Add lexer user code expansions to access lexer match text positions

assets/parser.c.erb

@@ -486,6 +486,22 @@ static size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%= @
     case P_SUCCESS:
         {
         <%= @grammar.prefix %>token_t token_to_accept = match_info.accepting_state->token;
+        /* Calculate the token length and start/end positions before invoking
+         * the lexer user code so that the user code can access them. The
+         * context input text position tracking is not updated until after the
+         * user code has run so that it is left unchanged if the user code
+         * requests to terminate the lexer. */
+        token_info.length = match_info.length;
+        if (match_info.end_delta_position.row != 0u)
+        {
+            token_info.end_position.row = token_info.position.row + match_info.end_delta_position.row;
+            token_info.end_position.col = match_info.end_delta_position.col;
+        }
+        else
+        {
+            token_info.end_position.row = token_info.position.row;
+            token_info.end_position.col = token_info.position.col + match_info.end_delta_position.col;
+        }
         if (match_info.accepting_state->code_id != INVALID_USER_CODE_ID)
         {
             uint8_t const * match = &context->input[context->input_index];
@@ -524,17 +540,6 @@ static size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%= @
             return P_DROP;
         }
         token_info.token = token_to_accept;
-        token_info.length = match_info.length;
-        if (match_info.end_delta_position.row != 0u)
-        {
-            token_info.end_position.row = token_info.position.row + match_info.end_delta_position.row;
-            token_info.end_position.col = match_info.end_delta_position.col;
-        }
-        else
-        {
-            token_info.end_position.row = token_info.position.row;
-            token_info.end_position.col = token_info.position.col + match_info.end_delta_position.col;
-        }
         *out_token_info = token_info;
         }
         return P_SUCCESS;

assets/parser.d.erb

@@ -665,6 +665,22 @@ private size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%=
     {
     case P_SUCCESS:
         <%= @grammar.prefix %>token_t token_to_accept = match_info.accepting_state.token;
+        /* Calculate the token length and start/end positions before invoking
+         * the lexer user code so that the user code can access them. The
+         * context input text position tracking is not updated until after the
+         * user code has run so that it is left unchanged if the user code
+         * requests to terminate the lexer. */
+        token_info.length = match_info.length;
+        if (match_info.end_delta_position.row != 0u)
+        {
+            token_info.end_position.row = token_info.position.row + match_info.end_delta_position.row;
+            token_info.end_position.col = match_info.end_delta_position.col;
+        }
+        else
+        {
+            token_info.end_position.row = token_info.position.row;
+            token_info.end_position.col = token_info.position.col + match_info.end_delta_position.col;
+        }
         if (match_info.accepting_state.code_id != INVALID_USER_CODE_ID)
         {
             string match = context.input[context.input_index..(context.input_index + match_info.length)];
@@ -703,17 +719,6 @@ private size_t attempt_lex_token(<%= @grammar.prefix %>context_t * context, <%=
             return P_DROP;
         }
         token_info.token = token_to_accept;
-        token_info.length = match_info.length;
-        if (match_info.end_delta_position.row != 0u)
-        {
-            token_info.end_position.row = token_info.position.row + match_info.end_delta_position.row;
-            token_info.end_position.col = match_info.end_delta_position.col;
-        }
-        else
-        {
-            token_info.end_position.row = token_info.position.row;
-            token_info.end_position.col = token_info.position.col + match_info.end_delta_position.col;
-        }
         *out_token_info = token_info;
         return P_SUCCESS;
 

doc/user_guide.md

@@ -151,12 +151,37 @@ needed by a `ptype` directive, for example:
 
 ### Lexer pattern code blocks
 
-#### C/C++
+Lexer code blocks appear between `<<` and `>>` markers following a `drop`,
+`token`, or pattern expression.
+User code in a lexer code block will be executed when the lexer matches the
+given pattern.
+Assignment to the `$$` symbol will associate a parser value with the lexed
+token.
+This parser value can then be used later in a parser rule.
+
+The input text positions of the matched token can also be accessed from within
+a lexer code block.
+Each of these positions is an instance of the `p_position_t` structure (see
+`${#p_position_t}`), which contains 1-based `row` and `col` fields.
+The start position of the matched token is accessed with `${position}`, and
+the end position of the matched token is accessed with `${end_position}`.
+
+Example:
+
+```
+token integer /\d+/ <<
+  printf("integer token on row %d, col %d\n",
+      ${position}.row, ${position}.col);
+  $$ = parse_integer(match, match_length);
+>>
+```
+
+#### C/C++ lexer code block arguments
 
 The lexer code block is passed the following arguments:
 
-  * `match` - a pointer points to the text matched by the lexer pattern.
-  * `match_length` - length of the matched text.
+  * `match` (`uint8_t const`) - the pointer points to the text matched by the lexer pattern.
+  * `match_length` (`size_t`) - length of the matched text.
 
 Example:
 
@@ -174,11 +199,11 @@ token integer /\d+/ <<
 >>
 ```
 
-#### D
+#### D lexer code block arguments
 
 The lexer code block is passed the following arguments:
 
-  * `match` - a slice containing the text matched by the lexer pattern.
+  * `match` (`string`) - a slice containing the text matched by the lexer pattern.
 
 ```
 ptype ulong;
@@ -194,13 +219,6 @@ token integer /\d+/ <<
 >>
 ```
 
-Lexer code blocks appear following a `drop`, `token`, or pattern expression.
-User code in a lexer code block will be executed when the lexer matches the
-given pattern.
-Assignment to the `$$` symbol will associate a parser value with the lexed
-token.
-This parser value can then be used later in a parser rule.
-
 ### Parser rule code blocks
 
 Example:

lib/propane/generator.rb

@@ -351,6 +351,22 @@ class Propane
             end
           end
         end
+        code = code.gsub(/\$\{position\}/) do |match|
+          case @language
+          when "c"
+            "out_token_info->position"
+          when "d"
+            "out_token_info.position"
+          end
+        end
+        code = code.gsub(/\$\{end_position\}/) do |match|
+          case @language
+          when "c"
+            "out_token_info->end_position"
+          when "d"
+            "out_token_info.end_position"
+          end
+        end
         code = code.gsub(/\$mode\(([a-zA-Z_][a-zA-Z_0-9]*)\)/) do |match|
           mode_name = $1
           mode_id = @lexer.mode_id(mode_name)

spec/propane_spec.rb

@@ -910,6 +910,57 @@ EOF
         expect(results.status).to eq 0
       end
 
+      it "allows lexer code blocks to access the matched token positions" do
+        case language
+        when "c", "cpp"
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+>>
+context_user_fields <<
+    p_position_t last_start;
+    p_position_t last_end;
+>>
+drop /\\s+/;
+token word /[a-z]+/ <<
+    ${context.last_start} = ${position};
+    ${context.last_end} = ${end_position};
+>>
+token stop /!/ <<
+    $terminate(42);
+>>
+Start -> Words;
+Words -> ;
+Words -> word Words;
+Words -> stop Words;
+EOF
+        when "d"
+          write_grammar <<EOF
+context_user_fields <<
+    p_position_t last_start;
+    p_position_t last_end;
+>>
+drop /\\s+/;
+token word /[a-z]+/ <<
+    ${context.last_start} = ${position};
+    ${context.last_end} = ${end_position};
+>>
+token stop /!/ <<
+    $terminate(42);
+>>
+Start -> Words;
+Words -> ;
+Words -> word Words;
+Words -> stop Words;
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_lexer_positions.#{language}", language: language)
+        results = run_test(language: language)
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
+
       it "allows the user to terminate the parser" do
         write_grammar <<EOF
 token a;

spec/test_lexer_positions.c

@@ -0,0 +1,50 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+
+int main()
+{
+    char const * input = "abc\n  defg hi\n!";
+    p_context_t * context = p_context_new((uint8_t const *)input, strlen(input));
+    p_token_info_t token_info;
+
+    /* First token "abc" on row 1, cols 1-3. */
+    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(token_info.token == TOKEN_word);
+    assert(context->last_start.row == 1u);
+    assert(context->last_start.col == 1u);
+    assert(context->last_end.row == 1u);
+    assert(context->last_end.col == 3u);
+    /* The lexer code block observed the same positions reported to the caller. */
+    assert(context->last_start.row == token_info.position.row);
+    assert(context->last_start.col == token_info.position.col);
+    assert(context->last_end.row == token_info.end_position.row);
+    assert(context->last_end.col == token_info.end_position.col);
+
+    /* Second token "defg" on row 2, cols 3-6. */
+    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(token_info.token == TOKEN_word);
+    assert(context->last_start.row == 2u);
+    assert(context->last_start.col == 3u);
+    assert(context->last_end.row == 2u);
+    assert(context->last_end.col == 6u);
+
+    /* Third token "hi" on row 2, cols 8-9. */
+    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(token_info.token == TOKEN_word);
+    assert(context->last_start.row == 2u);
+    assert(context->last_start.col == 8u);
+    assert(context->last_end.row == 2u);
+    assert(context->last_end.col == 9u);
+
+    /* The "!" stop token terminates the lexer. The context input text position
+     * must not be updated when the lexer user code requests termination, so it
+     * still points at the "!" token on row 3, col 1. */
+    assert(p_lex(context, &token_info) == P_USER_TERMINATED);
+    assert(p_user_terminate_code(context) == 42u);
+    assert(context->text_position.row == 3u);
+    assert(context->text_position.col == 1u);
+
+    p_context_delete(context);
+    return 0;
+}

spec/test_lexer_positions.d

@@ -0,0 +1,42 @@
+import testparser;
+import std.stdio;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "abc\n  defg hi\n!";
+    p_context_t * context = p_context_new(input);
+    p_token_info_t token_info;
+
+    /* First token "abc" on row 1, cols 1-3. */
+    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(token_info.token == TOKEN_word);
+    assert(context.last_start == p_position_t(1, 1));
+    assert(context.last_end == p_position_t(1, 3));
+    /* The lexer code block observed the same positions reported to the caller. */
+    assert(context.last_start == token_info.position);
+    assert(context.last_end == token_info.end_position);
+
+    /* Second token "defg" on row 2, cols 3-6. */
+    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(token_info.token == TOKEN_word);
+    assert(context.last_start == p_position_t(2, 3));
+    assert(context.last_end == p_position_t(2, 6));
+
+    /* Third token "hi" on row 2, cols 8-9. */
+    assert(p_lex(context, &token_info) == P_SUCCESS);
+    assert(token_info.token == TOKEN_word);
+    assert(context.last_start == p_position_t(2, 8));
+    assert(context.last_end == p_position_t(2, 9));
+
+    /* The "!" stop token terminates the lexer. The context input text position
+     * must not be updated when the lexer user code requests termination, so it
+     * still points at the "!" token on row 3, col 1. */
+    assert(p_lex(context, &token_info) == P_USER_TERMINATED);
+    assert(p_user_terminate_code(context) == 42u);
+    assert(context.text_position == p_position_t(3, 1));
+}