From 75dc36cd197ab47ffd1dbbad887dd7bac88c8add Mon Sep 17 00:00:00 2001
From: Aryadev Chavali <aryadev@aryadevchavali.com>
Date: Tue, 31 Oct 2023 20:39:26 +0000
Subject: Lexer now returns errors on failure

Currently only for invalid character literals, but still a possible
problem.
---
 asm/lexer.c | 69 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-------
 asm/lexer.h |  9 +++++++-
 asm/main.c  | 22 +++++++++++++++++++-
 3 files changed, 91 insertions(+), 9 deletions(-)

(limited to 'asm')

diff --git a/asm/lexer.c b/asm/lexer.c
index 51a8ec6..6a1f027 100644
--- a/asm/lexer.c
+++ b/asm/lexer.c
@@ -31,6 +31,20 @@ const char *token_type_as_cstr(token_type_t type)
   return "";
 }
 
+const char *lerr_as_cstr(lerr_t lerr)
+{
+  switch (lerr)
+  {
+  case LERR_INVALID_CHAR_LITERAL:
+    return "INVALID_CHAR_LITERAL";
+    break;
+  case LERR_OK:
+    return "OK";
+    break;
+  }
+  return "";
+}
+
 size_t space_left(buffer_t *buffer)
 {
   if (buffer->available == buffer->used)
@@ -95,10 +109,10 @@ token_t tokenise_char_literal(buffer_t *buffer, size_t *column)
   return token;
 }
 
-token_stream_t tokenise_buffer(buffer_t *buffer)
+lerr_t tokenise_buffer(buffer_t *buffer, token_stream_t *tokens_ptr)
 {
   size_t column = 0, line = 1;
-  buffer_t tokens = {0};
+  token_stream_t tokens = {0};
   darr_init(&tokens, sizeof(token_t));
   while (space_left(buffer) != 0)
   {
@@ -139,10 +153,50 @@ token_stream_t tokenise_buffer(buffer_t *buffer)
       t = tokenise_symbol(buffer, &column);
     else if (c == '\'')
     {
-      if (space_left(buffer) < 2 || buffer->data[buffer->used + 2] != '\'')
-        // TODO: Lex Error (INVALID_CHAR_LITERAL)
-        exit(1);
-      t = tokenise_char_literal(buffer, &column);
+      if (space_left(buffer) < 2)
+      {
+        free(tokens.data);
+        return LERR_INVALID_CHAR_LITERAL;
+      }
+      else if (buffer->data[buffer->used + 1] == '\\')
+      {
+        char escape = '\0';
+        if (space_left(buffer) < 3 || buffer->data[buffer->used + 3] != '\'')
+        {
+          free(tokens.data);
+          return LERR_INVALID_CHAR_LITERAL;
+        }
+        switch (buffer->data[buffer->used + 2])
+        {
+        case 'n':
+          escape = '\n';
+          break;
+        case 't':
+          escape = '\t';
+          break;
+        case 'r':
+          escape = '\r';
+          break;
+        case '\\':
+          escape = '\\';
+          break;
+        default:
+          column += 2;
+          free(tokens.data);
+          return LERR_INVALID_CHAR_LITERAL;
+          break;
+        }
+
+        t = (token_t){.type     = TOKEN_LITERAL_CHAR,
+                      .str      = malloc(1),
+                      .str_size = 1,
+                      .column   = column};
+        column += 4;
+        buffer->used += 4;
+        t.str[0] = escape;
+      }
+      else
+        t = tokenise_char_literal(buffer, &column);
     }
 
     if (is_token)
@@ -154,5 +208,6 @@ token_stream_t tokenise_buffer(buffer_t *buffer)
   size_t n_tokens  = tokens.used / sizeof(token_t);
   tokens.available = n_tokens;
   tokens.used      = 0;
-  return tokens;
+  *tokens_ptr      = tokens;
+  return LERR_OK;
 }
diff --git a/asm/lexer.h b/asm/lexer.h
index d2e0028..1e68d8b 100644
--- a/asm/lexer.h
+++ b/asm/lexer.h
@@ -30,11 +30,18 @@ typedef struct
   size_t str_size;
 } token_t;
 
+typedef enum
+{
+  LERR_OK = 0,
+  LERR_INVALID_CHAR_LITERAL,
+} lerr_t;
+const char *lerr_as_cstr(lerr_t);
+
 typedef darr_t buffer_t;
 typedef darr_t token_stream_t;
 #define TOKEN_STREAM_AT(STREAM_DATA, INDEX) (((token_t *)(STREAM_DATA))[INDEX])
 
 const char *token_type_as_cstr(token_type_t type);
-token_stream_t tokenise_buffer(buffer_t *);
+lerr_t tokenise_buffer(buffer_t *, token_stream_t *);
 
 #endif
diff --git a/asm/main.c b/asm/main.c
index 2f1102b..bfa12e0 100644
--- a/asm/main.c
+++ b/asm/main.c
@@ -41,7 +41,27 @@ int main(int argc, char *argv[])
   darr_t buffer = darr_read_file(fp);
   fclose(fp);
 
-  token_stream_t tokens = tokenise_buffer(&buffer);
+  token_stream_t tokens = {0};
+  lerr_t lex_error      = tokenise_buffer(&buffer, &tokens);
+  if (lex_error)
+  {
+    // Compute the line/newlines by hand
+    size_t column = 0, line = 1;
+    for (size_t i = 0; i < buffer.used; ++i)
+    {
+      if (buffer.data[i] == '\n')
+      {
+        column = 0;
+        ++line;
+      }
+      else
+        ++column;
+    }
+    fprintf(stderr, "%s:%lu:%lu: %s\n", source_file, line, column,
+            lerr_as_cstr(lex_error));
+    ret = 255 - lex_error;
+    goto end;
+  }
 #if VERBOSE >= 1
   printf("[%sTOKENISER%s]: %lu bytes -> %lu tokens\n", TERM_GREEN, TERM_RESET,
          buffer.used, tokens.available);
-- 
cgit v1.2.3-13-gbd6f