/*! A procedural macro to create Lemon-like parsers.

Pomelo is a port to Rust of the Lemon Parser Generator (from now on, Lemon\_C) originally written
by D. Richard Hipp for his SQLite parser.  It is based on a previous attempt to port Lemon to Rust
(Lemon\_Rust), but now it is written as a Rust procedural macro, so it does not contain any of the
original C code (although it uses the same algorithms). Thus the change in name to a different
citrus fruit.

This Pomelo guide is shamelessly based on the original [Lemon\_C guide](http://www.hwaci.com/sw/lemon/lemon.html).

*Pomelo* is an LALR(1) parser generator for Rust. It does the same job as *bison* and *yacc*. But *pomelo* is not another *bison* or *yacc* clone.
It uses a different grammar syntax which is designed to reduce the number of coding errors.
*Pomelo* also uses a more sophisticated parsing engine that is faster than *yacc* and *bison* and which is both reentrant and thread-safe.

## Example

```
# #[macro_use] extern crate pomelo;
pomelo! {
    %type input Vec<i32>;
    %type numbers Vec<i32>;
    %type Number i32;

    input ::= numbers?(A) { A.unwrap_or_else(Vec::new) };
    numbers ::= Number(N) { vec![N] }
    numbers ::= numbers(mut L) Comma Number(N) { L.push(N); L }
}
fn main() -> Result<(), ()> {
    use parser::{Parser, Token};
    // Real-world code would use a tokenizer.
    let tokens = vec![
        Token::Number(1),
        Token::Comma,
        Token::Number(2),
        Token::Comma,
        Token::Number(3),
    ];
    let mut p = Parser::new();
    for tok in tokens.into_iter() {
        p.parse(tok)?;
    }
    let data = p.end_of_input()?;
    assert_eq!(data, vec![1, 2, 3]);
    Ok(())
}
```


## Theory of Operation

The main goal of *pomelo* is to translate a context free grammar (CFG) for a particular language into Rust code that implements a parser for that language.

When using `pomelo!` you write the grammar specification into the macro and it will expand to a module with the following contents:

* A `Parser` struct that implements the parser logic.
* A `Token` enum that represents the input tokens to the parser.

## The Parser Interface

*Pomelo* doesn't generate a complete, working program. It only generates a Rust module, named
`parser` by default, that implement a parser. This section describes the interface to that crate.

Before a program begins using a *pomelo*-generated parser, the program must first create the parser. A new parser is created as follows:

```
# #[macro_use] extern crate pomelo;
# pomelo! { input ::=; }
# fn main() {
let mut parser = parser::Parser::new();
# }
```

Here, `parser` is the generated module. `Parser` is the struct that represents the parser and `new()` the function that creates and initializes a new parser.

The `new()`function may have an argument, depending on the grammar. If the grammar specification
file requests it (see `%extra_argument`), the `new()` function will have a parameter
that can be of any type chosen by the programmer. The parser doesn't do anything with this argument
except to pass a mutable reference to it to action routines. This is a convenient mechanism for
passing state information down to the action routines without having to use global variables.

After a parser has been created, the programmer must supply it with a sequence of tokens (terminal
symbols) to be parsed. This is accomplished by calling the following function once for each token:

```text
parser.parse(token)?;
```

The argument to the `parse()` function is a value of the generated `Token` enumeration that tells the parser the type of the next token in the data stream. There is one token variant for each terminal symbol in the grammar. Some variants will have an associated value, depending on the type of the token. Typically the token variant will be a broad category of tokens such as _identifier_ or _number_ and the data will be the name of the identifier or the value of the number. The return value of this function is `Result<(), Error>` being `Error` the error type of the grammar (see `%error`).

Note that this function will take ownership of the passed token, unless it implements the `Copy` trait (see `%token`).

When all the input has been consumed, the following function may be used to indicate end-of-input:

```text
parser.end_of_input()?;
```

This function actually consumes the parser and returns a value of type `Result<Output, Error>`. If there is no extra type defined, then `Output` is the type of the start symbol of the grammar, or `()` if it has no type. If there is an extra type, then `Output` is a tuple `(ExtraType, TypeOfStartSymbol)`.

A typical use of a *pomelo* parser might look something like the following:

```
# #[macro_use] extern crate pomelo;
# use std::io::BufRead;
# #[derive(Default)] pub struct Error;
# struct Tokenizer;
# pub struct Expression;
# pub struct State;
# impl Tokenizer { 
#     fn new<R: BufRead>(_read: &mut R) -> Self { Tokenizer }
#     fn next_token(&mut self) -> Option<parser::Token> { None }
# }
# impl State { fn new() -> Self { State } }
# pomelo! {
#     %error super::Error;
#     %parse_fail { super::Error }
#     %extra_argument super::State;
#     %type input super::Expression;
#     input ::= { super::Expression }
# }
fn parse_expression<R: BufRead>(read: &mut R) -> Result<Expression, Error> {
    let mut tokenizer = Tokenizer::new(read);
    let mut p = parser::Parser::new(State::new());
    while let Some(token) = tokenizer.next_token() {
        p.parse(token)?;
    }
    let (expr, _state) = p.end_of_input()?;
    Ok(expr)
}
# fn main() { }
```

This example shows a user-written routine that parses an input stream and returns an expression tree.
We assume the existence of some kind of tokenizer which is created using `Tokenizer::new()`. The
`Tokenizer::next_token()` function on retrieves the next token from the input file and returns an
`Option<parser::Token>`. The enum data is assumed to be some type that contains details about each
token, such as its complete text, what line it occurs on, etc.

This example also assumes the existence of a type `parser::State` that holds state
information about a particular parse. An instance of such a type is created with a call to
`parser::State::new()` and then passed into the parser upon initialization, as the optional argument.
The action routine specified by the grammar for the parser can use the this value to hold whatever
information is useful and appropriate. This value can be borrowed between tokens using the function
`parser.extra()` or moved out of the parser with `parser.into_extra()`.

## Differences with *yacc* and *bison*

Programmers who have previously used the *yacc* or *bison* parser generator will notice several
important differences between *yacc* and/or *bison* and *pomelo*.

 * In *yacc* and *bison*, the parser calls the tokenizer. In *pomelo*, the tokenizer calls the parser.
 * *pomelo* uses no global variables. *yacc* and *bison* use global variables to pass information between the tokenizer and parser.
 * *pomelo* allows multiple parsers to be running simultaneously. *yacc* and *bison* do not.

## Differences with *lemon*

The most obvious difference here is that *lemon* is written in C and generates C code, while
*pomelo* is written in Rust and produces Rust code. Many other differences arise from this fact:

 * Since there is no command to call, there are no command line switches.
 * No `%destructor` or `%default_destructor` or `%token_destructor` directives. Rust `drop` semantics should take care or everything.
 * No `%parse_accept` directive. If you want to run code after the end-of-input, just do it after calling Parser::end_of_input().
 * No `%token_type` directive. See below for details.
 * New `%extra_token` directive.
 * `%default_type` applies to all symbols. In *lemon* it applies only to non-terminals because all terminals has the same `%token_type`.
 * Rules are ended with a semi-colon instead of a point: it is more natural for a Rust programmer.
 * The `%stack_overflow` directive returns an error type, just like `%parse_fail`.
 * The `%stack_size` directive can specify an unlimited stack size, with 0, that makes the `%stack_overflow` code unreachable.
 * The `%stack_size` directive can specify a second argument: the type of the stack implementation. This allows to write a `no-std` or an allocation-free parser.

Another important difference is that in *lemon* the `%type` directive only applies to
non-terminals, while terminals all must have the same type, declared with `%token_type`. This is
necessary because the `Parse()` function must be declared with a type able to accept any token. In
*pomelo*, however, the input of the `parser()` function is an `enum` and the types of terminals and
non-terminals can be equally defined just with `%type`. If you want to add location information to
all terminal symbols you can use the `%extra_token` directive.

If the left-hand side of a grammar rule has an user defined type, then it must have a code block to produce its value. You can omit the code block if the rule has the following properties:

 * There is exactly one symbol on the right-hand side with type.
 * The type of that symbol is identical to the type of the left-hand side symbol's.
 * The symbols on the right-hand side do not have alias defined.

Then the rule code is auto-generated to just forward the value of that symbol. For example:

```
# #[macro_use] extern crate pomelo;
# pomelo! {
    %type expr String;
    %type Number String;
    expr ::= Number;
    expr ::= LParen Number RParen;
# }
# fn main() {}
```

is equivalent to:

```
# #[macro_use] extern crate pomelo;
# pomelo! {
    %type expr String;
    %type Number String;
    expr ::= Number(A) { A }
    expr ::= LParen Number(A) RParen { A }
# }
# fn main() {}
```

Another extension to the *lemon* syntax is the *optional* flag: in the right-hand side of a rule
you can add `?` at the end of a symbol to make it optional. The type of the alias is thus changed
from `T` to `Option<T>`. If the symbol is found its value will be `Some(_)`, if it is not found its
value will be `None`. For example:

```text
%type list T;

array ::= LParen list?(D) RParen { ... }
```

is actually converted to something like:

```text
%type list T;
%type optional_list Option<T>;

array ::= LParen optional_list(D) RParen { ... }
optional_list ::= list(D) { Some(D) }
optional_list ::= { None }
```


## Macro input

The main purpose of the `pomelo!` macro is to define the grammar for the parser. But it also
specifies additional information *pomelo* requires to do its job.

The grammar for *pomelo* is, for the most part, free format. It does not have sections or
divisions like *yacc* or *bison*. Any declaration can occur at any point in the macro. *Pomelo*
ignores whitespace (except where it is needed to separate tokens) and it honors the same commenting
conventions as Rust.

### Terminals and Nonterminals

A terminal symbol (token) is any string of alphanumeric and underscore characters that begins with an upper case letter. A terminal can contain lowercase letters after the first character. A nonterminal, on the other hand, is any string of alphanumeric and underscore characters that begins with a lower case letter.

In *pomelo*, terminal and nonterminal symbols do not need to be declared or identified in a separate section of the grammar. *Pomelo* is able to generate a list of all terminals and nonterminals by examining the grammar rules, and it can always distinguish a terminal from a nonterminal by checking the case of the first character of the name.

*Yacc* and *bison* allow terminal symbols to have either alphanumeric names or to be individual characters included in single quotes, like this: `)` or `$`. *Pomelo* does not allow this alternative form for terminal symbols. With *pomelo*, all symbols, terminals and nonterminals, must have alphanumeric names.

### Grammar Rules

The main component of a *pomelo* grammar is a sequence of grammar rules. Each grammar rule consists
of a nonterminal symbol followed by the special symbol `::=` and then a list of terminals and/or
nonterminals. The rule is terminated by a semi-colon. The list of terminals and nonterminals on the
right-hand side of the rule can be empty. Rules can occur in any order, except that the left-hand
side of the first rule is assumed to be the start symbol for the grammar (unless specified otherwise
using `%start`). A typical sequence of grammar rules might look something like this:

```
# #[macro_use]
# extern crate pomelo;
# pomelo! {
# %left Plus;
# %left Times;
    input ::= expr;
    expr ::= expr Plus expr;
    expr ::= expr Times expr;
    expr ::= LParen expr RParen;
    expr ::= Value;
# }
# fn main() {}
```

There is one non-terminal in this example, `expr`, and five terminal symbols or tokens: `Plus`, `Times`, `LParen`, `RParen` and `Value`.

Like *yacc* and *bison*, *pomelo* allows the grammar to specify a block of code that will be
executed whenever a grammar rule is reduced by the parser. In *pomelo*, this action is specified by
putting the code (contained within curly braces {...}) in place of the semi-colon that closes the
rule. For example:

```text
expr ::= expr Plus expr { println!("Doing an addition..."); }
```

In order to be useful, grammar actions must normally be linked to their associated grammar rules.
In *yacc* and *bison*, this is accomplished by embedding a `$$` in the action to stand for the
value of the left-hand side of the rule and symbols `$1`, `$2`, and so forth to stand for the value
of the terminal or nonterminal at position 1, 2 and so forth on the right-hand side of the rule.
This idea is very powerful, but it is also very error-prone. The single most common source of
errors in a *yacc* or *bison* grammar is to miscount the number of symbols on the right-hand side
of a grammar rule and say `$7` when you really mean `$8`.

*Pomelo* avoids the need to count grammar symbols by assigning symbolic names to each symbol in a
grammar rule and then using those symbolic names in the action. Moreover, the value to be assigned
to the left-hand side is simply the output value of the rule block. In *yacc* or *bison*, one would
write this:

```text
expr -> expr PLUS expr { $$ = $1 + $3; }
```

But in *pomelo*, the same rule becomes the following:

```text
expr ::= expr(A) Plus expr(B) { B + C }
```

In the *pomelo* rule, any symbol in parentheses after a grammar rule symbol becomes an irrefutable
pattern to match the corresponding value of that symbol.

The *pomelo* notation for linking a grammar rule with its reduce action is superior to *yacc* or
*bison* on several counts. First, as mentioned above, the *pomelo* method avoids the need to count
grammar symbols. Secondly, you cannot forget to assign to the left-hand side symbol: if the code block
does not have the same type as the left-hand side symbol, a compiler error will be raised.

If you have several terminal tokens that can be used in the same place you can put them all in the
same rule, separated with `|`.

```text
expr ::= SmallNumber|BigNumber(B) { B }
```

which is a shortcut of

```text
expr ::= SmallNumber(B) { B }
expr ::= BigNumber(B) { B }
```

If you use a symbolic name (`(B)` in the example) with such a compound token, then all these tokens
must be of the same type. However, if there is no symbolic name, then they may have different
types.

### Precedence Rules

*pomelo* resolves parsing ambiguities in exactly the same way as *yacc* and *bison*. A shift-reduce
conflict is resolved in favor of the shift, and a reduce-reduce conflict is resolved by reducing
whichever rule comes first in the grammar file.

Just like in *yacc* and *bison*, *pomelo* allows a measure of control over the resolution of
conflicts using precedence rules. A precedence value can be assigned to any terminal symbol using
the `%left`, `%right` or `%nonassoc` directives. Terminal symbols mentioned in earlier directives
have a lower precedence that terminal symbols mentioned in later directives. For example:

```text
%left And;
%left Or;
%nonassoc Eq Ne Gt Ge Lt Le;
%left Plus Minus;
%left Times Divide Mod;
%right Exp Not;
```

In the preceding sequence of directives, the `And` operator is defined to have the lowest
precedence. The `Or` operator is one precedence level higher. And so forth. Hence, the grammar
would attempt to group the ambiguous expression

```text
a And b Or c
```

like this

```text
a And (b Or c)
```

The associativity (left, right or nonassoc) is used to determine the grouping when the precedence is the same. `And` is left-associative in our example, so

```text
a And b And c
```

is parsed like this

```text
(a And b) And c
```

The `Exp` operator is right-associative, though, so

```text
a Exp b Exp c
```

is parsed like this

```text
a EXP (b EXP c)
```

The nonassoc precedence is used for non-associative operators. So

```text
a Eq b Eq c
```

is an error.

The precedence of non-terminals is transferred to rules as follows: The precedence of a grammar
rule is equal to the precedence of the left-most terminal symbol in the rule for which a precedence
is defined. This is normally what you want, but in those cases where you want to precedence of a
grammar rule to be something different, you can specify an alternative precedence symbol by putting
the symbol in square braces before the semi-colon or the rule code. For example:

```text
expr ::= Minus expr [Not];
```

This rule has a precedence equal to that of the Not symbol, not the Minus symbol as would have been the case by default.

With the knowledge of how precedence is assigned to terminal symbols and individual grammar rules,
we can now explain precisely how parsing conflicts are resolved in *pomelo*. Shift-reduce
conflicts are resolved as follows:

 * If either the token to be shifted or the rule to be reduced lacks precedence information, then resolve in favor of the shift, but report a parsing conflict.
 * If the precedence of the token to be shifted is greater than the precedence of the rule to reduce, then resolve in favor of the shift. No parsing conflict is reported.
 * If the precedence of the token it be shifted is less than the precedence of the rule to reduce, then resolve in favor of the reduce action. No parsing conflict is reported.
 * If the precedences are the same and the shift token is right-associative, then resolve in favor of the shift. No parsing conflict is reported.
 * If the precedences are the same the the shift token is left-associative, then resolve in favor of the reduce. No parsing conflict is reported.
 * Otherwise, resolve the conflict by doing the shift and report the parsing conflict.

Reduce-reduce conflicts are resolved this way:

 * If either reduce rule lacks precedence information, then resolve in favor of the rule that appears first in the grammar and report a parsing conflict.
 * If both rules have precedence and the precedence is different then resolve the dispute in favor of the rule with the highest precedence and do not report a conflict.
 * Otherwise, resolve the conflict by reducing by the rule that appears first in the grammar and report a parsing conflict.

### Special Directives

The input grammar to *pomelo* consists of grammar rules and special directives. We've described all
the grammar rules, so now we'll talk about the special directives.

Directives in *pomelo* can occur in any order. You can put them before the grammar rules, or after
the grammar rules, or in the midst of the grammar rules. It doesn't matter. The relative order of
directives used to assign precedence to terminals is important, but other than that, the order of
directives is arbitrary.

*Pomelo* supports the following special directives:

 * [`%module`](#the-module-directive)
 * [`%type`](#the-type-directive)
 * [`%include`](#the-include-directive)
 * [`%syntax_error`](#the-syntax_error-directive)
 * [`%parse_fail`](#the-parse_fail-directive)
 * [`%stack_overflow`](#the-stack_overflow-directive)
 * [`%stack_size`](#the-stack_size-directive)
 * [`%left`](#the-left-right-nonassoc-directives)
 * [`%right`](#the-left-right-nonassoc-directives)
 * [`%nonassoc`](#the-left-right-nonassoc-directives)
 * [`%default_type`](#the-default_type-directive)
 * [`%extra_argument`](#the-extra_argument-directive)
 * [`%error`](#the-error-directive)
 * [`%start_symbol`](#the-start_symbol-directive)
 * [`%fallback`](#the-fallback-directive)
 * [`%wildcard`](#the-wildcard-directive)
 * [`%token_class`](#the-token_class-directive)
 * [`%token`](#the-token-directive)
 * [`%parser`](#the-parser-directive)
 * [`%extra_token`](#the-extra_token-directive)
 * [`%verbose`](#the-verbose-directive)

#### The `%module` directive

This directive is used to specify the name of the module generated by the `pomelo!` macro. Fo example

```text
%module ident;
```

will create a module named `ident` instead of the default `parser`. This is specially useful if you want to create several parsers in the same module.

#### The `%type` directive

This directive is used to specify the data types for values on the parser's stack associated with
terminal and non-terminal symbols. The type of the terminal symbol is the value associated to the input token. The type associated to a non-terminal will be the type of the data associated
to the corresponding variant of the `Token` enumeration. For example:

```text
%type Value i32;
```

Then the `Token` enumeration will have a variant such as:

```text
pub Token {
    ...
    Value(i32),
}
```

Typically the data type of a non-terminal is a parse-tree structure that contains all information about that non-terminal. For example:

```text
%type expr ExprType;
```

Some Rust crates use derive macros in enums and require attributes in the Token enum variants. You can add them here, only for terminal tokens.
For example:
```text
%type #[serde(rename = "Amount")] Quantity i32;
```
will create a variant:
```text
pub Token {
    ...
    #[serde(rename = "Amount")] Quantity(i32);
}
```

The type of the symbol is optional: if you do not specify the variant will have no type. This is usually not needed because *pomelo* will
create typeless terminals automatically if they appear in the grammar rules, but it can be needed to specify the attributes.

Each entry on the parser's stack is actually an enum containing variants of all data types for
every symbol. *Pomelo* will automatically use the correct element of this enum depending on what
the corresponding symbol is. But the grammar designer should keep in mind that the size of the enum
will be the size of its largest element. So if you have a single non-terminal whose data type
requires 1K of storage, then your 100 entry parser stack will require 100K of heap space. If you
are willing and able to pay that price, fine. You just need to know.

#### The `%include` directive

The `%include` directive specifies Rust code that is included into the generated module. You can
include any Rust items you want. You can have multiple `%include` directives in your grammar.

The `%include` directive is very handy for using symbols declared elsewhere. For example:

```text
%include { use super::*; }
```

#### The `%syntax_error` directive

The `%syntax_error` directive specify code that will be called when a syntax error occurs. This code must evaluate to a value of type `Result<(), Error>`, and it is run in an function that returns the same type so you can also use the `?` operator. If it evaluates to `Ok(())`, the parser will try to recover and continue. If it evaluates to `Err(_)` or a `?` fails, the parser will fail with that error value. See the section [Error Processing](#error-processing) for more details.

In this code you have available `extra` as a mutable reference to the current `extra_argument`, and `token` as a type of value `Option<Token>` with the token that triggered the error. If the error is caused by the end-of-input, then `token` will be `None`.

By default it evaluates to `Err(Default::default())` so:

 * if `Error` implements `Default` it will fail with the default error.
 * if `Error` does not implement `Default` it will fail to compile and you *must* use this directive to create a meaningful one or to return `Ok(())` and ignore the error.

#### The `%parse_fail` directive

The `%parse_fail` directive specifies a block of Rust code that is executed whenever the parser
fails to complete. This code is not executed until the parser has tried and failed to resolve an
input error using is usual error recovery strategy. This block is only invoked when parsing is
unable to continue. It must evaluate to the defined `Error` type.

```text
%error String;
%parse_failure {
    "Giving up.  Parser is hopelessly lost...".to_string()
}
```

By default it will  return `Default::default()`. If your `Error` type does not implement `default()`
it will be a compiler error, so in this case you must use this directive.

After a parse failure this parser object must not be used again.

#### The `%stack_overflow` directive

The `%stack_overflow` directive specifies a block of Rust code that is executed whenever the
internal stack overflows. Beware of righ recursivity rules and right associativity! It must be evaluated to the defined `Error` type.

By default it will  return `Default::default()`. If your `Error` type does not implement `default()`
it will be a compiler error, so in this case you must use this directive.

However, if you set your `%stack_size` to `0` (unlimited), then the stack will never overflow, and
this directive is defaulted to `unreachable!()`.

```text
%error String;
%stack_overflow {
    "Parse stack overflow!".to_string()
}
```

After a stack overflow this parser object must not be used again.

See also the `%stack_size` directive for more details about the parser stack.

#### The `%stack_size` directive

If stack overflow is a problem and you can't resolve the trouble by using left-recursion, then you
might want to increase the size of the parser's stack using this directive. Put an positive integer
after the %stack_size directive and *pomelo* will generate a parse with a stack of the requested size.
The default value is 100.

```text
%stack_size 2000;
```

If you specify the limit as `0` then it means _unlimited_. Beware that if your parser input is
untrusted, having an unlimited stack may be an issue (it may eat up all your memory). If you do
this, the code in `%stack_overflow` directive is never called.

This directive has an additional optional parameter that is the type to implement the stack. It is
by default `std::vec::Vec`, but you may specify whatever type you want, as long as it complies the
following interface:
 * It must have a single generic type argument, without restrictions other than `Sized`. That is, if you write `%stack_size 100 Type;` then `Type<T>` must be a valid type.
 * It must implement the following member public functions, with the same signature and semantics as those in `Vec`:
    - `new()`
    - `push()`
    - `pop()`
    - `last()`
    - `is_empty()`
    - `clear()`
    - `len()`

You can use alternative types for the stack to make your parser `no-std` compliant. For example, using the `arrayvec` crate:

```text
%include {
    use arrayvec::ArrayVec;
    type Stack<T> = ArrayVec<[T; 32]>;
}
%stack_size 32 Stack;
```

#### The `%left`, `%right`, `%nonassoc` directives

The `%left`, `%right` and `%nonassoc` directives are used to declare precedences of terminal
symbols. Every terminal symbol whose name appears in one of those directives is given the same
associative precedence value. Subsequent directives have higher precedence. For example:

```text
%left And;
%left Or;
%nonassoc Eq Ne Gt Ge Lt Le;
%left Plus Minus;
%left Times Divide Mod;
%right Exp Not;
```

Note the semi-colon that terminates each `%left`, `%right` or `%nonassoc` directive.

LALR(1) grammars can get into a situation where they require a large amount of stack space if you
make heavy use or right-associative operators. For this reason, it is recommended that you use
`%left` rather than `%right` whenever possible.

#### The `%default_type` directive

This directive specifies a default type for all the symbols that do not specify a type.

#### The `%extra_argument` directive

The `%extra_argument` directive instructs *pomelo* to add a parameter to the `Parser::new()` function it generates. *Pomelo* doesn't do anything itself with this extra argument, but it does make the argument available to Rust-code action routines, and so forth, as a mutable refernce named `extra`. For example, if the grammar file contains:

```text
%extra_argument MyStruct;
```

Then the function generated will be of the form `Parser::new(extra: MyStruct)` and all action routines will have access to a variable as `extra: &mut MyStruct` that is the value of the stored argument.

Moreover, there will be the following extra public member functions in the `Parser` struct:

```text
pub fn into_extra(self) -> MyStruct;
pub fn extra(&self) -> &MyStruct;
pub fn extra_mut(&mut self) -> &mut MyStruct;
```

Also, if defined the `end_of_input()` member function will return a tuple, with the extra value as its second value.

#### The `%error` directive

This directive defines the type of the parser error. If not defined, it will default to `()`. Both functions of the `Parser` struct, `parse()` and `end_of_input()` return a `Result<_,Error>` with this type.

Also, any rule block can return an `Err(Error)` (usually with the `?` operator) to force a parser error.

For example:

```text
%error String;
```

If your error type does not implement `Default` then you must also write:
 * the `%syntax_error` directive to create a meaningful error (or ignore it).
 * the `%parse_fail` directive to return an error in case an unrecoverable error happens.
 * the `%stack_overflow` directive to return an error in case the internal stack overflows (unless you set the `%stack_size` to unlimited).

#### The `%start_symbol` directive

By default, the start symbol for the grammar that *pomelo* generates is the first non-terminal that
appears in the grammar file. But you can choose a different start symbol using the `%start_symbol`
directive.

```text
%start_symbol program;
```

#### The `%fallback` directive

This directive defines an alternative token that will be used instead of another if the original one cannot be parsed. For example:

```text
%fallback Id X Y Z;
```

declares the token `Id` as a fallback for any of the other tokens. If the input stream passes any
of these three tokens and they cannot be parsed, then the parser will try parsing an `Id` before
considering it an error.

The fallback token (`Id` in the example) must have the same type of every other token that it
replaces, or no type at all.

#### The `%wildcard` directive

This directive defines a token that will be used when any other token cannot be parsed. For example:

```text
%wildcard Any;
```

The wildcard token must not have a type.

#### The `%token_class` directive

This directive declares a compound token class. For example:

```text
%token_class number Integer Float Double;
```

is equivalent but more efficient than:

```text
number ::= Integer(A); { A }
number ::= Float(A);   { A }
number ::= Double(A);  { A }
```

or also:

```text
number ::= Integer|Float|Double(A) { A }
```

Naturally, if they use a symbolic name (`(A)` in the example), then all the tokens must have the same type.

#### The `%token` directive

This directive is used to customize the `Token` enumeration generated by *pomelo*. It must be followed by an enumeration declaration named `Token` without any variants (they will be filled in by the macro). It can be used to add auto-derive traits, change its visibility, add custom attributes, add generics... For example:

```text
%token #[derive(Copy,Clone,Debug)]
       pub enum Token {};
```

The default if it is this directive is not used is `pub enum Token {}`.

#### The `%parser` directive

This directive is used to customize the `Parser` struct generated by *pomelo*. It must be followed by a struct declaration named `Parser` without any fields (they will be filled in by the macro). It can be used to add auto-derive traits, change its visibility, add custom attributes, add generics... For example:

```text
%parser pub struct Parser<'a> {};
```

The default if it is this directive is not used is `pub struct Parser {}`, but with the added generic arguments from `%token` if any.

For more about generic arguments see the [Generic Parsers](#generic-parsers) section.

#### The `%extra_token` directive

Sometimes, all tokens share a common piece of data. This is usually some kind of location (line/column information). In those cases, instead of adding it to the type of all terminal tokens, you can use this directive that takes a single type as argument. For example:

```text
%extra_token Loc;
```

will change every terminal of type `T` into one of type `(Loc, T)`. Any terminal without a type will get a type of `Loc` instead. Non-terminal symbols are unchanged.

If you use this directive with a type `E`, then the `Token` enum gains the following member functions:

```text
fn into_extra(self) -> E;
fn extra(&self) -> &E;
fn extra_mut(&mut self) -> &mut E;
```

These are particularly useful in the `%syntax_error` code to build a meaninful error message.

#### The `%verbose` directive

This directive makes *pomelo* to dump the built states of the grammar to the console. This is mostly useful for diagnostics or for fine tuning your grammar.

### Error Processing

After extensive experimentation over several years, it has been discovered that the error recovery
strategy used by *yacc* is about as good as it gets. And so that is what *pomelo* uses.

When a *pomelo*-generated parser encounters a syntax error, it first invokes the code specified by
the `%syntax_error` directive, if any. It then enters its error recovery strategy. The error
recovery strategy is to begin popping the parsers stack until it enters a state where it is
permitted to shift a special non-terminal symbol named `error`. It then shifts this non-terminal
and continues parsing. But the `%syntax_error` routine will not be called again until at least
three new tokens have been successfully shifted.

If the parser pops its stack until the stack is empty, and it still is unable to shift the error
symbol, then the `%parse_fail` routine is invoked and the parser fails. This is what will happen at
the very first syntax error, of course, if there are no instances of the `error` non-terminal in
your grammar.

### Generic parsers

You can use generic arguments either in the `Token` enum, the `Parser` struct, or both. For that,
the directives `%token` and `%parser` are used. Note that, since the `Parser` contains partially
parsed tokens, every generic argument used in `Token` must also be specified in `Parser`. This is
the default if `%parser` is not used but must be taken into account if you use both directives.

Generic arguments in `Parser` can be used anywhere except terminal symbol types. That includes
non-terminal types, the `%extra_argument` type and the `%error` type.

In addition to those, generic arguments in `Token` can also be used in terminal symbol types and the
`%extra_token` directive.

Generic parsers can be very handy, as illustrated by the following two examples.

#### Generic lifetimes

Generic lifetime parameters allow the optional extra argument `%extra_argument` to reference shared data.
Additionally, tokens may refer to data slices within the input of the parser.

```
# #[macro_use] extern crate pomelo;
# use std::collections::HashMap;
pomelo! {
    %include { use std::collections::HashMap; }
    %parser pub struct Parser<'e>{};
    %token pub enum Token<'e>{};
    %extra_argument &'e HashMap<&'e[u8], i32>;

    %type input i32;
    %type numbers i32;
    %type number i32;
    %type Number i32;
    %type Var &'e[u8];
    input ::= numbers?(A) { A.unwrap_or(0) };
    number ::= Number(N) { N }
    number ::= Var(X) { match extra.get(X) { Some(v) => *v, _ => return Err(())} }
    numbers ::= number(N) { N }
    numbers ::= numbers(V) Comma number(N) { V + N }
}

fn main() -> Result<(), ()> {
    use parser::{Parser, Token};
    let input = b"1,x,2";
    // Real-world code would use a tokenizer.
    let x = &input[2..3];
    let tokens = vec![
        Token::Number(1),
        Token::Comma,
        Token::Var(x),
        Token::Comma,
        Token::Number(2),
    ];

    let mut values = HashMap::new();
    values.insert(x, 4);

    let mut p = Parser::new(&values);
    for tok in tokens.into_iter() {
        p.parse(tok)?;
    }
    let (value, _) = p.end_of_input()?;
    assert_eq!(value, 7);
    Ok(())
}
```

#### Generic production rules

Using a generic extra argument constrained by a trait makes it possible to change output types
and add side effects to production rules while re-using the same parser module.

```
# #[macro_use] extern crate pomelo;
pub trait Adder {
    type Value;
    fn from_int(&mut self, value: i32) -> Self::Value;
    fn add(&mut self, v1: Self::Value, v2: Self::Value) -> Self::Value;
}

pomelo! {
    %parser pub struct Parser<T: super::Adder> {};
    %extra_argument T;
    %token #[derive(Clone)] pub enum Token{};

    %type input T::Value;
    %type numbers T::Value;
    %type Number i32;
    input ::= numbers?(A) { A.unwrap_or_else(|| extra.from_int(0)) };
    numbers ::= Number(N) { extra.from_int(N) }
    numbers ::= numbers(V) Comma Number(N) { let w = extra.from_int(N); extra.add(V, w) }
}

struct OpsCounter(usize);

impl Adder for OpsCounter {
    type Value = i32;

    fn from_int(&mut self, value: i32) -> Self::Value {
        value
    }

    fn add(&mut self, v1: Self::Value, v2: Self::Value) -> Self::Value {
        self.0 += 1;
        v1 + v2
    }
}

struct LazyAdder;

impl Adder for LazyAdder {
    type Value = Vec<i32>;

    fn from_int(&mut self, value: i32) -> Self::Value {
        vec![value]
    }

    fn add(&mut self, mut v1: Self::Value, v2: Self::Value) -> Self::Value {
        v1.extend_from_slice(&v2);
        v1
    }
}

fn main() -> Result<(), ()> {
    use parser::{Parser, Token};
    // Real-world code would use a tokenizer.
    let tokens = vec![
        Token::Number(1),
        Token::Comma,
        Token::Number(2),
        Token::Comma,
        Token::Number(3),
    ];
    let mut p = Parser::new(OpsCounter(0));
    for tok in tokens.iter() {
        p.parse(tok.clone())?;
    }
    let (value, counter) = p.end_of_input()?;
    assert_eq!(value, 6);
    assert_eq!(counter.0, 2);

    let mut p = Parser::new(LazyAdder);
    for tok in tokens.into_iter() {
        p.parse(tok)?;
    }
    let (numbers, _) = p.end_of_input()?;
    assert_eq!(numbers, vec![1,2,3]);
    Ok(())
}
```

*/

#![no_std]

#[doc(hidden)]
pub use pomelo_impl::pomelo_impl;

/// The main macro of this crate. See the crate-level documentation for details.
#[macro_export]
macro_rules! pomelo { ($($t:tt)* ) => ( $crate::pomelo_impl!{$($t)*} ) }

#[cfg(feature = "doc_generated")]
pub mod generated;