Only calculate lookahead tokens when needed - #28

Lookahead tokens are only need if either:
(1) There is more than one rule that could be reduced in a given parser
state, or
(2) There are shift actions for a state and at least one rule that could
be reduced in the same state (to warn about shift/reduce conflicts).

CHANGELOG.md

@@ -2,7 +2,14 @@
 
 ### New Features
 
-- Track token position in AST Token node
+- Track start and end text positions for tokens and rules in AST node structures (#27)
+- Add warnings for shift/reduce conflicts to log file (#25)
+- Add -w command line switch to treat warnings as errors and output to stderr (#26)
+- Add rule field aliases (#24)
+
+### Improvements
+
+- Show line numbers of rules on conflict (#23)
 
 ## v1.4.0
 

README.md

@@ -31,9 +31,14 @@ Propane is typically invoked from the command-line as `./propane`.
 
     Usage: ./propane [options] <input-file> <output-file>
     Options:
-      --log LOG   Write log file
-      --version   Show program version and exit
-      -h, --help  Show this usage and exit
+      -h, --help  Show this usage and exit.
+      --log LOG   Write log file. This will show all parser states and their
+                  associated shifts and reduces. It can be helpful when
+                  debugging a grammar.
+      --version   Show program version and exit.
+      -w          Treat warnings as errors. This option will treat shift/reduce
+                  conflicts as fatal errors and will print them to stderr in
+                  addition to the log file.
 
 The user must specify the path to a Propane input grammar file and a path to an
 output file.

doc/user_guide.md

@@ -234,15 +234,15 @@ drop /\\s+/;
 
 Start -> Items;
 
-Items -> Item ItemsMore;
+Items -> Item:item ItemsMore;
 Items -> ;
 
-ItemsMore -> comma Item ItemsMore;
+ItemsMore -> comma Item:item ItemsMore;
 ItemsMore -> ;
 
 Item -> a;
 Item -> b;
-Item -> lparen Item rparen;
+Item -> lparen Item:item rparen;
 Item -> Dual;
 
 Dual -> One Two;
@@ -263,24 +263,24 @@ Start * start = p_result(&context);
 assert(start.pItems1 !is null);
 assert(start.pItems !is null);
 Items * items = start.pItems;
-assert(items.pItem !is null);
-assert(items.pItem.pToken1 !is null);
-assert_eq(TOKEN_a, items.pItem.pToken1.token);
-assert_eq(11, items.pItem.pToken1.pvalue);
+assert(items.item !is null);
+assert(items.item.pToken1 !is null);
+assert_eq(TOKEN_a, items.item.pToken1.token);
+assert_eq(11, items.item.pToken1.pvalue);
 assert(items.pItemsMore !is null);
 ItemsMore * itemsmore = items.pItemsMore;
-assert(itemsmore.pItem !is null);
-assert(itemsmore.pItem.pItem !is null);
-assert(itemsmore.pItem.pItem.pItem !is null);
-assert(itemsmore.pItem.pItem.pItem.pToken1 !is null);
-assert_eq(TOKEN_b, itemsmore.pItem.pItem.pItem.pToken1.token);
-assert_eq(22, itemsmore.pItem.pItem.pItem.pToken1.pvalue);
+assert(itemsmore.item !is null);
+assert(itemsmore.item.item !is null);
+assert(itemsmore.item.item.item !is null);
+assert(itemsmore.item.item.item.pToken1 !is null);
+assert_eq(TOKEN_b, itemsmore.item.item.item.pToken1.token);
+assert_eq(22, itemsmore.item.item.item.pToken1.pvalue);
 assert(itemsmore.pItemsMore !is null);
 itemsmore = itemsmore.pItemsMore;
-assert(itemsmore.pItem !is null);
-assert(itemsmore.pItem.pToken1 !is null);
-assert_eq(TOKEN_b, itemsmore.pItem.pToken1.token);
-assert_eq(22, itemsmore.pItem.pToken1.pvalue);
+assert(itemsmore.item !is null);
+assert(itemsmore.item.pToken1 !is null);
+assert_eq(TOKEN_b, itemsmore.item.pToken1.token);
+assert_eq(22, itemsmore.item.pToken1.pvalue);
 assert(itemsmore.pItemsMore is null);
 ```
 
@@ -607,6 +607,10 @@ This can be changed with the `start` statement.
 The grammar file must define a rule with the name of the start rule name which
 will be used as the top-level starting rule that the parser attempts to reduce.
 
+Rule statements are composed of the name of the rule, a `->` token, the fields
+defining the rule pattern that must be matched, and a terminating semicolon or
+user code block.
+
 Example:
 
 ```
@@ -635,9 +639,13 @@ E4 -> lparen E1 rparen << $$ = $2; >>
 
 This example uses the default start rule name of `Start`.
 
-A parser rule has zero or more terms on the right side of its definition.
-Each of these terms is either a token name or a rule name.
-A term can be immediately followed by a `?` character to signify that it is
+A parser rule has zero or more fields on the right side of its definition.
+Each of these fields is either a token name or a rule name.
+A field can optionally be followed by a `:` and then a field alias name.
+If present, the field alias name is used to refer to the field value in user
+code blocks, or if AST mode is active, the field alias name is used as the
+field name in the generated AST node structure.
+A field can be immediately followed by a `?` character to signify that it is
 optional.
 Another example:
 
@@ -647,14 +655,16 @@ token private;
 token int;
 token ident /[a-zA-Z_][a-zA-Z_0-9]*/;
 token semicolon /;/;
-IntegerDeclaration -> Visibility? int ident semicolon;
+IntegerDeclaration -> Visibility? int ident:name semicolon;
 Visibility -> public;
 Visibility -> private;
 ```
 
-In a parser rule code block, parser values for the right side terms are
-accessible as `$1` for the first term's parser value, `$2` for the second
-term's parser value, etc...
+In a parser rule code block, parser values for the right side fields are
+accessible as `$1` for the first field's parser value, `$2` for the second
+field's parser value, etc...
+For the `IntegerDeclaration` rule, the third field value can also be referred
+to as `${name}`.
 The `$$` symbol accesses the output parser value for this rule.
 The above examples demonstrate how the parser values for the rule components
 can be used to produce the parser value for the accepted rule.
@@ -849,6 +859,19 @@ If the first rule is matched, then `pOne1` and `pTwo2` will be non-null while
 `pTwo1` and `pOne2` will be null.
 If the second rule is matched instead, then the opposite would be the case.
 
+If a field alias is present in a rule definition, an additional field will be
+generated in the AST node with the field alias name.
+For example:
+
+```
+Exp -> Exp:left plus ExpB:right;
+```
+
+In the generated `Exp` structure, the fields `pExp`, `pExp1`, and `left` will
+all point to the same child node (an instance of the `Exp` structure), and the
+fields `pExpB`, `pExpB3`, and `right` will all point to the same child node
+(an instance of the `ExpB` structure).
+
 ##> Functions
 
 ### `p_context_init`

extra/vim/syntax/propane.vim

@@ -17,6 +17,9 @@ syn region propaneTarget matchgroup=propaneDelimiter start="<<" end=">>$" contai
 
 syn match propaneComment "#.*"
 syn match propaneOperator "->"
+syn match propaneFieldAlias ":[a-zA-Z0-9_]\+" contains=propaneFieldOperator
+syn match propaneFieldOperator ":" contained
+syn match propaneOperator "?"
 syn keyword propaneKeyword ast ast_prefix ast_suffix drop module prefix ptype start token tokenid
 
 syn region propaneRegex start="/" end="/" skip="\\/"
@@ -25,4 +28,6 @@ hi def link propaneComment Comment
 hi def link propaneKeyword Keyword
 hi def link propaneRegex String
 hi def link propaneOperator Operator
+hi def link propaneFieldOperator Operator
 hi def link propaneDelimiter Delimiter
+hi def link propaneFieldAlias Identifier

lib/propane/generator.rb

@@ -276,6 +276,19 @@ class Propane
             "statevalues[$-1-n_states+#{index}].pvalue.v_#{rule.components[index - 1].ptypename}"
           end
         end
+        code = code.gsub(/\$\{(\w+)\}/) do |match|
+          aliasname = $1
+          if index = rule.aliases[aliasname]
+            case @language
+            when "c"
+              "state_values_stack_index(statevalues, -(int)n_states + #{index})->pvalue.v_#{rule.components[index].ptypename}"
+            when "d"
+              "statevalues[$-n_states+#{index}].pvalue.v_#{rule.components[index].ptypename}"
+            end
+          else
+            raise Error.new("Field alias '#{aliasname}' not found")
+          end
+        end
       else
         code = code.gsub(/\$\$/) do |match|
           if @grammar.ast

lib/propane/grammar.rb

@@ -198,7 +198,7 @@ class Propane
         if @ast && ptypename
           raise Error.new("Multiple ptypes are unsupported in AST mode")
         end
-        md = consume!(/((?:#{IDENTIFIER_REGEX}\??\s*)*)\s*/, "expected rule component list")
+        md = consume!(/((?:#{IDENTIFIER_REGEX}(?::#{IDENTIFIER_REGEX})?\??\s*)*)\s*/, "expected rule component list")
         components = md[1].strip.split(/\s+/)
         if @ast
           consume!(/;/, "expected `;'")

lib/propane/parser.rb

@@ -39,7 +39,6 @@ class Propane
       end
 
       build_reduce_actions!
-      build_follow_sets!
       build_tables!
       write_log!
       if @warnings.size > 0 && @options[:warnings_as_errors]
@@ -66,10 +65,10 @@ class Propane
             state_id: state_id,
           }
         end
-        if item_set.reduce_actions
+        unless item_set.reduce_rules.empty?
           shift_entries.each do |shift_entry|
             token = shift_entry[:symbol]
-            if item_set.reduce_actions.include?(token)
+            if get_lookahead_reduce_actions_for_item_set(item_set).include?(token)
               rule = item_set.reduce_actions[token]
               @warnings << "Shift/Reduce conflict (state #{item_set.id}) between token #{token.name} and rule #{rule.name} (defined on line #{rule.line_number})"
             end
@@ -115,7 +114,7 @@ class Propane
     # @return [void]
     def build_reduce_actions!
       @item_sets.each do |item_set|
-        item_set.reduce_actions = build_reduce_actions_for_item_set(item_set)
+        build_reduce_actions_for_item_set(item_set)
       end
     end
 
@@ -124,27 +123,36 @@ class Propane
     # @param item_set [ItemSet]
     #   ItemSet (parser state)
     #
-    # @return [nil, Hash]
-    #   If no reduce actions are possible for the given item set, nil.
-    #   Otherwise, a mapping of lookahead Tokens to the Rules to reduce.
+    # @return [void]
     def build_reduce_actions_for_item_set(item_set)
       # To build the reduce actions, we start by looking at any
       # "complete" items, i.e., items where the parse position is at the
       # end of a rule. These are the only rules that are candidates for
       # reduction in the current ItemSet.
-      reduce_rules = Set.new(item_set.items.select(&:complete?).map(&:rule))
+      item_set.reduce_rules = Set.new(item_set.items.select(&:complete?).map(&:rule))
 
-      if reduce_rules.size == 1
-        item_set.reduce_rule = reduce_rules.first
+      if item_set.reduce_rules.size == 1
+        item_set.reduce_rule = item_set.reduce_rules.first
       end
 
-      if reduce_rules.size == 0
-        nil
-      else
-        build_lookahead_reduce_actions_for_item_set(item_set)
+      if item_set.reduce_rules.size > 1
+        # Force item_set.reduce_actions to be built to store the lookahead
+        # tokens for the possible reduce rules if there is more than one.
+        get_lookahead_reduce_actions_for_item_set(item_set)
       end
     end
 
+    # Get the reduce actions for a single item set (parser state).
+    #
+    # @param item_set [ItemSet]
+    #   ItemSet (parser state)
+    #
+    # @return [Hash]
+    #   Mapping of lookahead Tokens to the Rules to reduce.
+    def get_lookahead_reduce_actions_for_item_set(item_set)
+      item_set.reduce_actions ||= build_lookahead_reduce_actions_for_item_set(item_set)
+    end
+
     # Build the reduce actions for a single item set (parser state).
     #
     # @param item_set [ItemSet]
@@ -153,15 +161,13 @@ class Propane
     # @return [Hash]
     #   Mapping of lookahead Tokens to the Rules to reduce.
     def build_lookahead_reduce_actions_for_item_set(item_set)
-      reduce_rules = Set.new(item_set.items.select(&:complete?).map(&:rule))
-
       # We will be looking for all possible tokens that can follow instances of
       # these rules. Rather than looking through the entire grammar for the
       # possible following tokens, we will only look in the item sets leading
       # up to this one. This restriction gives us a more precise lookahead set,
       # and allows us to parse LALR grammars.
       item_sets = Set[item_set] + item_set.leading_item_sets
-      reduce_rules.reduce({}) do |reduce_actions, reduce_rule|
+      item_set.reduce_rules.reduce({}) do |reduce_actions, reduce_rule|
         lookahead_tokens_for_rule = build_lookahead_tokens_to_reduce(reduce_rule, item_sets)
         lookahead_tokens_for_rule.each do |lookahead_token|
           if existing_reduce_rule = reduce_actions[lookahead_token]
@@ -233,51 +239,6 @@ class Propane
       lookahead_tokens
     end
 
-    # Build the follow sets for each ItemSet.
-    #
-    # @return [void]
-    def build_follow_sets!
-      @item_sets.each do |item_set|
-        item_set.follow_set = build_follow_set_for_item_set(item_set)
-      end
-    end
-
-    # Build the follow set for the given ItemSet.
-    #
-    # @param item_set [ItemSet]
-    #   The ItemSet to build the follow set for.
-    #
-    # @return [Set]
-    #   Follow set for the given ItemSet.
-    def build_follow_set_for_item_set(item_set)
-      follow_set = Set.new
-      rule_sets_to_check_after = Set.new
-      item_set.items.each do |item|
-        (1..).each do |offset|
-          case symbol = item.next_symbol(offset)
-          when nil
-            rule_sets_to_check_after << item.rule.rule_set
-            break
-          when Token
-            follow_set << symbol
-            break
-          when RuleSet
-            follow_set += symbol.start_token_set
-            unless symbol.could_be_empty?
-              break
-            end
-          end
-        end
-      end
-      reduce_lookaheads = build_lookahead_reduce_actions_for_item_set(item_set)
-      reduce_lookaheads.each do |token, rule_set|
-        if rule_sets_to_check_after.include?(rule_set)
-          follow_set << token
-        end
-      end
-      follow_set
-    end
-
     def write_log!
       @log.puts Util.banner("Parser Rules")
       @grammar.rules.each do |rule|

lib/propane/parser/item_set.rb

@@ -2,7 +2,7 @@ class Propane
   class Parser
 
     # Represent a parser "item set", which is a set of possible items that the
-    # parser could currently be parsing.
+    # parser could currently be parsing. This is equivalent to a parser state.
     class ItemSet
 
       # @return [Set<Item>]
@@ -25,15 +25,15 @@ class Propane
       #   Rule to reduce if there is only one possibility.
       attr_accessor :reduce_rule
 
+      # @return [Set<Rule>]
+      #   Set of rules that could be reduced in this parser state.
+      attr_accessor :reduce_rules
+
       # @return [nil, Hash]
       #   Reduce actions, mapping lookahead tokens to rules, if there is
       #   more than one rule that could be reduced.
       attr_accessor :reduce_actions
 
-      # @return [Set<Token>]
-      #   Follow set for the ItemSet.
-      attr_accessor :follow_set
-
       # Build an ItemSet.
       #
       # @param items [Array<Item>]
@@ -50,7 +50,7 @@ class Propane
       # @return [Set<Token, RuleSet>]
       #   Set of next symbols for all Items in this ItemSet.
       def next_symbols
-        Set.new(@items.map(&:next_symbol).compact)
+        @_next_symbols ||= Set.new(@items.map(&:next_symbol).compact)
       end
 
       # Build a next ItemSet for the given next symbol.
@@ -99,6 +99,8 @@ class Propane
       # @return [Set<ItemSet>]
       #   Set of all ItemSets that lead up to this ItemSet.
       def leading_item_sets
+        @_leading_item_sets ||=
+          begin
             result = Set.new
             eval_sets = Set[self]
             evaled = Set.new
@@ -115,6 +117,7 @@ class Propane
             end
             result
           end
+      end
 
       # Represent the ItemSet as a String.
       #

lib/propane/rule.rb

@@ -6,6 +6,10 @@ class Propane
     #   Rule components.
     attr_reader :components
 
+    # @return [Hash]
+    #   Field aliases.
+    attr_reader :aliases
+
     # @return [String]
     #   User code associated with the rule.
     attr_reader :code
@@ -49,7 +53,19 @@ class Propane
     #   Line number where the rule was defined in the input grammar.
     def initialize(name, components, code, ptypename, line_number)
       @name = name
-      @components = components
+      @aliases = {}
+      @components = components.each_with_index.map do |component, i|
+        if component =~ /(\S+):(\S+)/
+          c, aliasname = $1, $2
+          if @aliases[aliasname]
+            raise Error.new("Error: duplicate field alias `#{aliasname}` for rule #{name} defined on line #{line_number}")
+          end
+          @aliases[aliasname] = i
+          c
+        else
+          component
+        end
+      end
       @rule_set_node_field_index_map = components.map {0}
       @code = code
       @ptypename = ptypename

lib/propane/rule_set.rb

@@ -100,8 +100,10 @@ class Propane
 
     # Finalize a RuleSet after adding all Rules to it.
     def finalize(grammar)
+      if grammar.ast
         build_ast_fields(grammar)
       end
+    end
 
     private
 
@@ -148,6 +150,18 @@ class Propane
             "#{grammar.ast_prefix}#{node_name}#{grammar.ast_suffix}"
         end
       end
+      # Now merge in the field aliases as given by the user in the
+      # grammar.
+      field_aliases = {}
+      @rules.each do |rule|
+        rule.aliases.each do |alias_name, index|
+          if field_aliases[alias_name] && field_aliases[alias_name] != index
+            raise Error.new("Error: conflicting AST node field positions for alias `#{alias_name}`")
+          end
+          field_aliases[alias_name] = index
+          @ast_fields[index][alias_name] = @ast_fields[index].first[1]
+        end
+      end
     end
 
   end

lib/propane/version.rb

@@ -1,3 +1,3 @@
 class Propane
-  VERSION = "1.4.0"
+  VERSION = "1.5.0"
 end

spec/propane_spec.rb

@@ -213,6 +213,42 @@ EOF
     expect(File.binread("spec/run/testparser.log")).to match %r{Shift/Reduce conflict \(state \d+\) between token b and rule As2\? \(defined on line 4\)}
   end
 
+  it "errors on duplicate field aliases in a rule" do
+    write_grammar <<EOF
+token a;
+token b;
+Start -> a:foo b:foo;
+EOF
+    results = run_propane(extra_args: %w[-w], capture: true)
+    expect(results.stderr).to match %r{Error: duplicate field alias `foo` for rule Start defined on line 3}
+    expect(results.status).to_not eq 0
+  end
+
+  it "errors when an alias is in different positions for different rules in a rule set when AST mode is enabled" do
+    write_grammar <<EOF
+ast;
+token a;
+token b;
+Start -> a:foo b;
+Start -> b b:foo;
+EOF
+    results = run_propane(extra_args: %w[-w], capture: true)
+    expect(results.stderr).to match %r{Error: conflicting AST node field positions for alias `foo`}
+    expect(results.status).to_not eq 0
+  end
+
+  it "does not error when an alias is in different positions for different rules in a rule set when AST mode is not enabled" do
+    write_grammar <<EOF
+token a;
+token b;
+Start -> a:foo b;
+Start -> b b:foo;
+EOF
+    results = run_propane(extra_args: %w[-w], capture: true)
+    expect(results.stderr).to eq ""
+    expect(results.status).to eq 0
+  end
+
   %w[d c].each do |language|
 
     context "#{language.upcase} language" do
@@ -1120,6 +1156,70 @@ EOF
         expect(results.stderr).to eq ""
         expect(results.status).to eq 0
       end
+
+      it "allows specifying field aliases in AST mode" do
+        write_grammar <<EOF
+ast;
+
+token a;
+token b;
+token c;
+drop /\\s+/;
+Start -> T:first T:second T:third;
+T -> a;
+T -> b;
+T -> c;
+EOF
+        run_propane(language: language)
+        compile("spec/test_ast_field_aliases.#{language}", language: language)
+        results = run_test
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+      end
+
+      it "allows specifying field aliases when AST mode is not enabled" do
+        if language == "d"
+          write_grammar <<EOF
+<<
+import std.stdio;
+>>
+ptype string;
+token id /[a-zA-Z_][a-zA-Z0-9_]*/ <<
+  $$ = match;
+>>
+drop /\\s+/;
+Start -> id:first id:second <<
+  writeln("first is ", ${first});
+  writeln("second is ", ${second});
+>>
+EOF
+        else
+          write_grammar <<EOF
+<<
+#include <stdio.h>
+#include <string.h>
+>>
+ptype char const *;
+token id /[a-zA-Z_][a-zA-Z0-9_]*/ <<
+  char * s = malloc(match_length + 1);
+  strncpy(s, (char const *)match, match_length);
+  s[match_length] = 0;
+  $$ = s;
+>>
+drop /\\s+/;
+Start -> id:first id:second <<
+  printf("first is %s\\n", ${first});
+  printf("second is %s\\n", ${second});
+>>
+EOF
+        end
+        run_propane(language: language)
+        compile("spec/test_field_aliases.#{language}", language: language)
+        results = run_test
+        expect(results.stderr).to eq ""
+        expect(results.status).to eq 0
+        expect(results.stdout).to match /first is foo1.*second is bar2/m
+      end
     end
   end
 end

spec/test_ast_field_aliases.c

@@ -0,0 +1,19 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "\na\nb\nc";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(&context);
+
+    assert_eq(TOKEN_a, start->first->pToken->token);
+    assert_eq(TOKEN_b, start->second->pToken->token);
+    assert_eq(TOKEN_c, start->third->pToken->token);
+
+    return 0;
+}

spec/test_ast_field_aliases.d

@@ -0,0 +1,21 @@
+import testparser;
+import std.stdio;
+import testutils;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "\na\nb\nc";
+    p_context_t context;
+    p_context_init(&context, input);
+    assert(p_parse(&context) == P_SUCCESS);
+    Start * start = p_result(&context);
+
+    assert_eq(TOKEN_a, start.first.pToken.token);
+    assert_eq(TOKEN_b, start.second.pToken.token);
+    assert_eq(TOKEN_c, start.third.pToken.token);
+}

spec/test_field_aliases.c

@@ -0,0 +1,13 @@
+#include "testparser.h"
+#include <assert.h>
+#include <string.h>
+#include "testutils.h"
+
+int main()
+{
+    char const * input = "foo1\nbar2";
+    p_context_t context;
+    p_context_init(&context, (uint8_t const *)input, strlen(input));
+    assert(p_parse(&context) == P_SUCCESS);
+    return 0;
+}

spec/test_field_aliases.d

@@ -0,0 +1,15 @@
+import testparser;
+import std.stdio;
+
+int main()
+{
+    return 0;
+}
+
+unittest
+{
+    string input = "foo1\nbar2";
+    p_context_t context;
+    p_context_init(&context, input);
+    assert(p_parse(&context) == P_SUCCESS);
+}