Running Parsers
There are different ways of running your parser on an input.
try() and tryString()
Most of the time, you'll want to use try
. It will run the parser on an input Stream
, return a Succeed
(which implements ParseResult
) on success, and throw a ParserHasFailed
exception if the input can't successfully be parsed.
The Stream
type generalises over a different ways of providing input. The simplest implementation is StringStream
. This is really a wrapper around a PHP multibyte string.
(In v0.6.0, StringStream
is also the only implementation of Stream
, but this will change.)
ParseResult
has an output()
method, which has the type T
for a Parser<T>
(see Mapping to Objects). It also has a remainder()
method, which gives you the part of the input that wasn't consumed by the parser.
ParserHasFailed
has the usual Exception
methods. It also gives you access to the Fail implements ParseResult
object. This contains all the relevant information about the failure, such as expected()
, got()
, and position()
.
run()
run
is mostly intended for internal use.
The main difference between run
and try
is that run
doesn't throw exceptions when parsing an input fails. (It might throw exceptions if your parser itself is incorrectly defined.) Instead, you'll always get a ParseResult
, and you can inspect it with the same methods as above. You'll also get isSuccess
and isFail
, so you know what you're dealing with.
Continue with a result
Using run
instead of try
is a good choice when you want to do something with the result, such as:
- Building your own combinators
- Interacting with
ParseResult
while in the middle of a parse flow
To do that, ParseResult
lets you continue parsing:
continueWith
takes another parser, and uses it to parse the remainder of the of the result. You may have noticed we didn't check for isSuccess
. That's becasue we don't need to. continueWith
is smart; if $parser1
fails, trying to continue parsing on the result will not have any effect. In fact, the example above will fail, because $parser1
doesn't take into account the space between "hello" and "world".