eval: Move some decoding functions to eval/decode.c

5 files changed, 798 insertions, 811 deletions

diff --git a/src/nvim/eval.c b/src/nvim/eval.c
index 7edc1402f5..38dc7c1034 100644
--- a/src/nvim/eval.c
+++ b/src/nvim/eval.c
@@ -74,6 +74,7 @@
 #include "nvim/version.h"
 #include "nvim/window.h"
 #include "nvim/eval/encode.h"
+#include "nvim/eval/decode.h"
 #include "nvim/os/os.h"
 #include "nvim/event/libuv_process.h"
 #include "nvim/event/pty_process.h"
@@ -89,7 +90,6 @@
 #include "nvim/os/input.h"
 #include "nvim/event/loop.h"
 #include "nvim/lib/queue.h"
-#include "nvim/lib/kvec.h"
 
 #define DICT_MAXNEST 100        /* maximum nesting of lists and dicts */
 
@@ -416,16 +416,6 @@ typedef struct {
   int status;
 } JobEvent;
 
-/// Helper structure for container_struct
-typedef struct {
-  size_t stack_index;  ///< Index of current container in stack.
-  typval_T container;  ///< Container. Either VAR_LIST, VAR_DICT or VAR_LIST
-                       ///< which is _VAL from special dictionary.
-} ContainerStackItem;
-
-typedef kvec_t(typval_T) ValuesStack;
-typedef kvec_t(ContainerStackItem) ContainerStack;
-
 #ifdef INCLUDE_GENERATED_DECLARATIONS
 # include "eval.c.generated.h"
 #endif
@@ -4142,7 +4132,7 @@ static int eval7(
     if (get_float) {
       float_T f;
 
-      *arg += string2float(*arg, &f);
+      *arg += string2float((char *) *arg, &f);
       if (evaluate) {
         rettv->v_type = VAR_FLOAT;
         rettv->vval.v_float = f;
@@ -6550,24 +6540,22 @@ failret:
   return OK;
 }
 
-/*
- * Convert the string "text" to a floating point number.
- * This uses strtod().  setlocale(LC_NUMERIC, "C") has been used to make sure
- * this always uses a decimal point.
- * Returns the length of the text that was consumed.
- */
-static int 
-string2float (
-    char_u *text,
-    float_T *value         /* result stored here */
-)
+/// Convert the string to a floating point number
+///
+/// This uses strtod().  setlocale(LC_NUMERIC, "C") has been used earlier to
+/// make sure this always uses a decimal point.
+///
+/// @param[in]  text  String to convert.
+/// @param[out]  ret_value  Location where conversion result is saved.
+///
+/// @return Length of the text that was consumed.
+size_t string2float(const char *const text, float_T *const ret_value)
+  FUNC_ATTR_NONNULL_ALL
 {
-  char        *s = (char *)text;
-  float_T f;
+  char *s = NULL;
 
-  f = strtod(s, &s);
-  *value = f;
-  return (int)((char_u *)s - text);
+  *ret_value = strtod(text, &s);
+  return (size_t) (s - text);
 }
 
 /// Get the value of an environment variable.
@@ -11571,498 +11559,6 @@ static void f_join(typval_T *argvars, typval_T *rettv)
     rettv->vval.v_string = NULL;
 }
 
-/// Helper function used for working with stack vectors used by JSON decoder
-///
-/// @param[in]  obj  New object.
-/// @param[out]  stack  Object stack.
-/// @param[out]  container_stack  Container objects stack.
-/// @param[in]  p  Position in string which is currently being parsed.
-///
-/// @return OK in case of success, FAIL in case of error.
-static inline int json_decoder_pop(typval_T obj, ValuesStack *const stack,
-                                   ContainerStack *const container_stack,
-                                   const char *const p)
-  FUNC_ATTR_NONNULL_ALL
-{
-  if (kv_size(*container_stack) == 0) {
-    kv_push(typval_T, *stack, obj);
-    return OK;
-  }
-  ContainerStackItem last_container = kv_last(*container_stack);
-  if (obj.v_type == last_container.container.v_type
-      // vval.v_list and vval.v_dict should have the same size and offset
-      && ((void *) obj.vval.v_list
-          == (void *) last_container.container.vval.v_list)) {
-    kv_pop(*container_stack);
-    last_container = kv_last(*container_stack);
-  }
-  if (last_container.container.v_type == VAR_LIST) {
-    listitem_T *obj_li = listitem_alloc();
-    obj_li->li_tv = obj;
-    list_append(last_container.container.vval.v_list, obj_li);
-  } else if (last_container.stack_index == kv_size(*stack) - 2) {
-    typval_T key = kv_pop(*stack);
-    if (key.v_type != VAR_STRING) {
-      assert(false);
-    } else if (key.vval.v_string == NULL || *key.vval.v_string == NUL) {
-      // TODO: fall back to special dict in case of empty key
-      EMSG(_("E474: Empty key"));
-      clear_tv(&obj);
-      return FAIL;
-    }
-    dictitem_T *obj_di = dictitem_alloc(key.vval.v_string);
-    clear_tv(&key);
-    if (dict_add(last_container.container.vval.v_dict, obj_di)
-        == FAIL) {
-      // TODO: fall back to special dict in case of duplicate keys
-      EMSG(_("E474: Duplicate key"));
-      dictitem_free(obj_di);
-      clear_tv(&obj);
-      return FAIL;
-    }
-    obj_di->di_tv = obj;
-  } else {
-    // Object with key only
-    if (obj.v_type != VAR_STRING) {
-      EMSG2(_("E474: Expected string key: %s"), p);
-      clear_tv(&obj);
-      return FAIL;
-    }
-    kv_push(typval_T, *stack, obj);
-  }
-  return OK;
-}
-
-/// Convert JSON string into VimL object
-///
-/// @param[in]  buf  String to convert. UTF-8 encoding is assumed.
-/// @param[in]  len  Length of the string.
-/// @param[out]  rettv  Location where to save results.
-///
-/// @return OK in case of success, FAIL otherwise.
-static int json_decode_string(const char *const buf, const size_t len,
-                              typval_T *rettv)
-  FUNC_ATTR_NONNULL_ALL FUNC_ATTR_WARN_UNUSED_RESULT
-{
-  vimconv_T conv;
-  convert_setup(&conv, (char_u *) "utf-8", p_enc);
-  conv.vc_fail = true;
-  int ret = OK;
-  ValuesStack stack;
-  kv_init(stack);
-  ContainerStack container_stack;
-  kv_init(container_stack);
-  rettv->v_type = VAR_UNKNOWN;
-  const char *const e = buf + len;
-  bool didcomma = false;
-  bool didcolon = false;
-#define POP(obj) \
-  do { \
-    if (json_decoder_pop(obj, &stack, &container_stack, p) == FAIL) { \
-      goto json_decode_string_fail; \
-    } \
-  } while (0)
-  const char *p = buf;
-  for (; p < e; p++) {
-    switch (*p) {
-      case '}':
-      case ']': {
-        if (kv_size(container_stack) == 0) {
-          EMSG2(_("E474: No container to close: %s"), p);
-          goto json_decode_string_fail;
-        }
-        ContainerStackItem last_container = kv_last(container_stack);
-        if (*p == '}' && last_container.container.v_type != VAR_DICT) {
-          EMSG2(_("E474: Closing list with figure brace: %s"), p);
-          goto json_decode_string_fail;
-        } else if (*p == ']' && last_container.container.v_type != VAR_LIST) {
-          EMSG2(_("E474: Closing dictionary with bracket: %s"), p);
-          goto json_decode_string_fail;
-        } else if (didcomma) {
-          EMSG2(_("E474: Trailing comma: %s"), p);
-          goto json_decode_string_fail;
-        } else if (didcolon) {
-          EMSG2(_("E474: Expected value after colon: %s"), p);
-          goto json_decode_string_fail;
-        } else if (last_container.stack_index != kv_size(stack) - 1) {
-          assert(last_container.stack_index < kv_size(stack) - 1);
-          EMSG2(_("E474: Expected value: %s"), p);
-          goto json_decode_string_fail;
-        }
-        if (kv_size(stack) == 1) {
-          p++;
-          kv_pop(container_stack);
-          goto json_decode_string_after_cycle;
-        } else {
-          typval_T obj = kv_pop(stack);
-          POP(obj);
-          break;
-        }
-      }
-      case ',': {
-        if (kv_size(container_stack) == 0) {
-          EMSG2(_("E474: Comma not inside container: %s"), p);
-          goto json_decode_string_fail;
-        }
-        ContainerStackItem last_container = kv_last(container_stack);
-        if (didcomma) {
-          EMSG2(_("E474: Duplicate comma: %s"), p);
-          goto json_decode_string_fail;
-        } else if (didcolon) {
-          EMSG2(_("E474: Comma after colon: %s"), p);
-          goto json_decode_string_fail;
-        } if (last_container.container.v_type == VAR_DICT
-            && last_container.stack_index != kv_size(stack) - 1) {
-          EMSG2(_("E474: Using comma in place of colon: %s"), p);
-          goto json_decode_string_fail;
-        } else if ((last_container.container.v_type == VAR_DICT
-                    && (last_container.container.vval.v_dict->dv_hashtab.ht_used
-                        == 0))
-                   || (last_container.container.v_type == VAR_LIST
-                       && last_container.container.vval.v_list->lv_len == 0)) {
-          EMSG2(_("E474: Leading comma: %s"), p);
-          goto json_decode_string_fail;
-        }
-        didcomma = true;
-        continue;
-      }
-      case ':': {
-        if (kv_size(container_stack) == 0) {
-          EMSG2(_("E474: Colon not inside container: %s"), p);
-          goto json_decode_string_fail;
-        }
-        ContainerStackItem last_container = kv_last(container_stack);
-        if (last_container.container.v_type != VAR_DICT) {
-          EMSG2(_("E474: Using colon not in dictionary: %s"), p);
-          goto json_decode_string_fail;
-        } else if (last_container.stack_index != kv_size(stack) - 2) {
-          EMSG2(_("E474: Unexpected colon: %s"), p);
-          goto json_decode_string_fail;
-        } else if (didcomma) {
-          EMSG2(_("E474: Colon after comma: %s"), p);
-          goto json_decode_string_fail;
-        } else if (didcolon) {
-          EMSG2(_("E474: Duplicate colon: %s"), p);
-          goto json_decode_string_fail;
-        }
-        didcolon = true;
-        continue;
-      }
-      case ' ':
-      case TAB:
-      case NL: {
-        continue;
-      }
-      case 'n': {
-        if (strncmp(p + 1, "ull", 3) != 0) {
-          EMSG2(_("E474: Expected null: %s"), p);
-          goto json_decode_string_fail;
-        }
-        p += 3;
-        POP(vimvars[VV_NULL].vv_di.di_tv);
-        break;
-      }
-      case 't': {
-        if (strncmp(p + 1, "rue", 3) != 0) {
-          EMSG2(_("E474: Expected true: %s"), p);
-          goto json_decode_string_fail;
-        }
-        p += 3;
-        POP(vimvars[VV_TRUE].vv_di.di_tv);
-        break;
-      }
-      case 'f': {
-        if (strncmp(p + 1, "alse", 4) != 0) {
-          EMSG2(_("E474: Expected false: %s"), p);
-          goto json_decode_string_fail;
-        }
-        p += 4;
-        POP(vimvars[VV_FALSE].vv_di.di_tv);
-        break;
-      }
-      case '"': {
-        size_t len = 0;
-        const char *s;
-        for (s = ++p; p < e && *p != '"'; p++) {
-          if (*p == '\\') {
-            p++;
-            if (p == e) {
-              EMSG2(_("E474: Unfinished escape sequence: %s"), buf);
-              goto json_decode_string_fail;
-            }
-            switch (*p) {
-              case 'u': {
-                if (p + 4 >= e) {
-                  EMSG2(_("E474: Unfinished unicode escape sequence: %s"), buf);
-                  goto json_decode_string_fail;
-                } else if (!ascii_isxdigit(p[1])
-                           || !ascii_isxdigit(p[2])
-                           || !ascii_isxdigit(p[3])
-                           || !ascii_isxdigit(p[4])) {
-                  EMSG2(_("E474: Expected four hex digits after \\u: %s"),
-                        p - 1);
-                  goto json_decode_string_fail;
-                }
-                // One UTF-8 character below U+10000 can take up to 3 bytes
-                len += 3;
-                p += 4;
-                break;
-              }
-              case '\\':
-              case '/':
-              case '"':
-              case 't':
-              case 'b':
-              case 'n':
-              case 'r':
-              case 'f': {
-                len++;
-                break;
-              }
-              default: {
-                EMSG2(_("E474: Unknown escape sequence: %s"), p - 1);
-                goto json_decode_string_fail;
-              }
-            }
-          } else {
-            len++;
-          }
-        }
-        if (*p != '"') {
-          EMSG2(_("E474: Expected string end: %s"), buf);
-          goto json_decode_string_fail;
-        }
-        char *str = xmalloc(len + 1);
-        uint16_t fst_in_pair = 0;
-        char *str_end = str;
-        for (const char *t = s; t < p; t++) {
-          if (t[0] != '\\' || t[1] != 'u') {
-            if (fst_in_pair != 0) {
-              str_end += utf_char2bytes((int) fst_in_pair, (char_u *) str_end);
-              fst_in_pair = 0;
-            }
-          }
-          if (*t == '\\') {
-            t++;
-            switch (*t) {
-              case 'u': {
-                char ubuf[] = { t[1], t[2], t[3], t[4], 0 };
-                t += 4;
-                unsigned long ch;
-                vim_str2nr((char_u *) ubuf, NULL, NULL, 0, 0, 2, NULL, &ch);
-                if (0xD800UL <= ch && ch <= 0xDB7FUL) {
-                  fst_in_pair = (uint16_t) ch;
-                } else if (0xDC00ULL <= ch && ch <= 0xDB7FUL) {
-                  if (fst_in_pair != 0) {
-                    int full_char = (
-                        (int) (ch - 0xDC00UL)
-                        + (((int) (fst_in_pair - 0xD800)) << 10)
-                    );
-                    str_end += utf_char2bytes(full_char, (char_u *) str_end);
-                  }
-                } else {
-                  str_end += utf_char2bytes((int) ch, (char_u *) str_end);
-                }
-                break;
-              }
-              case '\\':
-              case '/':
-              case '"':
-              case 't':
-              case 'b':
-              case 'n':
-              case 'r':
-              case 'f': {
-                static const char escapes[] = {
-                  ['\\'] = '\\',
-                  ['/'] = '/',
-                  ['"'] = '"',
-                  ['t'] = TAB,
-                  ['b'] = BS,
-                  ['n'] = NL,
-                  ['r'] = CAR,
-                  ['f'] = FF,
-                };
-                *str_end++ = escapes[(int) *t];
-                break;
-              }
-              default: {
-                assert(false);
-              }
-            }
-          } else {
-            *str_end++ = *t;
-          }
-        }
-        if (fst_in_pair != 0) {
-          str_end += utf_char2bytes((int) fst_in_pair, (char_u *) str_end);
-        }
-        if (conv.vc_type != CONV_NONE) {
-          size_t len = (str_end - str);
-          char *const new_str = (char *) string_convert(&conv, (char_u *) str,
-                                                        &len);
-          if (new_str == NULL) {
-            EMSG2(_("E474: Failed to convert string \"%s\" from UTF-8"), str);
-            xfree(str);
-            goto json_decode_string_fail;
-          }
-          xfree(str);
-          str = new_str;
-          str_end = new_str + len;
-        }
-        *str_end = NUL;
-        // TODO: return special string in case of NUL bytes
-        POP(((typval_T) {
-          .v_type = VAR_STRING,
-          .vval = { .v_string = (char_u *) str, },
-        }));
-        break;
-      }
-      case '-':
-      case '0':
-      case '1':
-      case '2':
-      case '3':
-      case '4':
-      case '5':
-      case '6':
-      case '7':
-      case '8':
-      case '9': {
-        // a.bE[+-]exp
-        const char *const s = p;
-        const char *ints = NULL;
-        const char *fracs = NULL;
-        const char *exps = NULL;
-        if (*p == '-') {
-          p++;
-        }
-        ints = p;
-        while (p < e && ascii_isdigit(*p)) {
-          p++;
-        }
-        if (p < e && *p == '.') {
-          p++;
-          fracs = p;
-          while (p < e && ascii_isdigit(*p)) {
-            p++;
-          }
-          if (p < e && (*p == 'e' || *p == 'E')) {
-            p++;
-            if (p < e && (*p == '-' || *p == '+')) {
-              p++;
-            }
-            exps = p;
-            while (p < e && ascii_isdigit(*p)) {
-              p++;
-            }
-          }
-        }
-        if (p == ints) {
-          EMSG2(_("E474: Missing number after minus sign: %s"), s);
-          goto json_decode_string_fail;
-        } else if (p == fracs) {
-          EMSG2(_("E474: Missing number after decimal dot: %s"), s);
-          goto json_decode_string_fail;
-        } else if (p == exps) {
-          EMSG2(_("E474: Missing exponent: %s"), s);
-          goto json_decode_string_fail;
-        }
-        typval_T tv = {
-          .v_type = VAR_NUMBER,
-          .v_lock = VAR_UNLOCKED,
-        };
-        if (fracs) {
-          // Convert floating-point number
-          (void) string2float((char_u *) s, &tv.vval.v_float);
-          tv.v_type = VAR_FLOAT;
-        } else {
-          // Convert integer
-          long nr;
-          vim_str2nr((char_u *) s, NULL, NULL, 0, 0, 0, &nr, NULL);
-          tv.vval.v_number = (varnumber_T) nr;
-        }
-        POP(tv);
-        p--;
-        break;
-      }
-      case '[': {
-        list_T *list = list_alloc();
-        list->lv_refcount++;
-        typval_T tv = {
-          .v_type = VAR_LIST,
-          .v_lock = VAR_UNLOCKED,
-          .vval = { .v_list = list },
-        };
-        kv_push(ContainerStackItem, container_stack, ((ContainerStackItem) {
-          .stack_index = kv_size(stack),
-          .container = tv,
-        }));
-        kv_push(typval_T, stack, tv);
-        break;
-      }
-      case '{': {
-        dict_T *dict = dict_alloc();
-        dict->dv_refcount++;
-        typval_T tv = {
-          .v_type = VAR_DICT,
-          .v_lock = VAR_UNLOCKED,
-          .vval = { .v_dict = dict },
-        };
-        kv_push(ContainerStackItem, container_stack, ((ContainerStackItem) {
-          .stack_index = kv_size(stack),
-          .container = tv,
-        }));
-        kv_push(typval_T, stack, tv);
-        break;
-      }
-      default: {
-        EMSG2(_("E474: Unidentified byte: %s"), p);
-        goto json_decode_string_fail;
-      }
-    }
-    didcomma = false;
-    didcolon = false;
-    if (kv_size(container_stack) == 0) {
-      p++;
-      break;
-    }
-  }
-#undef POP
-json_decode_string_after_cycle:
-  for (; p < e; p++) {
-    switch (*p) {
-      case NL:
-      case ' ':
-      case TAB: {
-        break;
-      }
-      default: {
-        EMSG2(_("E474: Trailing characters: %s"), p);
-        goto json_decode_string_fail;
-      }
-    }
-  }
-  if (kv_size(stack) > 1 || kv_size(container_stack)) {
-    EMSG2(_("E474: Unexpected end of input: %s"), buf);
-    goto json_decode_string_fail;
-  }
-  goto json_decode_string_ret;
-json_decode_string_fail:
-  ret = FAIL;
-  while (kv_size(stack)) {
-    clear_tv(&kv_pop(stack));
-  }
-json_decode_string_ret:
-  if (ret != FAIL) {
-    assert(kv_size(stack) == 1);
-    *rettv = kv_pop(stack);
-  }
-  kv_destroy(stack);
-  kv_destroy(container_stack);
-  return ret;
-}
-
 /// jsondecode() function
 static void f_jsondecode(typval_T *argvars, typval_T *rettv)
 {
@@ -12865,58 +12361,6 @@ static void f_mode(typval_T *argvars, typval_T *rettv)
   rettv->v_type = VAR_STRING;
 }
 
-/// Msgpack callback for writing to readfile()-style list
-static int msgpack_list_write(void *data, const char *buf, size_t len)
-{
-  if (len == 0) {
-    return 0;
-  }
-  list_T *const list = (list_T *) data;
-  const char *const end = buf + len;
-  const char *line_end = buf;
-  if (list->lv_last == NULL) {
-    list_append_string(list, NULL, 0);
-  }
-  listitem_T *li = list->lv_last;
-  do {
-    const char *line_start = line_end;
-    line_end = xmemscan(line_start, NL, end - line_start);
-    if (line_end == line_start) {
-      list_append_allocated_string(list, NULL);
-    } else {
-      const size_t line_length = line_end - line_start;
-      char *str;
-      if (li == NULL) {
-        str = xmemdupz(line_start, line_length);
-      } else {
-        const size_t li_len = (li->li_tv.vval.v_string == NULL
-                               ? 0
-                               : STRLEN(li->li_tv.vval.v_string));
-        li->li_tv.vval.v_string = xrealloc(li->li_tv.vval.v_string,
-                                           li_len + line_length + 1);
-        str = (char *) li->li_tv.vval.v_string + li_len;
-        memmove(str, line_start, line_length);
-        str[line_length] = 0;
-      }
-      for (size_t i = 0; i < line_length; i++) {
-        if (str[i] == NUL) {
-          str[i] = NL;
-        }
-      }
-      if (li == NULL) {
-        list_append_allocated_string(list, str);
-      } else {
-        li = NULL;
-      }
-      if (line_end == end - 1) {
-        list_append_allocated_string(list, NULL);
-      }
-    }
-    line_end++;
-  } while (line_end < end);
-  return 0;
-}
-
 /// "msgpackdump()" function
 static void f_msgpackdump(typval_T *argvars, typval_T *rettv)
   FUNC_ATTR_NONNULL_ALL
@@ -12930,7 +12374,7 @@ static void f_msgpackdump(typval_T *argvars, typval_T *rettv)
   if (list == NULL) {
     return;
   }
-  msgpack_packer *lpacker = msgpack_packer_new(ret_list, &msgpack_list_write);
+  msgpack_packer *lpacker = msgpack_packer_new(ret_list, &encode_list_write);
   const char *const msg = _("msgpackdump() argument, index %i");
   // Assume that translation will not take more then 4 times more space
   char msgbuf[sizeof("msgpackdump() argument, index ") * 4 + NUMBUFLEN];
@@ -12945,241 +12389,6 @@ static void f_msgpackdump(typval_T *argvars, typval_T *rettv)
   msgpack_packer_free(lpacker);
 }
 
-/// Convert msgpack object to a VimL one
-int msgpack_to_vim(const msgpack_object mobj, typval_T *const rettv)
-  FUNC_ATTR_NONNULL_ALL FUNC_ATTR_WARN_UNUSED_RESULT
-{
-#define INIT_SPECIAL_DICT(tv, type, val) \
-  do { \
-    dict_T *const dict = dict_alloc(); \
-    dictitem_T *const type_di = dictitem_alloc((char_u *) "_TYPE"); \
-    type_di->di_tv.v_type = VAR_LIST; \
-    type_di->di_tv.v_lock = 0; \
-    type_di->di_tv.vval.v_list = (list_T *) eval_msgpack_type_lists[type]; \
-    type_di->di_tv.vval.v_list->lv_refcount++; \
-    dict_add(dict, type_di); \
-    dictitem_T *const val_di = dictitem_alloc((char_u *) "_VAL"); \
-    val_di->di_tv = val; \
-    dict_add(dict, val_di); \
-    tv->v_type = VAR_DICT; \
-    dict->dv_refcount++; \
-    tv->vval.v_dict = dict; \
-  } while (0)
-  switch (mobj.type) {
-    case MSGPACK_OBJECT_NIL: {
-      INIT_SPECIAL_DICT(rettv, kMPNil, ((typval_T) {
-                                          .v_type = VAR_NUMBER,
-                                          .v_lock = 0,
-                                          .vval = { .v_number = 0 },
-                                        }));
-      break;
-    }
-    case MSGPACK_OBJECT_BOOLEAN: {
-      INIT_SPECIAL_DICT(rettv, kMPBoolean,
-                        ((typval_T) {
-                           .v_type = VAR_NUMBER,
-                           .v_lock = 0,
-                           .vval = {
-                             .v_number = (varnumber_T) mobj.via.boolean,
-                           },
-                         }));
-      break;
-    }
-    case MSGPACK_OBJECT_POSITIVE_INTEGER: {
-      if (mobj.via.u64 <= VARNUMBER_MAX) {
-        *rettv = (typval_T) {
-          .v_type = VAR_NUMBER,
-          .v_lock = 0,
-          .vval = { .v_number = (varnumber_T) mobj.via.u64 },
-        };
-      } else {
-        list_T *const list = list_alloc();
-        list->lv_refcount++;
-        INIT_SPECIAL_DICT(rettv, kMPInteger,
-                          ((typval_T) {
-                             .v_type = VAR_LIST,
-                             .v_lock = 0,
-                             .vval = { .v_list = list },
-                           }));
-        uint64_t n = mobj.via.u64;
-        list_append_number(list, 1);
-        list_append_number(list, (varnumber_T) ((n >> 62) & 0x3));
-        list_append_number(list, (varnumber_T) ((n >> 31) & 0x7FFFFFFF));
-        list_append_number(list, (varnumber_T) (n & 0x7FFFFFFF));
-      }
-      break;
-    }
-    case MSGPACK_OBJECT_NEGATIVE_INTEGER: {
-      if (mobj.via.i64 >= VARNUMBER_MIN) {
-        *rettv = (typval_T) {
-          .v_type = VAR_NUMBER,
-          .v_lock = 0,
-          .vval = { .v_number = (varnumber_T) mobj.via.i64 },
-        };
-      } else {
-        list_T *const list = list_alloc();
-        list->lv_refcount++;
-        INIT_SPECIAL_DICT(rettv, kMPInteger,
-                          ((typval_T) {
-                             .v_type = VAR_LIST,
-                             .v_lock = 0,
-                             .vval = { .v_list = list },
-                           }));
-        uint64_t n = -((uint64_t) mobj.via.i64);
-        list_append_number(list, -1);
-        list_append_number(list, (varnumber_T) ((n >> 62) & 0x3));
-        list_append_number(list, (varnumber_T) ((n >> 31) & 0x7FFFFFFF));
-        list_append_number(list, (varnumber_T) (n & 0x7FFFFFFF));
-      }
-      break;
-    }
-    case MSGPACK_OBJECT_FLOAT: {
-      *rettv = (typval_T) {
-        .v_type = VAR_FLOAT,
-        .v_lock = 0,
-        .vval = { .v_float = mobj.via.f64 },
-      };
-      break;
-    }
-    case MSGPACK_OBJECT_STR: {
-      list_T *const list = list_alloc();
-      list->lv_refcount++;
-      INIT_SPECIAL_DICT(rettv, kMPString,
-                        ((typval_T) {
-                           .v_type = VAR_LIST,
-                           .v_lock = 0,
-                           .vval = { .v_list = list },
-                         }));
-      if (msgpack_list_write((void *) list, mobj.via.str.ptr, mobj.via.str.size)
-          == -1) {
-        return FAIL;
-      }
-      break;
-    }
-    case MSGPACK_OBJECT_BIN: {
-      if (memchr(mobj.via.bin.ptr, NUL, mobj.via.bin.size) == NULL) {
-        *rettv = (typval_T) {
-          .v_type = VAR_STRING,
-          .v_lock = 0,
-          .vval = { .v_string = xmemdupz(mobj.via.bin.ptr, mobj.via.bin.size) },
-        };
-        break;
-      }
-      list_T *const list = list_alloc();
-      list->lv_refcount++;
-      INIT_SPECIAL_DICT(rettv, kMPBinary,
-                        ((typval_T) {
-                           .v_type = VAR_LIST,
-                           .v_lock = 0,
-                           .vval = { .v_list = list },
-                         }));
-      if (msgpack_list_write((void *) list, mobj.via.bin.ptr, mobj.via.bin.size)
-          == -1) {
-        return FAIL;
-      }
-      break;
-    }
-    case MSGPACK_OBJECT_ARRAY: {
-      list_T *const list = list_alloc();
-      list->lv_refcount++;
-      *rettv = (typval_T) {
-        .v_type = VAR_LIST,
-        .v_lock = 0,
-        .vval = { .v_list = list },
-      };
-      for (size_t i = 0; i < mobj.via.array.size; i++) {
-        listitem_T *const li = listitem_alloc();
-        li->li_tv.v_type = VAR_UNKNOWN;
-        list_append(list, li);
-        if (msgpack_to_vim(mobj.via.array.ptr[i], &li->li_tv) == FAIL) {
-          return FAIL;
-        }
-      }
-      break;
-    }
-    case MSGPACK_OBJECT_MAP: {
-      for (size_t i = 0; i < mobj.via.map.size; i++) {
-        if (mobj.via.map.ptr[i].key.type != MSGPACK_OBJECT_STR
-            || mobj.via.map.ptr[i].key.via.str.size == 0
-            || memchr(mobj.via.map.ptr[i].key.via.str.ptr, NUL,
-                      mobj.via.map.ptr[i].key.via.str.size) != NULL) {
-          goto msgpack_to_vim_generic_map;
-        }
-      }
-      dict_T *const dict = dict_alloc();
-      dict->dv_refcount++;
-      *rettv = (typval_T) {
-        .v_type = VAR_DICT,
-        .v_lock = 0,
-        .vval = { .v_dict = dict },
-      };
-      for (size_t i = 0; i < mobj.via.map.size; i++) {
-        dictitem_T *const di = xmallocz(offsetof(dictitem_T, di_key)
-                                        + mobj.via.map.ptr[i].key.via.str.size);
-        memcpy(&di->di_key[0], mobj.via.map.ptr[i].key.via.str.ptr,
-               mobj.via.map.ptr[i].key.via.str.size);
-        di->di_tv.v_type = VAR_UNKNOWN;
-        if (dict_add(dict, di) == FAIL) {
-          // Duplicate key: fallback to generic map
-          clear_tv(rettv);
-          xfree(di);
-          goto msgpack_to_vim_generic_map;
-        }
-        if (msgpack_to_vim(mobj.via.map.ptr[i].val, &di->di_tv) == FAIL) {
-          return FAIL;
-        }
-      }
-      break;
-msgpack_to_vim_generic_map: {}
-      list_T *const list = list_alloc();
-      list->lv_refcount++;
-      INIT_SPECIAL_DICT(rettv, kMPMap,
-                        ((typval_T) {
-                           .v_type = VAR_LIST,
-                           .v_lock = 0,
-                           .vval = { .v_list = list },
-                         }));
-      for (size_t i = 0; i < mobj.via.map.size; i++) {
-        list_T *const kv_pair = list_alloc();
-        list_append_list(list, kv_pair);
-        listitem_T *const key_li = listitem_alloc();
-        key_li->li_tv.v_type = VAR_UNKNOWN;
-        list_append(kv_pair, key_li);
-        listitem_T *const val_li = listitem_alloc();
-        val_li->li_tv.v_type = VAR_UNKNOWN;
-        list_append(kv_pair, val_li);
-        if (msgpack_to_vim(mobj.via.map.ptr[i].key, &key_li->li_tv) == FAIL) {
-          return FAIL;
-        }
-        if (msgpack_to_vim(mobj.via.map.ptr[i].val, &val_li->li_tv) == FAIL) {
-          return FAIL;
-        }
-      }
-      break;
-    }
-    case MSGPACK_OBJECT_EXT: {
-      list_T *const list = list_alloc();
-      list->lv_refcount++;
-      list_append_number(list, mobj.via.ext.type);
-      list_T *const ext_val_list = list_alloc();
-      list_append_list(list, ext_val_list);
-      INIT_SPECIAL_DICT(rettv, kMPExt,
-                        ((typval_T) {
-                           .v_type = VAR_LIST,
-                           .v_lock = 0,
-                           .vval = { .v_list = list },
-                         }));
-      if (msgpack_list_write((void *) ext_val_list, mobj.via.ext.ptr,
-                             mobj.via.ext.size) == -1) {
-        return FAIL;
-      }
-      break;
-    }
-  }
-#undef INIT_SPECIAL_DICT
-  return OK;
-}
-
 /// "msgpackparse" function
 static void f_msgpackparse(typval_T *argvars, typval_T *rettv)
   FUNC_ATTR_NONNULL_ALL
@@ -15927,7 +15136,7 @@ static void f_str2float(typval_T *argvars, typval_T *rettv)
 
   if (*p == '+')
     p = skipwhite(p + 1);
-  (void)string2float(p, &rettv->vval.v_float);
+  (void) string2float((char *) p, &rettv->vval.v_float);
   rettv->v_type = VAR_FLOAT;
 }
 
@@ -17851,6 +17060,18 @@ long get_vim_var_nr(int idx) FUNC_ATTR_PURE
   return vimvars[idx].vv_nr;
 }
 
+/// Get typval_T representing v: variable
+///
+/// @warning if v: variable has reference counter it is not increased.
+///
+/// @param[in]  idx  Variable index, @see VimVarIndex.
+typval_T get_vim_var_tv(VimVarIndex idx) FUNC_ATTR_PURE
+{
+  typval_T ret = vimvars[idx].vv_di.di_tv;
+  ret.v_lock = VAR_UNLOCKED;
+  return ret;
+}
+
 /*
  * Get string v: variable value.  Uses a static buffer, can only be used once.
  */
diff --git a/src/nvim/eval/decode.c b/src/nvim/eval/decode.c
new file mode 100644
index 0000000000..e2b2574d83
--- /dev/null
+++ b/src/nvim/eval/decode.c
@@ -0,0 +1,753 @@
+#include <stddef.h>
+
+#include <msgpack.h>
+
+#include "nvim/eval_defs.h"
+#include "nvim/eval.h"
+#include "nvim/eval/encode.h"
+#include "nvim/ascii.h"
+#include "nvim/message.h"
+#include "nvim/charset.h"  // vim_str2nr
+#include "nvim/lib/kvec.h"
+#include "nvim/vim.h"  // OK, FAIL
+
+/// Helper structure for container_struct
+typedef struct {
+  size_t stack_index;  ///< Index of current container in stack.
+  typval_T container;  ///< Container. Either VAR_LIST, VAR_DICT or VAR_LIST
+                       ///< which is _VAL from special dictionary.
+} ContainerStackItem;
+
+typedef kvec_t(typval_T) ValuesStack;
+typedef kvec_t(ContainerStackItem) ContainerStack;
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "eval/decode.c.generated.h"
+#endif
+
+/// Helper function used for working with stack vectors used by JSON decoder
+///
+/// @param[in]  obj  New object.
+/// @param[out]  stack  Object stack.
+/// @param[out]  container_stack  Container objects stack.
+/// @param[in]  p  Position in string which is currently being parsed.
+///
+/// @return OK in case of success, FAIL in case of error.
+static inline int json_decoder_pop(typval_T obj, ValuesStack *const stack,
+                                   ContainerStack *const container_stack,
+                                   const char *const p)
+  FUNC_ATTR_NONNULL_ALL
+{
+  if (kv_size(*container_stack) == 0) {
+    kv_push(typval_T, *stack, obj);
+    return OK;
+  }
+  ContainerStackItem last_container = kv_last(*container_stack);
+  if (obj.v_type == last_container.container.v_type
+      // vval.v_list and vval.v_dict should have the same size and offset
+      && ((void *) obj.vval.v_list
+          == (void *) last_container.container.vval.v_list)) {
+    kv_pop(*container_stack);
+    last_container = kv_last(*container_stack);
+  }
+  if (last_container.container.v_type == VAR_LIST) {
+    listitem_T *obj_li = listitem_alloc();
+    obj_li->li_tv = obj;
+    list_append(last_container.container.vval.v_list, obj_li);
+  } else if (last_container.stack_index == kv_size(*stack) - 2) {
+    typval_T key = kv_pop(*stack);
+    if (key.v_type != VAR_STRING) {
+      assert(false);
+    } else if (key.vval.v_string == NULL || *key.vval.v_string == NUL) {
+      // TODO: fall back to special dict in case of empty key
+      EMSG(_("E474: Empty key"));
+      clear_tv(&obj);
+      return FAIL;
+    }
+    dictitem_T *obj_di = dictitem_alloc(key.vval.v_string);
+    clear_tv(&key);
+    if (dict_add(last_container.container.vval.v_dict, obj_di)
+        == FAIL) {
+      // TODO: fall back to special dict in case of duplicate keys
+      EMSG(_("E474: Duplicate key"));
+      dictitem_free(obj_di);
+      clear_tv(&obj);
+      return FAIL;
+    }
+    obj_di->di_tv = obj;
+  } else {
+    // Object with key only
+    if (obj.v_type != VAR_STRING) {
+      EMSG2(_("E474: Expected string key: %s"), p);
+      clear_tv(&obj);
+      return FAIL;
+    }
+    kv_push(typval_T, *stack, obj);
+  }
+  return OK;
+}
+
+/// Convert JSON string into VimL object
+///
+/// @param[in]  buf  String to convert. UTF-8 encoding is assumed.
+/// @param[in]  len  Length of the string.
+/// @param[out]  rettv  Location where to save results.
+///
+/// @return OK in case of success, FAIL otherwise.
+int json_decode_string(const char *const buf, const size_t len,
+                       typval_T *const rettv)
+  FUNC_ATTR_NONNULL_ALL FUNC_ATTR_WARN_UNUSED_RESULT
+{
+  vimconv_T conv;
+  convert_setup(&conv, (char_u *) "utf-8", p_enc);
+  conv.vc_fail = true;
+  int ret = OK;
+  ValuesStack stack;
+  kv_init(stack);
+  ContainerStack container_stack;
+  kv_init(container_stack);
+  rettv->v_type = VAR_UNKNOWN;
+  const char *const e = buf + len;
+  bool didcomma = false;
+  bool didcolon = false;
+#define POP(obj) \
+  do { \
+    if (json_decoder_pop(obj, &stack, &container_stack, p) == FAIL) { \
+      goto json_decode_string_fail; \
+    } \
+  } while (0)
+  const char *p = buf;
+  for (; p < e; p++) {
+    switch (*p) {
+      case '}':
+      case ']': {
+        if (kv_size(container_stack) == 0) {
+          EMSG2(_("E474: No container to close: %s"), p);
+          goto json_decode_string_fail;
+        }
+        ContainerStackItem last_container = kv_last(container_stack);
+        if (*p == '}' && last_container.container.v_type != VAR_DICT) {
+          EMSG2(_("E474: Closing list with figure brace: %s"), p);
+          goto json_decode_string_fail;
+        } else if (*p == ']' && last_container.container.v_type != VAR_LIST) {
+          EMSG2(_("E474: Closing dictionary with bracket: %s"), p);
+          goto json_decode_string_fail;
+        } else if (didcomma) {
+          EMSG2(_("E474: Trailing comma: %s"), p);
+          goto json_decode_string_fail;
+        } else if (didcolon) {
+          EMSG2(_("E474: Expected value after colon: %s"), p);
+          goto json_decode_string_fail;
+        } else if (last_container.stack_index != kv_size(stack) - 1) {
+          assert(last_container.stack_index < kv_size(stack) - 1);
+          EMSG2(_("E474: Expected value: %s"), p);
+          goto json_decode_string_fail;
+        }
+        if (kv_size(stack) == 1) {
+          p++;
+          kv_pop(container_stack);
+          goto json_decode_string_after_cycle;
+        } else {
+          typval_T obj = kv_pop(stack);
+          POP(obj);
+          break;
+        }
+      }
+      case ',': {
+        if (kv_size(container_stack) == 0) {
+          EMSG2(_("E474: Comma not inside container: %s"), p);
+          goto json_decode_string_fail;
+        }
+        ContainerStackItem last_container = kv_last(container_stack);
+        if (didcomma) {
+          EMSG2(_("E474: Duplicate comma: %s"), p);
+          goto json_decode_string_fail;
+        } else if (didcolon) {
+          EMSG2(_("E474: Comma after colon: %s"), p);
+          goto json_decode_string_fail;
+        } if (last_container.container.v_type == VAR_DICT
+            && last_container.stack_index != kv_size(stack) - 1) {
+          EMSG2(_("E474: Using comma in place of colon: %s"), p);
+          goto json_decode_string_fail;
+        } else if ((last_container.container.v_type == VAR_DICT
+                    && (last_container.container.vval.v_dict->dv_hashtab.ht_used
+                        == 0))
+                   || (last_container.container.v_type == VAR_LIST
+                       && last_container.container.vval.v_list->lv_len == 0)) {
+          EMSG2(_("E474: Leading comma: %s"), p);
+          goto json_decode_string_fail;
+        }
+        didcomma = true;
+        continue;
+      }
+      case ':': {
+        if (kv_size(container_stack) == 0) {
+          EMSG2(_("E474: Colon not inside container: %s"), p);
+          goto json_decode_string_fail;
+        }
+        ContainerStackItem last_container = kv_last(container_stack);
+        if (last_container.container.v_type != VAR_DICT) {
+          EMSG2(_("E474: Using colon not in dictionary: %s"), p);
+          goto json_decode_string_fail;
+        } else if (last_container.stack_index != kv_size(stack) - 2) {
+          EMSG2(_("E474: Unexpected colon: %s"), p);
+          goto json_decode_string_fail;
+        } else if (didcomma) {
+          EMSG2(_("E474: Colon after comma: %s"), p);
+          goto json_decode_string_fail;
+        } else if (didcolon) {
+          EMSG2(_("E474: Duplicate colon: %s"), p);
+          goto json_decode_string_fail;
+        }
+        didcolon = true;
+        continue;
+      }
+      case ' ':
+      case TAB:
+      case NL: {
+        continue;
+      }
+      case 'n': {
+        if (strncmp(p + 1, "ull", 3) != 0) {
+          EMSG2(_("E474: Expected null: %s"), p);
+          goto json_decode_string_fail;
+        }
+        p += 3;
+        POP(get_vim_var_tv(VV_NULL));
+        break;
+      }
+      case 't': {
+        if (strncmp(p + 1, "rue", 3) != 0) {
+          EMSG2(_("E474: Expected true: %s"), p);
+          goto json_decode_string_fail;
+        }
+        p += 3;
+        POP(get_vim_var_tv(VV_TRUE));
+        break;
+      }
+      case 'f': {
+        if (strncmp(p + 1, "alse", 4) != 0) {
+          EMSG2(_("E474: Expected false: %s"), p);
+          goto json_decode_string_fail;
+        }
+        p += 4;
+        POP(get_vim_var_tv(VV_FALSE));
+        break;
+      }
+      case '"': {
+        size_t len = 0;
+        const char *s;
+        for (s = ++p; p < e && *p != '"'; p++) {
+          if (*p == '\\') {
+            p++;
+            if (p == e) {
+              EMSG2(_("E474: Unfinished escape sequence: %s"), buf);
+              goto json_decode_string_fail;
+            }
+            switch (*p) {
+              case 'u': {
+                if (p + 4 >= e) {
+                  EMSG2(_("E474: Unfinished unicode escape sequence: %s"), buf);
+                  goto json_decode_string_fail;
+                } else if (!ascii_isxdigit(p[1])
+                           || !ascii_isxdigit(p[2])
+                           || !ascii_isxdigit(p[3])
+                           || !ascii_isxdigit(p[4])) {
+                  EMSG2(_("E474: Expected four hex digits after \\u: %s"),
+                        p - 1);
+                  goto json_decode_string_fail;
+                }
+                // One UTF-8 character below U+10000 can take up to 3 bytes
+                len += 3;
+                p += 4;
+                break;
+              }
+              case '\\':
+              case '/':
+              case '"':
+              case 't':
+              case 'b':
+              case 'n':
+              case 'r':
+              case 'f': {
+                len++;
+                break;
+              }
+              default: {
+                EMSG2(_("E474: Unknown escape sequence: %s"), p - 1);
+                goto json_decode_string_fail;
+              }
+            }
+          } else {
+            len++;
+          }
+        }
+        if (*p != '"') {
+          EMSG2(_("E474: Expected string end: %s"), buf);
+          goto json_decode_string_fail;
+        }
+        char *str = xmalloc(len + 1);
+        uint16_t fst_in_pair = 0;
+        char *str_end = str;
+        for (const char *t = s; t < p; t++) {
+          if (t[0] != '\\' || t[1] != 'u') {
+            if (fst_in_pair != 0) {
+              str_end += utf_char2bytes((int) fst_in_pair, (char_u *) str_end);
+              fst_in_pair = 0;
+            }
+          }
+          if (*t == '\\') {
+            t++;
+            switch (*t) {
+              case 'u': {
+                char ubuf[] = { t[1], t[2], t[3], t[4], 0 };
+                t += 4;
+                unsigned long ch;
+                vim_str2nr((char_u *) ubuf, NULL, NULL, 0, 0, 2, NULL, &ch);
+                if (0xD800UL <= ch && ch <= 0xDB7FUL) {
+                  fst_in_pair = (uint16_t) ch;
+                } else if (0xDC00ULL <= ch && ch <= 0xDB7FUL) {
+                  if (fst_in_pair != 0) {
+                    int full_char = (
+                        (int) (ch - 0xDC00UL)
+                        + (((int) (fst_in_pair - 0xD800)) << 10)
+                    );
+                    str_end += utf_char2bytes(full_char, (char_u *) str_end);
+                  }
+                } else {
+                  str_end += utf_char2bytes((int) ch, (char_u *) str_end);
+                }
+                break;
+              }
+              case '\\':
+              case '/':
+              case '"':
+              case 't':
+              case 'b':
+              case 'n':
+              case 'r':
+              case 'f': {
+                static const char escapes[] = {
+                  ['\\'] = '\\',
+                  ['/'] = '/',
+                  ['"'] = '"',
+                  ['t'] = TAB,
+                  ['b'] = BS,
+                  ['n'] = NL,
+                  ['r'] = CAR,
+                  ['f'] = FF,
+                };
+                *str_end++ = escapes[(int) *t];
+                break;
+              }
+              default: {
+                assert(false);
+              }
+            }
+          } else {
+            *str_end++ = *t;
+          }
+        }
+        if (fst_in_pair != 0) {
+          str_end += utf_char2bytes((int) fst_in_pair, (char_u *) str_end);
+        }
+        if (conv.vc_type != CONV_NONE) {
+          size_t len = (size_t) (str_end - str);
+          char *const new_str = (char *) string_convert(&conv, (char_u *) str,
+                                                        &len);
+          if (new_str == NULL) {
+            EMSG2(_("E474: Failed to convert string \"%s\" from UTF-8"), str);
+            xfree(str);
+            goto json_decode_string_fail;
+          }
+          xfree(str);
+          str = new_str;
+          str_end = new_str + len;
+        }
+        *str_end = NUL;
+        // TODO: return special string in case of NUL bytes
+        POP(((typval_T) {
+          .v_type = VAR_STRING,
+          .vval = { .v_string = (char_u *) str, },
+        }));
+        break;
+      }
+      case '-':
+      case '0':
+      case '1':
+      case '2':
+      case '3':
+      case '4':
+      case '5':
+      case '6':
+      case '7':
+      case '8':
+      case '9': {
+        // a.bE[+-]exp
+        const char *const s = p;
+        const char *ints = NULL;
+        const char *fracs = NULL;
+        const char *exps = NULL;
+        if (*p == '-') {
+          p++;
+        }
+        ints = p;
+        while (p < e && ascii_isdigit(*p)) {
+          p++;
+        }
+        if (p < e && *p == '.') {
+          p++;
+          fracs = p;
+          while (p < e && ascii_isdigit(*p)) {
+            p++;
+          }
+          if (p < e && (*p == 'e' || *p == 'E')) {
+            p++;
+            if (p < e && (*p == '-' || *p == '+')) {
+              p++;
+            }
+            exps = p;
+            while (p < e && ascii_isdigit(*p)) {
+              p++;
+            }
+          }
+        }
+        if (p == ints) {
+          EMSG2(_("E474: Missing number after minus sign: %s"), s);
+          goto json_decode_string_fail;
+        } else if (p == fracs) {
+          EMSG2(_("E474: Missing number after decimal dot: %s"), s);
+          goto json_decode_string_fail;
+        } else if (p == exps) {
+          EMSG2(_("E474: Missing exponent: %s"), s);
+          goto json_decode_string_fail;
+        }
+        typval_T tv = {
+          .v_type = VAR_NUMBER,
+          .v_lock = VAR_UNLOCKED,
+        };
+        if (fracs) {
+          // Convert floating-point number
+          (void) string2float(s, &tv.vval.v_float);
+          tv.v_type = VAR_FLOAT;
+        } else {
+          // Convert integer
+          long nr;
+          vim_str2nr((char_u *) s, NULL, NULL, 0, 0, 0, &nr, NULL);
+          tv.vval.v_number = (varnumber_T) nr;
+        }
+        POP(tv);
+        p--;
+        break;
+      }
+      case '[': {
+        list_T *list = list_alloc();
+        list->lv_refcount++;
+        typval_T tv = {
+          .v_type = VAR_LIST,
+          .v_lock = VAR_UNLOCKED,
+          .vval = { .v_list = list },
+        };
+        kv_push(ContainerStackItem, container_stack, ((ContainerStackItem) {
+          .stack_index = kv_size(stack),
+          .container = tv,
+        }));
+        kv_push(typval_T, stack, tv);
+        break;
+      }
+      case '{': {
+        dict_T *dict = dict_alloc();
+        dict->dv_refcount++;
+        typval_T tv = {
+          .v_type = VAR_DICT,
+          .v_lock = VAR_UNLOCKED,
+          .vval = { .v_dict = dict },
+        };
+        kv_push(ContainerStackItem, container_stack, ((ContainerStackItem) {
+          .stack_index = kv_size(stack),
+          .container = tv,
+        }));
+        kv_push(typval_T, stack, tv);
+        break;
+      }
+      default: {
+        EMSG2(_("E474: Unidentified byte: %s"), p);
+        goto json_decode_string_fail;
+      }
+    }
+    didcomma = false;
+    didcolon = false;
+    if (kv_size(container_stack) == 0) {
+      p++;
+      break;
+    }
+  }
+#undef POP
+json_decode_string_after_cycle:
+  for (; p < e; p++) {
+    switch (*p) {
+      case NL:
+      case ' ':
+      case TAB: {
+        break;
+      }
+      default: {
+        EMSG2(_("E474: Trailing characters: %s"), p);
+        goto json_decode_string_fail;
+      }
+    }
+  }
+  if (kv_size(stack) > 1 || kv_size(container_stack)) {
+    EMSG2(_("E474: Unexpected end of input: %s"), buf);
+    goto json_decode_string_fail;
+  }
+  goto json_decode_string_ret;
+json_decode_string_fail:
+  ret = FAIL;
+  while (kv_size(stack)) {
+    clear_tv(&kv_pop(stack));
+  }
+json_decode_string_ret:
+  if (ret != FAIL) {
+    assert(kv_size(stack) == 1);
+    *rettv = kv_pop(stack);
+  }
+  kv_destroy(stack);
+  kv_destroy(container_stack);
+  return ret;
+}
+
+/// Convert msgpack object to a VimL one
+int msgpack_to_vim(const msgpack_object mobj, typval_T *const rettv)
+  FUNC_ATTR_NONNULL_ALL FUNC_ATTR_WARN_UNUSED_RESULT
+{
+#define INIT_SPECIAL_DICT(tv, type, val) \
+  do { \
+    dict_T *const dict = dict_alloc(); \
+    dictitem_T *const type_di = dictitem_alloc((char_u *) "_TYPE"); \
+    type_di->di_tv.v_type = VAR_LIST; \
+    type_di->di_tv.v_lock = 0; \
+    type_di->di_tv.vval.v_list = (list_T *) eval_msgpack_type_lists[type]; \
+    type_di->di_tv.vval.v_list->lv_refcount++; \
+    dict_add(dict, type_di); \
+    dictitem_T *const val_di = dictitem_alloc((char_u *) "_VAL"); \
+    val_di->di_tv = val; \
+    dict_add(dict, val_di); \
+    tv->v_type = VAR_DICT; \
+    dict->dv_refcount++; \
+    tv->vval.v_dict = dict; \
+  } while (0)
+  switch (mobj.type) {
+    case MSGPACK_OBJECT_NIL: {
+      INIT_SPECIAL_DICT(rettv, kMPNil, ((typval_T) {
+                                          .v_type = VAR_NUMBER,
+                                          .v_lock = 0,
+                                          .vval = { .v_number = 0 },
+                                        }));
+      break;
+    }
+    case MSGPACK_OBJECT_BOOLEAN: {
+      INIT_SPECIAL_DICT(rettv, kMPBoolean,
+                        ((typval_T) {
+                           .v_type = VAR_NUMBER,
+                           .v_lock = 0,
+                           .vval = {
+                             .v_number = (varnumber_T) mobj.via.boolean,
+                           },
+                         }));
+      break;
+    }
+    case MSGPACK_OBJECT_POSITIVE_INTEGER: {
+      if (mobj.via.u64 <= VARNUMBER_MAX) {
+        *rettv = (typval_T) {
+          .v_type = VAR_NUMBER,
+          .v_lock = 0,
+          .vval = { .v_number = (varnumber_T) mobj.via.u64 },
+        };
+      } else {
+        list_T *const list = list_alloc();
+        list->lv_refcount++;
+        INIT_SPECIAL_DICT(rettv, kMPInteger,
+                          ((typval_T) {
+                             .v_type = VAR_LIST,
+                             .v_lock = 0,
+                             .vval = { .v_list = list },
+                           }));
+        uint64_t n = mobj.via.u64;
+        list_append_number(list, 1);
+        list_append_number(list, (varnumber_T) ((n >> 62) & 0x3));
+        list_append_number(list, (varnumber_T) ((n >> 31) & 0x7FFFFFFF));
+        list_append_number(list, (varnumber_T) (n & 0x7FFFFFFF));
+      }
+      break;
+    }
+    case MSGPACK_OBJECT_NEGATIVE_INTEGER: {
+      if (mobj.via.i64 >= VARNUMBER_MIN) {
+        *rettv = (typval_T) {
+          .v_type = VAR_NUMBER,
+          .v_lock = 0,
+          .vval = { .v_number = (varnumber_T) mobj.via.i64 },
+        };
+      } else {
+        list_T *const list = list_alloc();
+        list->lv_refcount++;
+        INIT_SPECIAL_DICT(rettv, kMPInteger,
+                          ((typval_T) {
+                             .v_type = VAR_LIST,
+                             .v_lock = 0,
+                             .vval = { .v_list = list },
+                           }));
+        uint64_t n = -((uint64_t) mobj.via.i64);
+        list_append_number(list, -1);
+        list_append_number(list, (varnumber_T) ((n >> 62) & 0x3));
+        list_append_number(list, (varnumber_T) ((n >> 31) & 0x7FFFFFFF));
+        list_append_number(list, (varnumber_T) (n & 0x7FFFFFFF));
+      }
+      break;
+    }
+    case MSGPACK_OBJECT_FLOAT: {
+      *rettv = (typval_T) {
+        .v_type = VAR_FLOAT,
+        .v_lock = 0,
+        .vval = { .v_float = mobj.via.f64 },
+      };
+      break;
+    }
+    case MSGPACK_OBJECT_STR: {
+      list_T *const list = list_alloc();
+      list->lv_refcount++;
+      INIT_SPECIAL_DICT(rettv, kMPString,
+                        ((typval_T) {
+                           .v_type = VAR_LIST,
+                           .v_lock = 0,
+                           .vval = { .v_list = list },
+                         }));
+      if (encode_list_write((void *) list, mobj.via.str.ptr, mobj.via.str.size)
+          == -1) {
+        return FAIL;
+      }
+      break;
+    }
+    case MSGPACK_OBJECT_BIN: {
+      if (memchr(mobj.via.bin.ptr, NUL, mobj.via.bin.size) == NULL) {
+        *rettv = (typval_T) {
+          .v_type = VAR_STRING,
+          .v_lock = 0,
+          .vval = { .v_string = xmemdupz(mobj.via.bin.ptr, mobj.via.bin.size) },
+        };
+        break;
+      }
+      list_T *const list = list_alloc();
+      list->lv_refcount++;
+      INIT_SPECIAL_DICT(rettv, kMPBinary,
+                        ((typval_T) {
+                           .v_type = VAR_LIST,
+                           .v_lock = 0,
+                           .vval = { .v_list = list },
+                         }));
+      if (encode_list_write((void *) list, mobj.via.bin.ptr, mobj.via.bin.size)
+          == -1) {
+        return FAIL;
+      }
+      break;
+    }
+    case MSGPACK_OBJECT_ARRAY: {
+      list_T *const list = list_alloc();
+      list->lv_refcount++;
+      *rettv = (typval_T) {
+        .v_type = VAR_LIST,
+        .v_lock = 0,
+        .vval = { .v_list = list },
+      };
+      for (size_t i = 0; i < mobj.via.array.size; i++) {
+        listitem_T *const li = listitem_alloc();
+        li->li_tv.v_type = VAR_UNKNOWN;
+        list_append(list, li);
+        if (msgpack_to_vim(mobj.via.array.ptr[i], &li->li_tv) == FAIL) {
+          return FAIL;
+        }
+      }
+      break;
+    }
+    case MSGPACK_OBJECT_MAP: {
+      for (size_t i = 0; i < mobj.via.map.size; i++) {
+        if (mobj.via.map.ptr[i].key.type != MSGPACK_OBJECT_STR
+            || mobj.via.map.ptr[i].key.via.str.size == 0
+            || memchr(mobj.via.map.ptr[i].key.via.str.ptr, NUL,
+                      mobj.via.map.ptr[i].key.via.str.size) != NULL) {
+          goto msgpack_to_vim_generic_map;
+        }
+      }
+      dict_T *const dict = dict_alloc();
+      dict->dv_refcount++;
+      *rettv = (typval_T) {
+        .v_type = VAR_DICT,
+        .v_lock = 0,
+        .vval = { .v_dict = dict },
+      };
+      for (size_t i = 0; i < mobj.via.map.size; i++) {
+        dictitem_T *const di = xmallocz(offsetof(dictitem_T, di_key)
+                                        + mobj.via.map.ptr[i].key.via.str.size);
+        memcpy(&di->di_key[0], mobj.via.map.ptr[i].key.via.str.ptr,
+               mobj.via.map.ptr[i].key.via.str.size);
+        di->di_tv.v_type = VAR_UNKNOWN;
+        if (dict_add(dict, di) == FAIL) {
+          // Duplicate key: fallback to generic map
+          clear_tv(rettv);
+          xfree(di);
+          goto msgpack_to_vim_generic_map;
+        }
+        if (msgpack_to_vim(mobj.via.map.ptr[i].val, &di->di_tv) == FAIL) {
+          return FAIL;
+        }
+      }
+      break;
+msgpack_to_vim_generic_map: {}
+      list_T *const list = list_alloc();
+      list->lv_refcount++;
+      INIT_SPECIAL_DICT(rettv, kMPMap,
+                        ((typval_T) {
+                           .v_type = VAR_LIST,
+                           .v_lock = 0,
+                           .vval = { .v_list = list },
+                         }));
+      for (size_t i = 0; i < mobj.via.map.size; i++) {
+        list_T *const kv_pair = list_alloc();
+        list_append_list(list, kv_pair);
+        listitem_T *const key_li = listitem_alloc();
+        key_li->li_tv.v_type = VAR_UNKNOWN;
+        list_append(kv_pair, key_li);
+        listitem_T *const val_li = listitem_alloc();
+        val_li->li_tv.v_type = VAR_UNKNOWN;
+        list_append(kv_pair, val_li);
+        if (msgpack_to_vim(mobj.via.map.ptr[i].key, &key_li->li_tv) == FAIL) {
+          return FAIL;
+        }
+        if (msgpack_to_vim(mobj.via.map.ptr[i].val, &val_li->li_tv) == FAIL) {
+          return FAIL;
+        }
+      }
+      break;
+    }
+    case MSGPACK_OBJECT_EXT: {
+      list_T *const list = list_alloc();
+      list->lv_refcount++;
+      list_append_number(list, mobj.via.ext.type);
+      list_T *const ext_val_list = list_alloc();
+      list_append_list(list, ext_val_list);
+      INIT_SPECIAL_DICT(rettv, kMPExt,
+                        ((typval_T) {
+                           .v_type = VAR_LIST,
+                           .v_lock = 0,
+                           .vval = { .v_list = list },
+                         }));
+      if (encode_list_write((void *) ext_val_list, mobj.via.ext.ptr,
+                             mobj.via.ext.size) == -1) {
+        return FAIL;
+      }
+      break;
+    }
+  }
+#undef INIT_SPECIAL_DICT
+  return OK;
+}
diff --git a/src/nvim/eval/decode.h b/src/nvim/eval/decode.h
new file mode 100644
index 0000000000..5c25a64f7a
--- /dev/null
+++ b/src/nvim/eval/decode.h
@@ -0,0 +1,13 @@
+#ifndef NVIM_EVAL_DECODE_H
+#define NVIM_EVAL_DECODE_H
+
+#include <stddef.h>
+
+#include <msgpack.h>
+
+#include "nvim/eval_defs.h"
+
+#ifdef INCLUDE_GENERATED_DECLARATIONS
+# include "eval/decode.h.generated.h"
+#endif
+#endif  // NVIM_EVAL_DECODE_H
diff --git a/src/nvim/eval/encode.c b/src/nvim/eval/encode.c
index e23e68dc62..359c9b3de7 100644
--- a/src/nvim/eval/encode.c
+++ b/src/nvim/eval/encode.c
@@ -65,7 +65,7 @@ const char *const encode_special_var_names[] = {
 #endif
 
 /// Msgpack callback for writing to readfile()-style list
-int msgpack_list_write(void *data, const char *buf, size_t len)
+int encode_list_write(void *data, const char *buf, size_t len)
 {
   if (len == 0) {
     return 0;
diff --git a/src/nvim/shada.c b/src/nvim/shada.c
index f402b75ad6..a18fd725d2 100644
--- a/src/nvim/shada.c
+++ b/src/nvim/shada.c
@@ -31,7 +31,6 @@
 #include "nvim/misc2.h"
 #include "nvim/ex_getln.h"
 #include "nvim/search.h"
-#include "nvim/eval.h"
 #include "nvim/regexp.h"
 #include "nvim/eval_defs.h"
 #include "nvim/version.h"
@@ -40,6 +39,7 @@
 #include "nvim/strings.h"
 #include "nvim/quickfix.h"
 #include "nvim/eval/encode.h"
+#include "nvim/eval/decode.h"
 #include "nvim/lib/khash.h"
 #include "nvim/lib/kvec.h"