I'm looking for a parser generator for a reasonably complex language (similar in complexity to Python itself) which works with Python3. If it can generate an AST automatically, this would be a bonus, but I'm fine if it just calls rules while parsing. I have no special requirements, nor does it have to be very efficient/fast.
LEPL isn't exactly a parser generator - it's better! The parsers are defined in Python code and constructed at runtime (hence some inefficiency, but much easier to use). It uses operator overloading to construct a quite readable DSL. Things like c = a & b | b & c for the BNF c := a b | b c..
You can pass the results of a (sub-)parser to an abritary callable, and this is very usable for AST generation (also useful for converting e.g. number literals to Python-level number objects). It's a recursive descent parser, so you better avoid left recursion in the grammar (there are memoization objets that can make left recursion work, but "Lepl's support for them has historically been unreliable (buggy)").
ANTLR can generate a lexer and/or parser in Python. You can also use it to create AST's and iterator-like structures to walk the AST (called tree grammars).
See ANTLR get and split lexer content for an ANTLR demo that produces an AST with the Python target.
Related
It is often believed that functional programming languages such as Ocaml or F# have a type system allowing us to spend more time in writing code, and less time in debugging, compared with coding in dynamic languages like Python or JavaScript.
Now I am writing a parser specification to be used with FsLexYacc, the F# Lexer and Parser package. The parser specification has something like this, for parsing integers and identifier names:
%token <int> CSTINT
%token <string> NAME
...
Expr:
NAME {VAR $1}
|CSTINT {CSTI $1}
Such kind of code is no more written in function programming languages, and therefore, they are not "protected" by the type system any more. Strange bugs can easily slip in I guess (not quite sure though).
Question: Is there any work (theoretical or practical) that tries to address this issue by providing a kind of type system also for the parser/lexer specification?
Yes, there has been a certain amount of research into this area.
It should be noted that although the semantic value stack used in LR parsing algorithms has heterogenous types, the algorithm itself is typesafe. But of course there could be mistakes in the implementation of the algorithm, so it would be ideal if the compiler itself could verify that the code produced by the parser generator is correctly typed. That turns out to be possible, and a number of implementations have been produced.
I don't have a complete bibliography handy, but I do have at hand two papers both published in 2006, which seem relevant:
Towards Efficient, Typed LR Parsers by François Pottier and Yann Régis-Gianas:
The LR parser generators that are bundled with many functional programming language implementations produce code that is untyped, needlessly inefficient, or both. We show that, using generalized algebraic data types, it is possible to produce parsers that are well-typed (so they cannot unexpectedly crash or fail) and nevertheless efficient.
Derivation of a Typed Functional LR Parser by Ralf Hinze and Ross Paterson:
This paper describes a purely functional implementation of LR parsing.
We formally derive our parsers in a series of steps starting from the inverse of printing. In contrast to traditional implementations of LR parsing, the resulting parsers are fully typed, stackless and table-free. The parsing functions pursue alternatives in parallel with each alternative represented by a continuation argument. The direct implementation presents many opportunities for optimization and initial measurements show excellent performance in comparison with conventional table-driven parsers.
You could define your parser using FParsec. This way you'll have all the benefits of using the F# type system. The above could be described as follows using FParsec in F#:
type Expression = StringExpr of String | NumberExpr of int
let alphanumericString = letter .>>. (manyChars (letter <|> digit))
>>= (fun (c,s)-> preturn (StringExpr(c.ToString()+s)))
let number = manyChars digit
>>= (fun n -> preturn (NumberExpr (Int32.Parse n)))
let expression = alphanumericString <|> number
(Background: Inspired by Is C++ context-free or context-sensitive?, while I am writing a simple compiler using jflex/cup myself. )
If they are written using a lexer/parser generator, how do we specify the grammar?
Since code like
a b(c);
could be interpreted as either a function declaration or a local variable definition, how could we handle it in the grammar definition file?
Another example could be the token ">>" in the following code:
std::vector<std::vector<int>> foo;
int a = 1000 >> 4;
Thanks
Are the compilers of C++ written using a lexer/parser generator?
It depends. Some are, some aren't.
GCC originally did use GNU bison, but was re-written a couple of years ago with a hand-written parser. If I have understood that correctly, the main reason was that writing the parser by hand gives you more control over the parser state, and specifically, how much "extraneous" data to keep in there, so that you can generate better error messages.
If they are written using a lexer/parser generator, how do we specify the grammar?
This depends on which parser generator you are using.
Since code like
a b(c);
could be interpreted as either a function declaration or a local variable definition, how could we handle it in the grammar definition file?
Some parser generators may be powerful enough to handle this directly.
Some aren't. Some parser generators which aren't powerful enough have a concept of semantic action that allow you to attach code written in an arbitrarily powerful language to parser rules. E.g. yacc allows you to attach C code to rules.
Otherwise, you will have to handle it during semantic analysis.
I'm looking for a mature parser library, either for Scala or Haskell.
The most important point is, that the library can handle ambiguity.
If an expression is ambiguous, I want every possible abstract syntax tree, that matches the expression.
Simple example: The expression a ⊗ b ⊗ c can be seen as (a ⊗ b) ⊗ c or a ⊗ (b ⊗ c), and I need both variants.
Thanks!
I feel like the old guy for remembering when Walder's papers like Comprehending Monads (the precursor to the do notation) were exciting and new. The idea is that you (to quote) replace a failure by a list of successes, meaning maintain a list of all the possible parses. At the end you normally just take the first match, but with this setup, you can take all of them.
These aren't all that efficient for a deterministic parser, which is why they're less in fashion, but they are what you need.
Have a look at polyparse, and in particular Text.ParserCombinators.HuttonMeijer and Text.ParserCombinators.HuttonMeijerWallace.
(Hutton & Meijer translated the parser library to Haskell (from Gofer) and Wallace added extra features.)
Make sure you check it out on simple cases like parsing "aaaa" with
testP = do
a <- many $ char 'a'
b <- many $ char 'a'
return (a,b)
to see if it has the semantics you seek.
You asked for mature. These libraries are part of pure functional programming's heritage! Having said that, I'd call parsec more mature, even though it's younger.
(Speculation: I don't think parsec can do what you want. Its standard choice combinator is deterministic. I haven't looked into tweaking or replacing that behaviour, and I wouldn't want to I'm afraid.)
This question immediately reminded me of the Yacc is dead / No, it's not debate from the end of 2010. The authors of the Yacc is dead paper provide a library in Scala (unmaintained), Haskell and Racket. In the Yacc is alive response, Russ Cox points out that the code runs in exponential time for ambiguous grammars.
It's well-known that it is possible to parse ambiguous grammars in O(n^3), although obviously it can take exponential time to enumerate all the parse trees in the case that there are exponentially many of them -- and there will be in the case of x1 + x2 + x3 ... + xn. bison implements the GLR algorithm which does so; unfortunately, while bison is certainly mature (if not actually moribund), it is written neither in Haskell nor in Scala.
Daniel Spiewak implemented a GLL parser in Scala IIRC, but last time I looked at it, it suffered from some performance issues. So I'm not sure that it could be described as mature, either.
I can't speak to how mature it is or give you any usage examples, but I've had the scala gll-combinators library open in a tab for a few days. It handles ambiguous grammars and looks pretty nifty.
At the end the the choice fell on the Syntax Definition Formalism (SDF2)
with an sdf table generator here
and JSGLR as parser generator.
ie: http://en.wikipedia.org/wiki/Wirth_syntax_notation
It seems like most use BNF / EBNF ...
The distinction made by the Wikipedia article looks to me like it is splitting hairs. "BNF/EBNF" has long meant writing grammar rules in roughly the following form:
nonterminal = right_hand_side end_rule_marker
As with other silly langauge differences ( "{" in C, begin in Pascal, endif vs. fi) you can get very different looking but identical meaning by choosing different, er, syntax for end_rule_marker and what you are allowed to say for the right_hand_side.
Normally people allow literal tokens in (your choice! of) quotes, other nonterminal names, and for EBNF, various "choice" operators typically | or / for alternatives, * or + for "repeat", [ ... ] or ? for optional, etc.
Because people designing language syntax are playing with syntax, they seem to invent their own every time they write some down. (Check the various syntax formalisms in the language standards; none of them are the same). Yes, we'd all be better off if there were one standard way to write this stuff. But we don't do that for C or C++ or C# (not even Microsoft could follow their own standard); why should BNF be any different?
Because people that build parser generators usually use it to parse their own syntax, they can and so easily define their own for each parser generator. I've never seen one that did precisely the "WSN" version at Wikipedia and I doubt I ever will in practice.
Does it matter much? Well, no. What really matters is the power of the parser generator behind the notation. You have to bend most grammars to match the power (well, weakness) of the parser generator; for instance, for LL-style generators, your grammar rules can't have left recursion (WSN allows that according to my reading). People building parser generators also want to make it convenient to express non-parser issues (such as, "how to build tree nodes") and so they also add extra notation for non-parsing issues.
So what really drives the parser generator syntax are weaknesses of the parser generator to handle arbitrary languages, extra goodies deemed valuable for the parser generator, and the mood of the guy implementing it.
I've been reading a bit about how interpreters/compilers work, and one area where I'm getting confused is the difference between an AST and a CST. My understanding is that the parser makes a CST, hands it to the semantic analyzer which turns it into an AST. However, my understanding is that the semantic analyzer simply ensures that rules are followed. I don't really understand why it would actually make any changes to make it abstract rather than concrete.
Is there something that I'm missing about the semantic analyzer, or is the difference between an AST and CST somewhat artificial?
A concrete syntax tree represents the source text exactly in parsed form. In general, it conforms to the context-free grammar defining the source language.
However, the concrete grammar and tree have a lot of things that are necessary to make source text unambiguously parseable, but do not contribute to actual meaning. For example, to implement operator precedence, your CFG usually has several levels of expression components (term, factor, etc.), with the operators connecting them at the different levels (you add terms to get expressions, terms are composed of factors optionally multipled, etc.). To actually interpret or compile the language, however, you don't need this; you just need Expression nodes that have operators and operands. The abstract syntax tree is the result of simplifying the concrete syntax tree down to the things actually needed to represent the meaning of the program. This tree has a much simpler definition and is thus easier to process in the later stages of execution.
You usually don't need to actually build a concrete syntax tree. The action routines in your YACC (or Antlr, or Menhir, or whatever...) grammar can directly build the abstract syntax tree, so the concrete syntax tree only exists as a conceptual entity representing the parse structure of your source text.
A concrete syntax tree matches what the grammar rules say is the syntax. The purpose of the abstract syntax tree is have a "simple" representation of what's essential in "the syntax tree".
A real value in the AST IMHO is that it is smaller than the CST, and therefore takes less time to process. (You might say, who cares? But I work with a tool where we have
tens of millions of nodes live at once!).
Most parser generators that have any support for building syntax trees insist that you personally specify exactly how they get built under the assumption that your tree nodes will be "simpler" than the CST (and in that, they are generally right, as programmers are pretty lazy). Arguably it means you have to code fewer tree visitor functions, and that's valuable, too, in that it minimizes engineering energy. When you have 3500 rules (e.g., for COBOL) this matters. And this "simpler"ness leads to the good property of "smallness".
But having such ASTs creates a problem that wasn't there: it doesn't match the grammar, and now you have to mentally track both of them. And when there are 1500 AST nodes for a 3500 rule grammar, this matters a lot. And if the grammar evolves (they always do!), now you have two giant sets of things to keep in synch.
Another solution is to let the parser simply build CST nodes for you and just use those. This is a huge advantage when building the grammars: there's no need to invent 1500 special AST nodes to model 3500 grammar rules. Just think about the tree being isomorphic to the grammar. From the point of view of the grammar engineer this is completely brainless, which lets him focus on getting the grammar right and hacking at it to his heart's content. Arguably you have to write more node visitor rules, but that can be managed. More on this later.
What we do with the DMS Software Reengineering Toolkit is to automatically build a CST based on the results of a (GLR) parsing process. DMS then automatically constructs an "compressed" CST for space efficiency reasons, by eliminating non-value carrying terminals (keywords, punctation), semantically useless unary productions, and forming directly-indexable lists for grammar rule pairs that are list like:
L = e ;
L = L e ;
L2 = e2 ;
L2 = L2 ',' e2 ;
and a wide variety of variations of such forms. You think in terms of the grammar rules and the virtual CST; the tool operates on the compressed representation. Easy on your brain, faster/smaller at runtime.
Remarkably, the compressed CST built this way looks a lot an AST that you might have designed by hand (see link at end to examples). In particular, the compressed CST doesn't carry any nodes that are just concrete syntax.
There are minor bits of awkwardness: for example while the concrete nodes for '(' and ')' classically found in expression subgrammars are not in the tree, a "parentheses node" does appear in the compressed CST and has to be handled. A true AST would not have this. This seems like a pretty small price to pay for the convenience of not have to specify the AST construction, ever. And the documentation for the tree is always available and correct: the grammar is the documentation.
How do we avoid "extra visitors"? We don't entirely, but DMS provides an AST library that walks the AST and handles the differences between the CST and the AST transparently. DMS also offers an "attribute grammar" evaluator (AGE), which is a method for passing values computed at nodes up and down the tree; the AGE handles all the tree representation issues and so the tool engineer only worries about writing computations effectively directly on the grammar rules themselves. Finally, DMS also provides "surface-syntax" patterns, which allows code fragments from the grammar to used to find specific types of subtrees, without knowing most of the node types involved.
One of the other answers observes that if you want to build tools that can regenerate source, your AST will have to match the CST. That's not really right, but it is far easier to regenerate the source if you have CST nodes. DMS generates most of the prettyprinter automatically because it has access to both :-}
Bottom line: ASTs are good for small, both phyiscal and conceptual. Automated AST construction from the CST provides both, and lets you avoid the problem of tracking two different sets.
EDIT March 2015: Link to examples of CST vs. "AST" built this way
This is based on the Expression Evaluator grammar by Terrence Parr.
The grammar for this example:
grammar Expr002;
options
{
output=AST;
ASTLabelType=CommonTree; // type of $stat.tree ref etc...
}
prog : ( stat )+ ;
stat : expr NEWLINE -> expr
| ID '=' expr NEWLINE -> ^('=' ID expr)
| NEWLINE ->
;
expr : multExpr (( '+'^ | '-'^ ) multExpr)*
;
multExpr
: atom ('*'^ atom)*
;
atom : INT
| ID
| '('! expr ')'!
;
ID : ('a'..'z' | 'A'..'Z' )+ ;
INT : '0'..'9'+ ;
NEWLINE : '\r'? '\n' ;
WS : ( ' ' | '\t' )+ { skip(); } ;
Input
x=1
y=2
3*(x+y)
Parse Tree
The parse tree is a concrete representation of the input. The parse tree retains all of the information of the input. The empty boxes represent whitespace, i.e. end of line.
AST
The AST is an abstract representation of the input. Notice that parens are not present in the AST because the associations are derivable from the tree structure.
EDIT
For a more through explanation see Compilers and Compiler Generators pg. 23
This blog post may be helpful.
It seems to me that the AST "throws away" a lot of intermediate grammatical/structural information that wouldn't contribute to semantics. For example, you don't care that 3 is an atom is a term is a factor is a.... You just care that it's 3 when you're implementing the exponentiation expression or whatever.
The concrete syntax tree follows the rules of the grammar of the language. In the grammar, "expression lists" are typically defined with two rules
expression_list can be: expression
expression_list can be: expression, expression_list
Followed literally, these two rules gives a comb shape to any expression list that appears in the program.
The abstract syntax tree is in the form that's convenient for further manipulation. It represents things in a way that makes sense for someone that understand the meaning of programs, and not just the way they are written. The expression list above, which may be the list of arguments of a function, may conveniently be represented as a vector of expressions, since it's better for static analysis to have the total number of expression explicitly available and be able to access each expression by its index.
Simply, AST only contains semantics of the code, Parse tree/CST also includes information on how exactly code was written.
The concrete syntax tree contains all information like superfluous parenthesis and whitespace and comments, the abstract syntax tree abstracts away from this information.
NB: funny enough, when you implement a refactoring engine your AST will again contain all the concrete information, but you'll keep referring to it as an AST because that has become the standard term in the field (so one could say it has long ago lost its original meaning).
CST(Concrete Syntax Tree) is a tree representation of the Grammar(Rules of how the program should be written).
Depending on compiler architecture, it can be used by the Parser to produce an AST.
AST(Abstract Syntax Tree) is a tree representation of Parsed source, produced by the Parser part of the compiler. It stores information about tokens+grammar.
Depending on architecture of your compiler, The CST can be used to produce an AST. It is fair to say that CST evolves into AST. Or, AST is a richer CST.
More explanations can be found on this link: http://eli.thegreenplace.net/2009/02/16/abstract-vs-concrete-syntax-trees#id6
It is a difference which doesn't make a difference.
An AST is usually explained as a way to approximate the semantics of a programming language expression by throwing away lexical content. For example in a context free grammar you might write the following EBNF rule
term: atom (('*' | '/') term )*
whereas in the AST case you only use mul_rule and div_rule which expresses the proper arithmetic operations.
Can't those rules be introduced in the grammar in the first place? Of course. You can rewrite the above compact and abstract rule by breaking it into a more concrete rules used to mimic the mentioned AST nodes:
term: mul_rule | div_rule
mul_rule: atom ('*' term)*
div_rule: atom ('/' term)*
Now, when you think in terms of top-down parsing then the second term introduces a FIRST/FIRST conflict between mul_rule and div_rule something an LL(1) parser cannot deal with. The first rule form was the left factored version of the second one which effectively eliminated structure. You have to pay some prize for using LL(1) here.
So ASTs are an ad hoc supplement used to fix the deficiencies of grammars and parsers. The CST -> AST transformation is a refactoring move. No one ever bothered when an additional comma or colon is stored in the syntax tree. On the contrary some authors retrofit them into ASTs because they like to use ASTs for doing refactorings instead of maintaining various trees at the same time or write an additional inference engine. Programmers are lazy for good reasons. Actually they store even line and column information gathered by lexical analysis in ASTs for error reporting. Very abstract indeed.