PHP

Under the Hood

/haskellcamargo
marcelocamargo@linuxmail.org

About me

  • PHP community contributor
  • Creator of Quack programming language
  • Free software translator
  • Author of "Skype for Linux" 2015
  • Author of "Gemidão do Zap" ✆ 
  • Compilers developer
  • Functional programming evangelist
  • A cat lover!

YOU PROBABLY

CODE PHP EVERYDAY

But did you wonder how it works or still think it's dark magic?

THE INPUT PROGRAM

factorial(n: int): int
1 * 2 * 3 * 4 * 5

THE SERVER REQUEST

THE

LEXER

What is a lexer?

  • Receives characters and outputs words
  • Words are easier to analyse than characters!

Hello, how are you?

[HELLO][,][HOW][ARE][YOU][?]

And PHP also does that to your source!

It breaks everything in "tokens"

And it also exposes:

token_get_all(string): array

THE

PARSER

The parser

  • Receives tokens, PHP <= 5.6: emits opcodes
  • Receives tokens, PHP >= 7: generates AST
  • Uses a formal "grammar"
  • A grammar is a definition of rules of the language

Every implemented compiler has a parser

The parser does:

  • Resolve precedence rules

The parser does:

  • Resolve associativity rules

It detects syntax errors

THE SYMBOL TABLE

PHP uses hashtables to store variables and control scope

Variable Value
$_GET array(...)
$_POST array(...)
$_GLOBALS array(...)

Global scope

Variable Value
$n 5
$i 1
$result 1

Factorial scope

THE VM AND OPCODES

The 💖 of PHP

PHP runs in a register based VM

  • The Zend Virtual Machine
  • It handles opcodes
  • Something like a "machine independent" assembly
  • It has a lot of instructions to make things easier!

Introducing opcodes

Opcodes are simple instructions that may take operands

Know the VM

and write faster code

DEBUG IT!

Variable Value
IP 0
Variable Value
IP 0
!0 5
Variable Value
IP 1
!0 5
!1 1
Variable Value
IP 2
!0 5
!1 1
!2 1
Variable Value
IP 3
!0 5
!1 1
!2 1
Variable Value
IP 7
!0 5
!1 1
!2 1
~7 1 
1 <= 5
Variable Value
IP 8
!0 5
!1 1
!2 1
~7 1

If ~7 is not 0, go to instruction 4

Variable Value
IP 4
!0 5
!1 1
!2 1
~7 1 
1 * 1
Variable Value
IP 5
!0 5
!1 1
!2 2
~7 1
~6 1
Variable Value
IP 6
!0 5
!1 1
!2 2
~7 1 
2 <= 5
Variable Value
IP 7
!0 5
!1 1
!2 2
~7 1
Variable Value
IP 8
!0 5
!1 1
!2 2
~7 1
Variable Value
IP 4
!0 5
!1 2
!2 2
~7 1 
1 * 2
Variable Value
IP 5
!0 5
!1 2
!2 3
~7 1
~6 2
Variable Value
IP 6
!0 5
!1 2
!2 3
~7 1
Variable Value
IP 7
!0 5
!1 2
!2 3
~7 1 
3 <= 5
Variable Value
IP 4
!0 5
!1 6
!2 3
~7 1 
2 * 3
Variable Value
IP 5
!0 5
!1 6
!2 4
~7 1
~6 3
Variable Value
IP 6
!0 5
!1 6
!2 4
~7 1
Variable Value
IP 7
!0 5
!1 6
!2 4
~7 1 
4 <= 5
Variable Value
IP 8
!0 5
!1 6
!2 4
~7 1
Variable Value
IP 4
!0 5
!1 24
!2 4
~7 1 
6 * 4
Variable Value
IP 5
!0 5
!1 24
!2 5
~7 1
~6 4
Variable Value
IP 6
!0 5
!1 24
!2 5
~7 1
Variable Value
IP 7
!0 5
!1 24
!2 5
~7 1 
5 <= 5
Variable Value
IP 8
!0 5
!1 24
!2 5
~7 1
Variable Value
IP 4
!0 5
!1 120
!2 5
~7 1 
24 * 5
Variable Value
IP 5
!0 5
!1 120
!2 6
~7 1
~6 5
Variable Value
IP 6
!0 5
!1 120
!2 6
~7 1
Variable Value
IP 7
!0 5
!1 120
!2 6
~7 0 
6 <= 5
Variable Value
IP 8
!0 5
!1 120
!2 6
~7 0
Variable Value
IP 9
!0 5
!1 120
!2 6
~7 0 
120

Do you know opcache?

Resources to study

Resources to study

  • PHP Internals book
  • Compilers: principles, techniques and tools
  • PHP: Zend Engine 2 Opcodes
  • PHP: PHP at the Core: A Hacker's Guide
  • Kaleidoscope: Implementing a Language with LLVM

Ok, thanks, bye!

😂😂👌👌👌🔝🔝🔝

PHP Under the Hood

By Marcelo Camargo

Made with Slides.com

PHP Under the Hood

PHP Under the Hood

  • 3,872

More from Marcelo Camargo