace(Text_t str, Text_t pat, Text_t replacement);
 array_t Text$split(Text_t text, Text_t pattern);
 Int_t Text$find(Text_t text, Text_t pattern, Int_t i, int64_t *match_length);
+array_t Text$find_all(Text_t text, Text_t pattern);
 bool Text$has(Text_t text, Text_t pattern);
 const char *Text$as_c_string(Text_t text);
 public Text_t Text$format(const char *fmt, ...);
diff --git a/environment.c b/environment.c
index 09935f11..d4ed6c8d 100644
--- a/environment.c
+++ b/environment.c
@@ -247,7 +247,8 @@ env_t *new_compilation_unit(CORD *libname)
             {"by", "Range$by", "func(range:Range, step:Int)->Range"},
         )},
         {"Text", TEXT_TYPE, "Text_t", "$Text", TypedArray(ns_entry_t,
-            // {"find", "Text$find", "func(text:Text, pattern:Text)->FindResult"},
+            {"find", "Text$find", "func(text:Text, pattern:Text)->Int"},
+            {"find_all", "Text$find_all", "func(text:Text, pattern:Text)->[Text]"},
             {"as_c_string", "CORD_to_char_star", "func(text:Text)->CString"},
             {"codepoint_names", "Text$codepoint_names", "func(text:Text)->[Text]"},
             {"from_bytes", "Text$from_bytes", "func(bytes:[Int8])->Text"},
diff --git a/test/text.tm b/test/text.tm
index cf084f8a..0dd5f2ec 100644
--- a/test/text.tm
+++ b/test/text.tm
@@ -155,3 +155,21 @@ func main():
 
 	>> "":split()
 	= []
+
+	>> " one  two three   ":find_all("[..alpha]")
+	= ["one", "two", "three"]
+
+	>> " one  two three   ":find_all("[..!space]")
+	= ["one", "two", "three"]
+
+	>> "    ":find_all("[..alpha]")
+	= []
+
+	>> " foo(baz(), 1)  doop() ":find_all("[..id](?)")
+	= ["foo(baz(), 1)", "doop()"]
+
+	>> "":find_all("")
+	= []
+
+	>> "Hello":find_all("")
+	= []
-- 
cgit v1.2.3