From 08b4a34a8ece5917baf25d2b594f03d43aa1c437 Mon Sep 17 00:00:00 2001
From: Nicolas Williams <nico@cryptonector.com>
Date: Thu, 28 Nov 2013 14:48:46 -0600
Subject: [PATCH] Add jv string utility functions
MIME-Version: 1.0
Content-Type: text/plain; charset=utf8
Content-Transfer-Encoding: 8bit

    jv_string_empty()
        -> return an empty string with given allocated length (for fast
           appends)
    jv_string_append_codepoint
        -> append a single codepoint (int) to the given string
    jv_string_explode
        -> return an array of codepoints making up a string
    jv_string_implode
        -> return the UTF-8 encoding of an array of codepoint numbers
---
 jv.c | 53 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 jv.h |  4 ++++
 2 files changed, 57 insertions(+)

diff --git a/jv.c b/jv.c
index 7182445..92c4860 100644
--- a/jv.c
+++ b/jv.c
@@ -417,6 +417,14 @@ static jv_nontrivial jvp_string_new(const char* data, uint32_t length) {
   return r;
 }
 
+static jv_nontrivial jvp_string_empty_new(uint32_t length) {
+  jvp_string* s = jvp_string_alloc(length);
+  s->length_hashed = 0;
+  memset(s->data, 0, length);
+  jv_nontrivial r = {&s->refcnt, {0,0}};
+  return r;
+}
+
 static void jvp_string_free(jv_nontrivial* s) {
   if (jvp_refcnt_dec(s)) {
     jvp_string* str = jvp_string_ptr(s);
@@ -560,6 +568,13 @@ jv jv_string_sized(const char* str, int len) {
   return j;
 }
 
+jv jv_string_empty(int len) {
+  jv j;
+  j.kind = JV_KIND_STRING;
+  j.val.nontrivial = jvp_string_empty_new(len);
+  return j;
+}
+
 jv jv_string(const char* str) {
   return jv_string_sized(str, strlen(str));
 }
@@ -581,6 +596,37 @@ int jv_string_length_codepoints(jv j) {
   return len;
 }
 
+jv jv_string_explode(jv j) {
+  assert(jv_get_kind(j) == JV_KIND_STRING);
+  const char* i = jv_string_value(j);
+  int len = jv_string_length_bytes(jv_copy(j));
+  const char* end = i + len;
+  jv a = jv_array_sized(len);
+  int c;
+  while ((i = jvp_utf8_next(i, end, &c)))
+    a = jv_array_append(a, jv_number(c));
+  jv_free(j);
+  return a;
+}
+
+jv jv_string_implode(jv j) {
+  assert(jv_get_kind(j) == JV_KIND_ARRAY);
+  int len = jv_array_length(jv_copy(j));
+  jv s = jv_string_empty(len);
+  int i;
+
+  assert(len >= 0);
+
+  for (i = 0; i < len; i++) {
+    jv n = jv_array_get(j, i);
+    assert(jv_get_kind(n) == JV_KIND_NUMBER);
+    s = jv_string_append_codepoint(s, jv_number_value(n));
+  }
+
+  jv_free(j);
+  return s;
+}
+
 unsigned long jv_string_hash(jv j) {
   assert(jv_get_kind(j) == JV_KIND_STRING);
   uint32_t hash = jvp_string_hash(jvp_string_ptr(&j.val.nontrivial));
@@ -612,6 +658,13 @@ jv jv_string_append_buf(jv a, const char* buf, int len) {
   return a;
 }
 
+jv jv_string_append_codepoint(jv a, int c) {
+  char buf[5];
+  int len = jvp_utf8_encode(c, buf);
+  jvp_string_append(&a.val.nontrivial, buf, len);
+  return a;
+}
+
 jv jv_string_append_str(jv a, const char* str) {
   return jv_string_append_buf(a, str, strlen(str));
 }
diff --git a/jv.h b/jv.h
index 0f50359..50efbe6 100644
--- a/jv.h
+++ b/jv.h
@@ -75,14 +75,18 @@ jv jv_array_slice(jv, int, int);
 
 jv jv_string(const char*);
 jv jv_string_sized(const char*, int);
+jv jv_string_empty(int len);
 int jv_string_length_bytes(jv);
 int jv_string_length_codepoints(jv);
 unsigned long jv_string_hash(jv);
 const char* jv_string_value(jv);
 jv jv_string_concat(jv, jv);
 jv jv_string_fmt(const char*, ...);
+jv jv_string_append_codepoint(jv a, int c);
 jv jv_string_append_buf(jv a, const char* buf, int len);
 jv jv_string_append_str(jv a, const char* str);
+jv jv_string_explode(jv j);
+jv jv_string_implode(jv j);
 
 jv jv_object();
 jv jv_object_get(jv object, jv key);
-- 
2.40.0