From d88358a88ca7bc5065f69c2f5dbce516ab15d142 Mon Sep 17 00:00:00 2001
From: "Prof. Ronald Moore" <ronald.moore@h-da.de>
Date: Tue, 15 Jun 2021 15:40:37 +0200
Subject: [PATCH] imported comments from pull request from Laura-Marie Henning
-- see
https://code.fbi.h-da.de/compilerconstruction/codesamples/-/merge_requests/1
Signed-off-by: Prof. Ronald Moore <ronald.moore@h-da.de>
---
flexBison/cplusplus/calc++.cc | 5 +++++
flexBison/cplusplus/driver.cc | 8 ++++---
flexBison/cplusplus/driver.hh | 20 ++++++++++++-----
flexBison/cplusplus/parser.yy | 39 ++++++++++++++++++++++++++++------
flexBison/cplusplus/scanner.ll | 14 +++++++++---
5 files changed, 69 insertions(+), 17 deletions(-)
diff --git a/flexBison/cplusplus/calc++.cc b/flexBison/cplusplus/calc++.cc
index 7c8440c..311e4eb 100644
--- a/flexBison/cplusplus/calc++.cc
+++ b/flexBison/cplusplus/calc++.cc
@@ -6,6 +6,8 @@
int
main (int argc, char *argv[])
{
+ // we build an interpreter,
+ // hence we return an int as the result of the expressions we evaluate
int res = 0;
driver drv;
for (int i = 1; i < argc; ++i)
@@ -14,8 +16,11 @@ main (int argc, char *argv[])
else if (argv[i] == std::string ("-s"))
drv.trace_scanning = true;
else if (!drv.parse (argv[i]))
+ // parse() returns 1 if it ran into an error
+ // so with !drv.parse we are here if everything was fine
std::cout << drv.result << '\n';
else
+ // error while parsing
res = 1;
return res;
}
diff --git a/flexBison/cplusplus/driver.cc b/flexBison/cplusplus/driver.cc
index fa0450b..5f5e104 100644
--- a/flexBison/cplusplus/driver.cc
+++ b/flexBison/cplusplus/driver.cc
@@ -5,6 +5,8 @@
driver::driver ()
: trace_parsing (false), trace_scanning (false)
{
+ // These are hardcoded variables that are now available inside expressions you write
+ // e.g. 'y = one + 1'
variables["one"] = 1;
variables["two"] = 2;
}
@@ -13,11 +15,11 @@ int
driver::parse (const std::string &f)
{
file = f;
- location.initialize (&file);
+ location.initialize (&file); // set location to beginning of file or stdin
scan_begin ();
- yy::parser parse (*this);
+ yy::parser parse (*this); // set the parsing context to *this (remember, we are inside driver.cc class)
parse.set_debug_level (trace_parsing);
- int res = parse ();
+ int res = parse (); //this evaluates the file (or stdin, see scan_begin() definition at the end of scanner.ll)
scan_end ();
return res;
}
diff --git a/flexBison/cplusplus/driver.hh b/flexBison/cplusplus/driver.hh
index 92a0e31..8137cc3 100644
--- a/flexBison/cplusplus/driver.hh
+++ b/flexBison/cplusplus/driver.hh
@@ -5,23 +5,30 @@
# include <map>
# include "parser.hh"
-// Give Flex the prototype of yylex we want ...
+// Give Flex the prototype of yylex we want:
+// because of %define api.value.type variant and %define api.token.constructor in parser.yy
+// the parser defines the type symbol_type, and expects yylex to have the following prototype.
# define YY_DECL \
yy::parser::symbol_type yylex (driver& drv)
// ... and declare it for the parser's sake.
+// yylex now returns a complete symbol, aggregating its type (i.e., the traditional value returned by yylex),
+// its semantic value, and possibly its location.
+// see https://www.gnu.org/software/bison/manual/html_node/Complete-Symbols.html
YY_DECL;
// Conducting the whole scanning and parsing of Calc++.
+// This is the parsing context, an interface to the parser and scanner
+// Contains all the data to exchange -> opens file to scan, instantiates the parser etc.
class driver
{
public:
driver ();
- std::map<std::string, int> variables;
+ std::map<std::string, int> variables; // this holds the variables and their values we find on the way
- int result;
+ int result; // the evaluated expession result
- // Run the parser on file F. Return 0 on success.
+ // Run the parser on file F. Return 0 on success and 1 on failure.
int parse (const std::string& f);
// The name of the file being parsed.
std::string file;
@@ -29,11 +36,14 @@ public:
bool trace_parsing;
// Handling the scanner.
+ // These functions are declared at the end of scanner.ll
void scan_begin ();
void scan_end ();
+
// Whether to generate scanner debug traces.
bool trace_scanning;
- // The token's location used by the scanner.
+ // The token's location used by the scanner.
+ // made available by 'yy::location& loc = drv.location;' in scanner.ll
yy::location location;
};
#endif // ! DRIVER_HH
diff --git a/flexBison/cplusplus/parser.yy b/flexBison/cplusplus/parser.yy
index 188873b..a517f96 100644
--- a/flexBison/cplusplus/parser.yy
+++ b/flexBison/cplusplus/parser.yy
@@ -1,20 +1,30 @@
// See https://www.gnu.org/software/bison/manual/html_node/Calc_002b_002b-Parser.html
+// needed for creation of parser.hh (C++ style)
%skeleton "lalr1.cc" /* -*- C++ -*- */
%require "3.5"
%defines
+// In C we would define a %union to use as our symbol type
+// In C++ and newer bison we can tell it to use a std::variant for a symbol type
+// We declare our intention to use is with
%define api.token.constructor
%define api.value.type variant
%define parse.assert
+// and then just use whatever types we want. 'parse.assert' makes sure we use the right types in symbols
%code requires {
+ // place your includes here
# include <string>
class driver;
}
-// The parsing context.
+// The parsing context is passed to flex and bison.
+// Equivalent to both %lex-param and %parse-param
+// You could also define 'int yylex (driver& drv)' and 'int yyparse (driver& drv);'
%param { driver& drv }
+//request location tracking
+// made available by 'yy::location& loc = drv.location;' in scanner.ll
%locations
%define parse.trace
@@ -24,6 +34,7 @@
# include "driver.hh"
}
+// DRY - tokens shall begin with "TOK_"
%define api.token.prefix {TOK_}
%token
END 0 "end of file"
@@ -36,29 +47,45 @@
RPAREN ")"
;
+// these are our variant types
+// bison generates functions like
+// 'symbol_type make_NUMBER (const std::string&, const location_type&);' for us, which can be overwritten
%token <std::string> IDENTIFIER "identifier"
%token <int> NUMBER "number"
+// declare nonterminal expression
+// compare to %type: https://www.gnu.org/software/bison/manual/html_node/Type-Decl.html
+// tl;dr: use nterm explicitly for nonterminals, type can also be a terminal
+// int: an exp resolves/ is evaluated to an int
+// replace e.g. with AST node types
%nterm <int> exp
+// unit etc. do not need to be declared here, because they have no type to be evaluated to,
+// as they are evaluated to other nterms
-%printer { yyo << $$; } <*>;
+%printer { yyo << $$; } <*>; //print values using their operator<<
%%
-%start unit;
-unit: assignments exp { drv.result = $2; };
+%start unit; // this is our root node, so to speak
+unit: assignments exp { drv.result = $2; };
+// here we set the final outcome of exp to the result of our driver
+// sort of to keep it when the compiling is done
+// imagine you could also put a pointer to the root node of you ast here :)
assignments:
%empty {}
| assignments assignment {};
assignment:
- "identifier" ":=" exp { drv.variables[$1] = $3; };
+ "identifier" ":=" exp { drv.variables[$1] = $3; };
+ // we save our variable inside the driver because we want to store it for later use
+// you will find this everytime mathematical expressions are used
+// it declares operator precedence; see here https://www.gnu.org/software/bison/manual/html_node/Precedence-Decl.html
%left "+" "-";
%left "*" "/";
exp:
"number"
| "identifier" { $$ = drv.variables[$1]; }
-| exp "+" exp { $$ = $1 + $3; }
+| exp "+" exp { $$ = $1 + $3; } // these are real mathematical expressions being evaluated and stored back in exp ($$)
| exp "-" exp { $$ = $1 - $3; }
| exp "*" exp { $$ = $1 * $3; }
| exp "/" exp { $$ = $1 / $3; }
diff --git a/flexBison/cplusplus/scanner.ll b/flexBison/cplusplus/scanner.ll
index 9cf9417..39d2ef7 100644
--- a/flexBison/cplusplus/scanner.ll
+++ b/flexBison/cplusplus/scanner.ll
@@ -78,6 +78,9 @@
// A number symbol corresponding to the value in S.
yy::parser::symbol_type
make_NUMBER (const std::string &s, const yy::parser::location_type& loc);
+ // Note: the symbol_type is definded in parser.yy by declaring
+ // %define api.token.constructor and following lines. See explanation there or
+ // https://www.gnu.org/software/bison/manual/html_node/Complete-Symbols.html
%}
id [a-zA-Z][a-zA-Z_0-9]*
@@ -86,13 +89,17 @@ blank [ \t\r]
%{
// Code run each time a pattern is matched.
+ // The macro YY_USER_ACTION can be defined to provide an action which is always executed prior to the matched rule's action.
+ // http://dinosaur.compilertools.net/flex/flex_14.html
+ // yyleng is a lex global variable that is always the length of the token you just read in
# define YY_USER_ACTION loc.columns (yyleng);
%}
%%
%{
// A handy shortcut to the location held by the driver.
yy::location& loc = drv.location;
- // Code run each time yylex is called.
+ // Code run each time yylex is called.
+ // The lexical analyzer function, yylex, recognizes tokens from the input stream and returns them to the parser.
loc.step ();
%}
{blank}+ loc.step ();
@@ -106,8 +113,8 @@ blank [ \t\r]
")" return yy::parser::make_RPAREN (loc);
":=" return yy::parser::make_ASSIGN (loc);
-{int} return make_NUMBER (yytext, loc);
-{id} return yy::parser::make_IDENTIFIER (yytext, loc);
+{int} return make_NUMBER (yytext, loc); // make_NUMBER is defined below and overwrites the bison generated one
+{id} return yy::parser::make_IDENTIFIER (yytext, loc); // the function is generated by bison
. {
throw yy::parser::syntax_error
(loc, "invalid character: " + std::string(yytext));
@@ -115,6 +122,7 @@ blank [ \t\r]
<<EOF>> return yy::parser::make_END (loc);
%%
+// overwrite the bison generated make_NUMBER function, because we want to add some functionality
yy::parser::symbol_type
make_NUMBER (const std::string &s, const yy::parser::location_type& loc)
{
--
GitLab