comment | ||
errors | ||
node | ||
parser | ||
php5 | ||
php7 | ||
position | ||
printer | ||
scanner | ||
visitor | ||
walker | ||
.gitignore | ||
CODE_OF_CONDUCT.md | ||
CONTRIBUTING.md | ||
doc.go | ||
ISSUE_TEMPLATE.md | ||
LICENSE | ||
main.go | ||
Makefile | ||
parser.jpg | ||
README.md |
PHP Parser written in Go
This project uses goyacc and golex libraries to parse PHP sources into AST. It can be used to write static analysis, refactoring, metrics, code style formatting tools.
Try it online: demo
Features:
- Fully support PHP 5 and PHP 7 syntax
- Abstract syntax tree (AST) representation
- Traversing AST
- Namespace resolver
- Able to parse syntax-invalid PHP files
Roadmap
- Pretty printer
- Control Flow Graph (CFG)
- PhpDocComment parser
- Stabilize api
Install
go get github.com/z7zmey/php-parser
CLI
php-parser [flags] <path> ...
flag | type | description |
---|---|---|
-d | string | dump format: [custom, go, json, pretty-json] |
-prof | string | start profiler: [cpu, mem] |
-p | bool | show positions |
-c | bool | show comments |
-r | bool | resolve names |
-php5 | bool | parse as PHP5 |
Dump AST to stdout.
Example
package main
import (
"fmt"
"bytes"
"os"
"github.com/z7zmey/php-parser/php7"
"github.com/z7zmey/php-parser/visitor"
)
func main() {
src := bytes.NewBufferString(`<? echo "Hello world";`)
parser := php7.NewParser(src, "example.php")
parser.Parse()
for _, e := range parser.GetErrors() {
fmt.Println(e)
}
visitor := visitor.Dumper{
Writer: os.Stdout,
Indent: "",
Comments: parser.GetComments(),
Positions: parser.GetPositions(),
}
rootNode := parser.GetRootNode()
rootNode.Walk(visitor)
}
Namespace resolver
Namespace resolver is a visitor that resolves nodes fully qualified name and saves into map[node.Node]string
structure
- For
Class
,Interface
,Trait
,Function
,Constant
nodes it saves name with current namespace. - For
Name
,Relative
,FullyQualified
nodes it resolvesuse
aliases and saves a fully qualified name.
Parsing syntax-invalid PHP files
If we try to parse $a$b;
then the parser triggers error 'syntax error: unexpected T_VARIABLE'. Token $b
is unexpected, but parser recovers parsing process and returns $b;
statement to AST, because it is syntactically correct.
Pretty printer [work in progress]
nodes := &stmt.StmtList{
Stmts: []node.Node{
&stmt.Namespace{
NamespaceName: &name.Name{
Parts: []node.Node{
&name.NamePart{Value: "Foo"},
},
},
},
&stmt.Class{
Modifiers: []node.Node{
&node.Identifier{Value: "abstract"},
},
ClassName: &name.Name{
Parts: []node.Node{
&name.NamePart{Value: "Bar"},
},
},
Extends: &stmt.ClassExtends{
ClassName: &name.Name{
Parts: []node.Node{
&name.NamePart{
Value: "Baz"
},
},
},
},
Stmts: []node.Node{
&stmt.ClassMethod{
Modifiers: []node.Node{
&node.Identifier{Value: "public"},
},
MethodName: &node.Identifier{Value: "greet"},
Stmt: &stmt.StmtList{
Stmts: []node.Node{
&stmt.Echo{
Exprs: []node.Node{
&scalar.String{Value: "'Hello world'"},
},
},
},
},
},
},
},
},
}
file := os.Stdout
p := printer.NewPrinter(file, " ")
p.Print(nodes)
It prints to stdout:
<?php
namespace Foo;
abstract class Bar extends Baz
{
public function greet()
{
echo 'Hello world';
}
}