parser: Better AST primitives

commit 1588e7b46d
Author: Aryadev Chavali <aryadev@aryadevchavali.com>
Date:   Sat Jan 24 02:55:12 2026 +0000

    parser/parser: parse_symbol now supports primitives

    parse_symbol now investigates if the parsed symbol data is actually
    just a primitive (linear search through all primitives).  If it is,
    return a primitive first.  Otherwise, generate a symbol as per
    previous form of routine.

commit 62c91990c4
Author: Aryadev Chavali <aryadev@aryadevchavali.com>
Date:   Sat Jan 24 02:40:26 2026 +0000

    parser/ast: Added support for node level primitives

    These are just an enumeration of primitives we already expect to be
    present within a program.  Instead of leaving everything as a symbol,
    we can compile certain symbols into the enumeration ahead of time to
    make later stages easier.

src/arl/parser/ast.c

@@ -5,15 +5,38 @@
  * Commentary: See ast.h.
  */
 
+#include <arl/lib/base.h>
 #include <arl/lib/vec.h>
 #include <arl/parser/ast.h>
 
+const char *ast_prim_to_cstr(ast_prim_t prim)
+{
+  switch (prim)
+  {
+  case AST_PRIM_NIL:
+    return "nil";
+  case AST_PRIM_PRINTLN:
+    return "println";
+  default:
+    FAIL("Unexpected AST primitive value: %d\n", prim);
+  }
+}
+
+ast_node_t ast_node_prim(u64 byte, ast_prim_t primitive)
+{
+  return (ast_node_t){
+      .byte_location = byte,
+      .type          = AST_NODE_TYPE_PRIMITIVE,
+      .value         = {.as_prim = primitive},
+  };
+}
+
 ast_node_t ast_node_string(u64 byte, sv_t string)
 {
   return (ast_node_t){
       .byte_location = byte,
       .type          = AST_NODE_TYPE_STRING,
-      .value         = {string},
+      .value         = {.as_string = string},
   };
 }
 
@@ -22,25 +45,31 @@ ast_node_t ast_node_symbol(u64 byte, sv_t symbol)
   return (ast_node_t){
       .byte_location = byte,
       .type          = AST_NODE_TYPE_SYMBOL,
-      .value         = {symbol},
+      .value         = {.as_symbol = symbol},
   };
 }
 
-void ast_node_print(FILE *fp, ast_node_t *obj)
+void ast_node_print(FILE *fp, ast_node_t *node)
 {
-  if (!obj)
+  if (!node)
   {
     fprintf(fp, "NIL");
     return;
   }
-  switch (obj->type)
+  switch (node->type)
   {
+  case AST_NODE_TYPE_PRIMITIVE:
+    fprintf(fp, "PRIMITIVE(%s)", ast_prim_to_cstr(node->value.as_prim));
+    break;
   case AST_NODE_TYPE_SYMBOL:
-    fprintf(fp, "SYMBOL(" PR_SV ")", SV_FMT(obj->value.as_symbol));
+    fprintf(fp, "SYMBOL(" PR_SV ")", SV_FMT(node->value.as_symbol));
     break;
   case AST_NODE_TYPE_STRING:
-    fprintf(fp, "STRING(" PR_SV ")", SV_FMT(obj->value.as_string));
+    fprintf(fp, "STRING(" PR_SV ")", SV_FMT(node->value.as_string));
     break;
+  case NUM_AST_NODE_TYPES:
+  default:
+    FAIL("Unexpected node type: %d\n", node->type);
   }
 }
 

src/arl/parser/ast.h

@@ -15,10 +15,24 @@
 /// Types the AST can encode
 typedef enum
 {
-  AST_NODE_TYPE_SYMBOL = 0,
+  AST_NODE_TYPE_PRIMITIVE = 0,
+  AST_NODE_TYPE_SYMBOL,
   AST_NODE_TYPE_STRING,
+
+  NUM_AST_NODE_TYPES,
 } ast_node_type_t;
 
+/// Primitives (values, callables, etc) as symbols
+typedef enum
+{
+  AST_PRIM_NIL = 0,
+  AST_PRIM_PRINTLN,
+
+  NUM_AST_PRIMS,
+} ast_prim_t;
+
+const char *ast_prim_to_cstr(ast_prim_t);
+
 /// Node of the AST as a tagged union
 typedef struct
 {
@@ -26,14 +40,16 @@ typedef struct
   ast_node_type_t type;
   union
   {
-    sv_t as_string;
+    ast_prim_t as_prim;
     sv_t as_symbol;
+    sv_t as_string;
   } value;
 } ast_node_t;
 
-ast_node_t ast_node_string(u64 byte, sv_t string);
+ast_node_t ast_node_prim(u64 byte, ast_prim_t primitive);
 ast_node_t ast_node_symbol(u64 byte, sv_t symbol);
-void ast_node_print(FILE *fp, ast_node_t *obj);
+ast_node_t ast_node_string(u64 byte, sv_t string);
+void ast_node_print(FILE *fp, ast_node_t *node);
 
 /// The AST as a flat collection of nodes
 typedef struct

src/arl/parser/parser.c

@@ -5,10 +5,11 @@
  * Commentary: See parser.h
  */
 
-#include "arl/lib/sv.h"
 #include <ctype.h>
 #include <string.h>
 
+#include <arl/lib/sv.h>
+#include <arl/parser/ast.h>
 #include <arl/parser/parser.h>
 
 /// Expected characters in a symbol
@@ -27,7 +28,7 @@ const char *parse_err_to_string(parse_err_t err)
   case PARSE_ERR_UNKNOWN_CHAR:
     return "UNKNOWN_CHAR";
   default:
-    FAIL("Unexpected value for parse_err_t: %d\n", err);
+    FAIL("Unexpected parse_err_t value: %d\n", err);
   }
 }
 
@@ -119,10 +120,27 @@ parse_err_t parse_string(parse_stream_t *stream, ast_node_t *ret)
 parse_err_t parse_symbol(parse_stream_t *stream, ast_node_t *ret)
 {
   sv_t current_contents = sv_chop_left(stream->contents, stream->byte);
-  u64 symbol_size       = sv_while(current_contents, SYMBOL_CHARS);
-  // Generate symbol
-  *ret = ast_node_symbol(stream->byte, SV(current_contents.data, symbol_size));
-  stream_advance(stream, symbol_size);
+  sv_t symbol =
+      SV(current_contents.data, sv_while(current_contents, SYMBOL_CHARS));
+
+  // see if symbol is one of the AST primitives we can parse AOT
+  static_assert(NUM_AST_PRIMS == 2, "Expected number of AST primitives");
+  for (ast_prim_t i = 0; i < NUM_AST_PRIMS; ++i)
+  {
+    const char *possible_prim = ast_prim_to_cstr(i);
+    if (strlen(possible_prim) == symbol.size &&
+        strncmp(possible_prim, symbol.data, symbol.size) == 0)
+    {
+      // Found a matching primitive
+      *ret = ast_node_prim(stream->byte, i);
+      goto end;
+    }
+  }
+
+  // otherwise, it must be a fresh symbol i.e. user defined
+  *ret = ast_node_symbol(stream->byte, symbol);
+end:
+  stream_advance(stream, symbol.size);
   return PARSE_ERR_OK;
 }