php / php-src Goto Github PK

The PHP Interpreter

License: Other

JavaScript 0.20% C 67.38% Makefile 0.03% Shell 0.32% PHP 30.22% DTrace 0.01% C++ 0.60% Awk 0.01% HTML 0.01% GAP 0.02% XSLT 0.01% Batchfile 0.01% Yacc 0.09% M4 0.41% GDB 0.01% Lua 0.29% Lex 0.10% Roff 0.03% Assembly 0.25% Python 0.01%

php-src's Introduction

The PHP Interpreter

PHP is a popular general-purpose scripting language that is especially suited to web development. Fast, flexible and pragmatic, PHP powers everything from your blog to the most popular websites in the world. PHP is distributed under the PHP License v3.01.

Documentation

The PHP manual is available at php.net/docs.

Installation

Prebuilt packages and binaries

Prebuilt packages and binaries can be used to get up and running fast with PHP.

For Windows, the PHP binaries can be obtained from windows.php.net. After extracting the archive the *.exe files are ready to use.

For other systems, see the installation chapter.

Building PHP source code

For Windows, see Build your own PHP on Windows.

For a minimal PHP build from Git, you will need autoconf, bison, and re2c. For a default build, you will additionally need libxml2 and libsqlite3.

On Ubuntu, you can install these using:

sudo apt install -y pkg-config build-essential autoconf bison re2c \
                    libxml2-dev libsqlite3-dev

On Fedora, you can install these using:

sudo dnf install re2c bison autoconf make libtool ccache libxml2-devel sqlite-devel

Generate configure:

./buildconf

Configure your build. --enable-debug is recommended for development, see ./configure --help for a full list of options.

# For development
./configure --enable-debug
# For production
./configure

Build PHP. To speed up the build, specify the maximum number of jobs using -j:

make -j4

The number of jobs should usually match the number of available cores, which can be determined using nproc.

Testing PHP source code

PHP ships with an extensive test suite, the command make test is used after successful compilation of the sources to run this test suite.

It is possible to run tests using multiple cores by setting -jN in TEST_PHP_ARGS:

make TEST_PHP_ARGS=-j4 test

Shall run make test with a maximum of 4 concurrent jobs: Generally the maximum number of jobs should not exceed the number of cores available.

The qa.php.net site provides more detailed info about testing and quality assurance.

Installing PHP built from source

After a successful build (and test), PHP may be installed with:

make install

Depending on your permissions and prefix, make install may need super user permissions.

PHP extensions

Extensions provide additional functionality on top of PHP. PHP consists of many essential bundled extensions. Additional extensions can be found in the PHP Extension Community Library - PECL.

Contributing

The PHP source code is located in the Git repository at github.com/php/php-src. Contributions are most welcome by forking the repository and sending a pull request.

Discussions are done on GitHub, but depending on the topic can also be relayed to the official PHP developer mailing list [email protected].

New features require an RFC and must be accepted by the developers. See Request for comments - RFC and Voting on PHP features for more information on the process.

Bug fixes don't require an RFC. If the bug has a GitHub issue, reference it in the commit message using GH-NNNNNN. Use #NNNNNN for tickets in the old bugs.php.net bug tracker.

Fix GH-7815: php_uname doesn't recognise latest Windows versions
Fix #55371: get_magic_quotes_gpc() throws deprecation warning

See Git workflow for details on how pull requests are merged.

Guidelines for contributors

See further documents in the repository for more information on how to contribute:

Credits

For the list of people who've put work into PHP, please see the PHP credits page.

php-src's People

Contributors

Stargazers

Watchers

Forkers

katakode iss logicsurge jeffy23 azatoth doap cujo sarang25491 filp jayzeng conf samnrique jpauli ajohnstone smarr dsp crodas skoop mfrost503 flyingmana olleyyy ircmaxell sasezaki dave1010 smalyshev jaytaph clayhinson bvibber rsky cordoval paulbarbu salathe ihumanable elijan stephenfarrell lucassouza1 samfailing janus dmitry-kabanov laruence idler oerdnj wahyusumartha renepardon tony2001 ermantaylan wez michaelhood c77cc anewczs ahyong gauravpadia vg-yokoyama afaolek kprog adampastusiak juil andreia masahide jns-software-llc johannes nikic hjr3 cjbj xandermld ikarl adoy kpx-dev bihicheng slumbers99 lalitjsraks srhazi-imvu taojuntjpp derickr saper epriestley jille reeze saulooliveiro theserb86 gmphp djpate enriquepiatti cpriest rrstoyanov icio jellydoughnut jeffreypalermo rossbg cataphract zhuzhuaicoding tom13beck yangphere campino aholmes colder asterixz kinncj tychay ralphschindler

php-src's Issues

file_put_contents() doesn't create a file, if $data is null or its length is 0

From manual page: https://php.net/function.file-put-contents

showing warning oci8 install through pecl

installing oci8 extension through pecl
sudo pecl -d php_suffix=8.1 install oci8

In function ‘php_oci_column_to_zval’:
/tmp/pear/temp/oci8/oci8.c:1708:72: warning: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘zend_long’ {aka ‘long int’} [-Wformat=]
1708 | php_error_docref(NULL, E_WARNING, "Unable to find LOB descriptor #%d", column->descid->handle);

Datetime created from negative float timestamp is not monotonic (fraction part is added instead of substracted)

Description

The following code:

foreach ([0.4, 0, -0.4, -1, -1.4] as $ts) {
    $date = new DateTime('@' . $ts);
    print_r($date);
}

https://3v4l.org/EWpSj

Diff of current output vs expected:

 DateTime Object
 (
     [date] => 1970-01-01 00:00:00.400000
     [timezone_type] => 1
     [timezone] => +00:00
 )
 DateTime Object
 (
     [date] => 1970-01-01 00:00:00.000000
     [timezone_type] => 1
     [timezone] => +00:00
 )
 DateTime Object
 (
-    [date] => 1970-01-01 00:00:00.400000
+    [date] => 1969-12-31 23:59:59.600000
     [timezone_type] => 1
     [timezone] => +00:00
 )
 DateTime Object
 (
     [date] => 1969-12-31 23:59:59.000000
     [timezone_type] => 1
     [timezone] => +00:00
 )
 DateTime Object
 (
-    [date] => 1969-12-31 23:59:59.400000
+    [date] => 1969-12-31 23:59:58.600000
     [timezone_type] => 1
     [timezone] => +00:00
 )

PHP Version

present since PHP 8.0, should be fixed starting this version

it seems the issue is present only when TS is parsed, getTimestamp / format('u') seems correct across all PHP versions - https://3v4l.org/CjJLf

PHP 8.2: unexpected deprecation for dynamic property set via magic method

Description

The following code:

<?php
class Foo {
    public function __set( $name, $value ) {
        $this->$name = $value;
    }
}

$obj = new Foo();
$obj->test = 'test';

Resulted in this output:

Deprecated: Creation of dynamic property Foo::$test is deprecated in /in/Ds0gN on line 5

https://3v4l.org/Ds0gN/rfc#vgit.master

But I expected this output instead:

No error

I realize that this is an "unconventional" interpretation of how magic methods should be implemented, but at the same time, the RFC explicitly excluded dynamic properties being set via magic methods from the deprecation, so I'd really like to invite some discussion on this.

PHP Version

8.2/nightly

Operating System

No response

mcrypt extension not installable on php 8.1

Description

It looks like the mcrypt extension is not available (yet?) on php 8.1. I couldn't find anything about whether or not there's already efforts underway to change that, hence this issue.

Dockerfile to reproduce:

FROM php:8.1-fpm-buster
RUN pecl install mcrypt-1.0.4

Results in:

pecl/mcrypt is not compatible with PHP version 8.1.0
No valid packages found
install failed

Not sure if this is the right place to report this, the bugtracker for the mcrypt page from pecl told me to open an issue here.

README.md steps for reporting bugs should be updated to mention github issues instead of bugs.php.net

Description

The README currently says:

Bug fixes do not require an RFC but require a bug tracker ticket. Open a
ticket at bugs.php.net and reference the bug id using
#NNNNNN.

Fix #55371: get_magic_quotes_gpc() throws deprecation warning

After removing magic quotes, the get_magic_quotes_gpc function caused a
deprecated warning. get_magic_quotes_gpc can be used to detect the
magic_quotes behavior and therefore should not raise a warning at any time.
The patch removes this warning.

https://wiki.php.net/rfc/github_issues#proposal changed this:

Bug reports and feature requests are filed on GitHub with custom issue forms. Security issues point to bugs.php.net instead, because GitHub currently does not support private issue reports.

It may be useful to also update the README in older minor versions of php that still get released?

https://bugs.php.net/report.php mentions the new steps.

Should creating an issue be optional if there's a GitHub PR fixing the bug?

PHP Version

N/A

Operating System

N/A

Match allow Lambda Functions

Description

Currently match() not allow use lambda functions.

If you try

$test = 'test'; $a = match($test) { 'test' => function($test) { print_r($test); } };

Return
-- Today: Nothing is returned.
-- New implementation: "test"

htmlspecialchars converts ' to #039 on PHP 8.1

Description

The following code:

<?php
echo htmlspecialchars("'");

Resulted in this output on PHP 8.1:

#039

But I expected this output instead (like it works on PHP 7.4 or PHP 8):

htmlspecialchars("'", ENT_SUBSTITUTE) makes it work on all PHP versions.

PHP Version

PHP 8.1

Operating System

No response

When a zip file is left empty, a warning is triggered when script end

Description

The following code:

<?php
$zip = new ZipArchive();
$out = "test.zip";
$zip->open($out, ZipArchive::CREATE | ZipArchive::OVERWRITE);
?>

Resulted in this warning:

PHP Warning:  Unknown: Cannot destroy the zip context: Can't remove file: No such file or directory in Unknown on line 0

But I expected no warnings. This does not happen if flag ZipArchive::OVERWRITE is not specified.

I guess ZipArchive has a registered shutdown function to clear temp zip file when script ends, but that file is created when the first file is added.

PHP Version

PHP 7.x

Operating System

Linux Ubuntu 16.04

Add nullsafe call and index access

Description

Proposal

// Nullsafe call
$res = $optionalClosure?->(1, 2, 3);
$res = ($obj->optionalClosure)?->(1, 2, 3);
// or even?
$res = $obj->optionalClosure?->(1, 2, 3);

// Nullsafe index access
$res = $optionalArray?->[42];
$res = $obj->optionalArray?->[42];

Backward Incompatible Changes

No.

Motivation

Nullsafe access operator ?-> already exists and whole expression consists of existing tokens only;
Syntax like $fn?() and $arr?[42] is not possible since it's a part of already valid syntax ?: ;
Syntax like $fn?‗() and $arr?‗[42] (where ‗ is anything else) most probably will never be added because ‗ is literally something else and leads to inconsistency with $obj?->member.

Open Issues

Must $obj->optionalClosure require parentheses ($obj->optionalClosure)?->() like ($obj->closure)() does?

Future Scope

Probably it may be a small step forward to $array->methods() like [1,2,3]->map(fn ($x) => $x * 10) which itself may be a step to more features like (-42)->abs(), $str->trim()->replace(...)->replace(...).

Can't use inherited constant value via 'self' in enum case value

Description

The following code:

<?php

interface I {
    public const DEFAULT_VALUE = 'gambrinus';
}

enum Beer: string implements I {
    // case GAMBRINUS = I::DEFAULT_VALUE;
     case GAMBRINUS = self::DEFAULT_VALUE;
}
var_dump(Beer::GAMBRINUS);

Resulted in this output:

Fatal error: Enum case value must be compile-time evaluatable in /in/Oqm7f on line 9

But I expected this output instead:
No errors, because commented case GAMBRINUS = I::DEFAULT_VALUE produces no error

PHP Version

8.1.1

Operating System

No response

libphp.so: oci8.php_oci_cleanup_global_handles segfaults at second call

Description

When httpd is terminating, function oci8.php_oci_cleanup_global_handles is sometimes called twice. It symbol ZTS is defined, the second call fails, as the thread specific data-area has already been freed by that time.

Suggested change for php_oci8_int.h
before:

542 #ifdef ZTS
543 #define OCI_G(v) TSRMG(oci_globals_id, zend_oci_globals *, v)
544 #else
545 #define OCI_G(v) (oci_globals.v)
546 #endif

after:

542 #ifdef ZTS
543 # define OCI_GLOBALS TSRMG_BULK(oci_globals_id, zend_oci_globals *)
544 #else
545 # define OCI_GLOBALS (&oci_globals)
546 #endif
547 #define OCI_G(v) (OCI_GLOBALS->v)

Suggested change for oci8.c

before:

246 static void php_oci_cleanup_global_handles(void)
247 {
248     if (OCI_G(err)) {
249         PHP_OCI_CALL(OCIHandleFree, ((dvoid *) OCI_G(err), OCI_HTYPE_ERROR));
250         OCI_G(err) = NULL;
251     }
252
253     if (OCI_G(env)) {
254         PHP_OCI_CALL(OCIHandleFree, ((dvoid *) OCI_G(env), OCI_HTYPE_ENV));
255         OCI_G(env) = NULL;
256     }
257 }

after:

246 static void php_oci_cleanup_global_handles(void)
247 {
248     if (OCI_GLOBALS && OCI_G(err)) {
249         PHP_OCI_CALL(OCIHandleFree, ((dvoid *) OCI_G(err), OCI_HTYPE_ERROR));
250         OCI_G(err) = NULL;
251     }
252
253     if (OCI_GLOBALS && OCI_G(env)) {
254         PHP_OCI_CALL(OCIHandleFree, ((dvoid *) OCI_G(env), OCI_HTYPE_ENV));
255         OCI_G(env) = NULL;
256     }
257 }

or perhaps:

246 static void php_oci_cleanup_global_handles(void)
247 {
248     zend_oci_globals *og= OCI_GLOBALS;
249     if (og) {
250         if (og->err) {
251             PHP_OCI_CALL(OCIHandleFree, ((dvoid *) og->err, OCI_HTYPE_ERROR));
252             og->err = NULL;
253         }
254
255         if (og->env) {
256             PHP_OCI_CALL(OCIHandleFree, ((dvoid *) og->env, OCI_HTYPE_ENV));
257             og->env = NULL;
258         }
259     }
260 }

PHP Version

all with ZTS defined

Operating System

All having pthread

openssl_seal()/_open() is not able to handle gcm cipers, e.g. aes-256-gcm

Description

The following code:

<?php

const NUM_KEYS = 2;

for ($i = 0; $i < NUM_KEYS; $i++) {
  $keys[] = $key = openssl_pkey_new();
  $details = openssl_pkey_get_details($key);
  $pubKeys[] = $details['key'];
  $privKeys[] = openssl_pkey_get_private($key);
}

$data = 'Test Data String';

$ciphers = [
  'aes-256-ctr',
  'aes-256-gcm',
];

foreach ($ciphers as $cipherAlgo) {

  echo "*** TESTING $cipherAlgo ***" . PHP_EOL . PHP_EOL;
  $iv = \random_bytes(openssl_cipher_iv_length($cipherAlgo));
  $result = openssl_seal($data, $sealedData, $sealedKeys, $pubKeys, $cipherAlgo, $iv);

  echo "DATA  :   " . strlen($data) . ' ' . $data . PHP_EOL;
  echo "IV-LEN:   " . openssl_cipher_iv_length($cipherAlgo) . PHP_EOL;
  echo "IV    :   " . bin2hex($iv) . PHP_EOL;
  echo "ENC DATA: " . strlen($sealedData) . ' ' . bin2hex($sealedData) . PHP_EOL;
  echo "RESULT:   " . ($result ? 'true' : 'false') . PHP_EOL;
  echo PHP_EOL;

  // Try decrypt
  foreach ($keys as $i => $key) {
    $decrypted = null; // ;)
    $result = openssl_open($sealedData, $decrypted, $sealedKeys[$i], $key, $cipherAlgo, $iv);
    echo "OPEN:    " . $decrypted . PHP_EOL;
    echo "RESULT:  " . ($result ? 'true' : 'false') . PHP_EOL;

    $result = openssl_private_decrypt($sealedKeys[$i], $unsealedKey, $key);
    echo "UNSEAL:  " . bin2hex($unsealedKey) . PHP_EOL;
    echo "RESULT:  " . ($result ? 'true' : 'false') . PHP_EOL;
    echo "DECRYPT: " . openssl_decrypt($sealedData, $cipherAlgo, $unsealedKey, OPENSSL_RAW_DATA, $iv) . PHP_EOL;
  }

  echo PHP_EOL;
}

Resulted in this output:

$ ./openssl-seal-test.php
*** TESTING aes-256-ctr ***

DATA  :   16 Test Data String
IV-LEN:   16
IV    :   37ff67041864b987b8ec2dcd0879f55e
ENC DATA: 16 2a748aa6ae176caf59cf22868304d65a
RESULT:   true

OPEN:    Test Data String
RESULT:  true
UNSEAL:  8745e00bdf3d88bd1caa8ec0150a9b369eade92e57fc2d26ae283f224c66af14
RESULT:  true
DECRYPT: Test Data String
OPEN:    Test Data String
RESULT:  true
UNSEAL:  8745e00bdf3d88bd1caa8ec0150a9b369eade92e57fc2d26ae283f224c66af14
RESULT:  true
DECRYPT: Test Data String

*** TESTING aes-256-gcm ***

DATA  :   16 Test Data String
IV-LEN:   12
IV    :   033b1589f62bebbfffa3adda
ENC DATA: 16 a531f88fc51e373235c928cae1eb7c29
RESULT:   true

OPEN:    
RESULT:  false
UNSEAL:  600cf6e76cf4ae9d48a881e0649af4682aa9065e63bec515a6b8c9678e398d36
RESULT:  true
DECRYPT: 
OPEN:    
RESULT:  false
UNSEAL:  600cf6e76cf4ae9d48a881e0649af4682aa9065e63bec515a6b8c9678e398d36
RESULT:  true
DECRYPT:

But I expected this output instead:

[...]
*** TESTING aes-256-gcm ***

DATA  :   16 Test Data String
IV-LEN:   12
IV    :   033b1589f62bebbfffa3adda
ENC DATA: 16 a531f88fc51e373235c928cae1eb7c29
RESULT:   true

OPEN:    Test Data String
RESULT:  true
UNSEAL:  600cf6e76cf4ae9d48a881e0649af4682aa9065e63bec515a6b8c9678e398d36
RESULT:  false
DECRYPT: 
OPEN:    Test Data String
RESULT:  true
UNSEAL:  600cf6e76cf4ae9d48a881e0649af4682aa9065e63bec515a6b8c9678e398d36
RESULT:  false
DECRYPT:

The point is that the gcm-cipher need the authentication tag for decrypting the data. This however does not seem to be saved by the openssl_seal() function. I have also placed a question in the OpenSSL issue tracker: openssl/openssl#17235

PHP Version

PHP 8.0.13 (cli) (built: Nov 27 2021 17:17:19) ( ZTS )

Operating System

Gentoo Linux 5.15.6, but should not matter

PHP 7.4 Windows does not include Apache module; VC15 x64 Thread Safe (2021-Nov-16 20:36:58)

Description

The latest version of PHP 7.4 for Windows does not include the Apache dll module (e.g. php7apache2_4.dll). Seriously? 🙄︀

PHP Version

7.4

Operating System

Windows

Incorrect file/line blamed for errors caused by `new SomeClass`

Description

https://3v4l.org/UtjRl

This code should blame line 4, not line 7, since that's where the bogus constant is actually declared.

An example of where this is more problematic:

SomeClass.php

<?php

class SomeClass{
	public const INGOT = IDontExist::INGOT;
}

test.php

<?php

require 'SomeClass.php';

new SomeClass();

Resulted in this output:

Fatal error: Uncaught Error: Class "IDontExist" not found in C:\stable\test4.php:5

But I expected this output instead:

Fatal error: Uncaught Error: Class "IDontExist" not found in C:\stable\SomeClass.php:4

This frequently causes issues with my telemetry because incorrect subsystems are getting blamed for errors when third-party plugin code comes into play.

PHP Version

8.0.13

Operating System

Windows (doubtful if relevant)

RecursiveRegexIterator

Description

For example, __DIR__ contain symfony/console.

<?php
$iterator = new RecursiveRegexIterator(
    new RecursiveDirectoryIterator( __DIR__, RecursiveDirectoryIterator::UNIX_PATHS),
    '/((?:[A-Z][a-zA-Z\d]+\/)*[A-Z][a-zA-Z\d]+)\.php$/',
    RegexIterator::GET_MATCH,
    RegexIterator::USE_KEY
);

Resulted in this output :
Matches all class files only in root directory

But I expected this output instead:
Matches all class files in directory and sub-directories

PHP Version

tested only in PHP 7.3

Operating System

No response

gethostbyaddr still not working in 8.1.1

Description

The following code:

D:\>php -v
PHP 8.1.1 (cli) (built: Dec 15 2021 10:31:43) (ZTS Visual C++ 2019 x64)
Copyright (c) The PHP Group
Zend Engine v4.1.1, Copyright (c) Zend Technologies

D:\>php
<?php
$ip = '192.88.99.1';
var_dump( gethostbyaddr($ip) );
^Z
string(6) "p�(�"

gethostbyaddr still not working for me - I expect not a binary string (I expect empty string '')

PHP Version

PHP 8.1.1

Operating System

Windows 10

Zip::extractTo return false on actual zip file

Description

The following code:

<?php
$zip = new \ZipArchive();
$zip->open('subtitle.zip');

$zip->extractTo('.'); // false

The following file is real zip archive because the open method return true, But the extractTo for some reason return false.

subtitle.zip

Is is possible to find the cause ?

PHP Version

PHP 8.0.0

Operating System

Windows 10

variable losing reference after being set in function

Description

I ran into the below scenario, which might be correct, but it wasn't something I expected, and I wanted to make sure it's the correct behavior.

In the below scenario, why do $a and $b lose their references, but $c retains the reference? I was expecting $a and $b to keep their references like $c does. It's almost like they are being passed by value since they revert to their original value even though they are passed by reference.

Is this correct behavior?

https://3v4l.org/uHaGq

Thanks for your help.

The following code:

<?php
$a = null;
$b = 'test';
$c = null;

$test1 = function() use (&$a, &$b, &$c)
{
	$value = 1;

	$a = &$value;

	$b = &$value;

	$c = array(
		'c' => &$value,
	);
};

$test1();

echo 'a: ';
var_dump($a);
echo 'b: ';
var_dump($b);
echo 'c: ';
var_dump($c);
exit;

Resulted in this output:

a: NULL
b: string(4) "test"
c: array(1) {
  ["c"]=>
  int(1)
}

But I expected this output instead:

a: int(1)
b: int(1)
c: array(1) {
  ["c"]=>
  int(1)
}

PHP Version

PHP 7.4.26

Operating System

No response

Enum could not be converted to string with \array_intersect

Description

The following code:

<?php
class EnumTest extends Unit
{
    public function testArrayIntersect(): void
    {
        $array1 = [TestEnum::TEST];
        $array2 = [TestEnum::TEST];

        $actual = \array_intersect($array1, $array2);
        self::assertCount(1, $actual);
    }
}

enum TestEnum {
    case TEST;
}

Resulted in this output:

Object of class TestEnum could not be converted to string

But I expected the test to pass and internally \array_intersect to be able to work with enums.
The same goes for \array_diff as well, so most probably other array functions, too

PHP Version

PHP 8.1.0

Operating System

php:8.1-fpm-buster

IntlPartsIterator key is wrong for KEY_LEFT/KEY_RIGHT

Description

When creating an IntlPartsIterator with a type of IntlPartsIterator::KEY_LEFT or IntlPartsIterator::KEY_RIGHT, the returned key value is off by one after the first key. So for the string "ABC" you would get 0, 2, 3, instead of the expected 0, 1, 2.

https://3v4l.org/O8374
https://3v4l.org/SNUAX

Originally filed as https://bugs.php.net/bug.php?id=78308
Initial fix attempt #4446

PHP Version

8.1.0

Operating System

No response

Provide more information if writing session with custom save handler fails

Description

The following code:

<?php
class MySessionHandler implements SessionHandlerInterface, SessionUpdateTimestampHandlerInterface {
    public function write($sessID, $sessData, $lock = false): bool {
        if(...){
            return false;
        }

        if(...){
            return false;
        }
        return true;
    }

    public function updateTimestamp($sessID, $val){
        return false;
    }
}

$handler = new MySessionHandler();
session_set_save_handler($handler, true);
session_start();

Resulted in this output:

session_write_close(): Failed to write session data using user defined save handler. (session.save_path: /var/lib/php)

But I expected this output instead:
Having this "fake" session handler, it is impossible to say where session write failed and find the cause. An installed error handler just reports
errorHandler::errorHandler called at [:]

PHP Version

php 8.0.14

Operating System

Mageia 8

add creating object constructor from Jason

Description

add this signature and syntax to PHP core to enjoy developer from fast developing

class MyObject{

__json_constructor( {"AdditionAmount": 0 // default value,
                "CashRegisterDeviceRef": null,
                "CustomerAddressRef": null,
                "DeliveryDesc": "",
                "Description": null,
                "DeviceReceiptCode": null,
                "DocumentRef": 0,
                "EffectiveNetPrice": 11678899,
                "ExcludedPolicies": "",
                "Fee": 11678899.000000,
                "FreeProductBusinessPolicyConditionRowRef": null,
                "FreeProductBusinessPolicyRef": null,
                "HasReservation": false,
              
            })


}

//create new that instance :
$obj= new MyObject({"AdditionAmount": 0,
                "CashRegisterDeviceRef": null,
                "CustomerAddressRef": null,
                "DeliveryDesc": "",
                "Description": null,
                "DeviceReceiptCode": null,
                "DocumentRef": 0,
                "EffectiveNetPrice": 11678899,
                "ExcludedPolicies": "",
                "Fee": 11678899.000000,
                "FreeProductBusinessPolicyConditionRowRef": null,
                "FreeProductBusinessPolicyRef": null,
                "HasReservation": false,
              
            })

Access behavior of SimpleXMLElement changes when element has sibling

Description

This was a hard bug to name, and hence it was hard to search for. I couldn't find a similar report, but if this was already reported, my apologies.

The behavior when accessing a property in a SimpleXMLElement object seems to be influenced by the element having 'siblings' or not. This behavior can be observed with xdebug, as well as using print_r or var_dump.

In the example, when accessing $xml_correct->bar the result is as expected.
However, adding a sibling 'fiz' to element 'bar' in $xml_bug will change the result when accessing $xml_bug->bar. Suddenly, $xml_bug->bar becomes an array. It only does this when $xml_bug->bar is accessed directly. When accessing $xml_bug (or viewing it in the debugger), the 'bar' property appears as expected.

I have tested this in php 7.2, 7.3, 7.4 and 8.0, all show the same result. 3v4l.org shows the same output: https://3v4l.org/0EGhe

This might be a bug in libxml, or php, I'm not sure.

<?php

/**
 * This demos a (possible) bug in PHP, which is observed in PHP 7.2, 7.3, 7.4 and 8.0 (not yet tested on 8.1).
 *
 * Access behaviour of a SimpleXMLElement changes if that node has a sibling.
 *
 * In this example, adding 'fiz' to 'foo' (which is the outer node) will change behaviour when accessing 'bar' (an inner node).
 */
$xml_correct = new SimpleXMLElement('<foo><bar><baz/></bar></foo>');

print_r($xml_correct);        // As expected
print_r($xml_correct->bar);   // As expected, $bar contains property $baz, which is an empty SimpleXMLElement

$xml_bug = new SimpleXMLElement('<foo><bar><baz/></bar><fiz/></foo>');

print_r($xml_bug);        // As expected, property $bar is equal to that observed in print_r($xml_correct)
print_r($xml_bug->bar);   // UNEXPECTED: property $bar is an array, containing one SimpleXMLElement with property $baz which is an empty SimpleXMLElement

Resulted in this output:

// output of print_r($xml_bug->bar):

SimpleXMLElement Object
(
    [0] => SimpleXMLElement Object
        (
            [baz] => SimpleXMLElement Object
                (
                )
        )
)

But I expected this output instead:

// output of print_r($xml_correct->bar):

SimpleXMLElement Object
(
    [baz] => SimpleXMLElement Object
        (
        )

)

PHP Version

7.4 8.0

Operating System

Linux

JIT segmentation fault in PHP 8.1

Description

PHP 8.1.0 + 8.1.1 produces segfault, randomly. Downgrading to 8.0 solves the issue.

--core dump---

BFD: Warning: coredump-php-fpm.30267 is truncated: expected core file size >= 5413076992, found: 35983360.
[New LWP 30267]
[New LWP 1887]
[New LWP 1886]
[New LWP 1888]
Cannot access memory at address 0x7f277dbb3128
Cannot access memory at address 0x7f277dbb3120
Failed to read a valid object file image from memory.
Core was generated by `php-fpm: pool xxxxxx '.

Program terminated with signal 11, Segmentation fault.
#0 0x000055bbf04c0f25 in ZEND_NEW_SPEC_CONST_UNUSED_HANDLER () at /usr/src/debug/php-8.1.1/Zend/zend_vm_execute.h:10137
10137 ce = CACHED_PTR(opline->op2.num);
(gdb) bt
Python Exception <class 'gdb.MemoryError'> Cannot access memory at address 0x7ffdc940e3a8:
(gdb) bt
#0 0x000055bbf04c0f25 in ZEND_NEW_SPEC_CONST_UNUSED_HANDLER () at /usr/src/debug/php-8.1.1/Zend/zend_vm_execute.h:10137
Cannot access memory at address 0x7ffdc940e3a8
(gdb) frame 0
#0 0x000055bbf04c0f25 in ZEND_NEW_SPEC_CONST_UNUSED_HANDLER () at /usr/src/debug/php-8.1.1/Zend/zend_vm_execute.h:10137
10137 ce = CACHED_PTR(opline->op2.num);
(gdb) info frame
Stack level 0, frame at 0x7ffdc940e3b0:
rip = 0x55bbf04c0f25 in ZEND_NEW_SPEC_CONST_UNUSED_HANDLER (/usr/src/debug/php-8.1.1/Zend/zend_vm_execute.h:10137); saved rip Cannot access memory at address 0x7ffdc940e3a8

this is the only available info in the core dump.

PHP Version

PHP 8.1.0 + 8.1.1

Operating System

CentOS 7

gethostbyaddr outputs binary string

Description

PHP 8.1.0
The following code:

<?php
var_dump(gethostbyaddr('89.222.225.0'));
var_dump(gethostbyaddr('192.88.99.1'));

Resulted in this output:

But I expected this output instead:

string(12) "89.222.225.0" (which is still shown when I use PHP 8.0.3)
string(12) "192.88.99.1"

PHP Version

PHP 8.1.0

Operating System

Windows 10 (x64)

Built-in web server prints HTML tags in terminal when address is wrong

Description

I tried to start the built-in web server, but I mistyped the address and then the PHP cli printed HTML tags in the terminal.

- php -S 0.0.0.0:8080
+ php -S 0.0..0:8080

Resulted in this output:

$ php -S 0.0..0:8080
[Wed Dec 15 11:54:39 2021] PHP Warning:  Unknown: php_network_getaddresses: getaddrinfo for 0.0..0 failed: Name or service not known in Unknown on line 0
<br />
<b>Warning</b>:  Unknown: php_network_getaddresses: getaddrinfo for 0.0..0 failed: Name or service not known in <b>Unknown</b> on line <b>0</b><br />
[Wed Dec 15 11:54:39 2021] Failed to listen on 0.0..0:8080 (reason: php_network_getaddresses: getaddrinfo for 0.0..0 failed: Name or service not known)

But I expected this output instead:

$ php -S 0.0..0:8080
[Wed Dec 15 12:01:43 2021] PHP Warning:  Unknown: php_network_getaddresses: getaddrinfo for 0.0..0 failed: Name or service not known in Unknown on line 0
[Wed Dec 15 12:01:43 2021] Failed to listen on 0.0..0:8080 (reason: php_network_getaddresses: getaddrinfo for 0.0..0 failed: Name or service not known)

PHP version:

PHP 8.1.0 (cli) (built: Nov 25 2021 20:48:52) (NTS)
Copyright (c) The PHP Group
Zend Engine v4.1.0, Copyright (c) Zend Technologies
    with Zend OPcache v8.1.0, Copyright (c), by Zend Technologies
    with Xdebug v3.1.1, Copyright (c) 2002-2021, by Derick Rethans
    with blackfire v1.72.0~linux-x64-non_zts81, https://blackfire.io, by Blackfire

PHP Version

8.1.0

Operating System

Debian Bookworm

Cloning a faked SplFileInfo object may segfault

Description

We have some code that worked without any issues on PHP 8.0, but it's failing with a segmentation fault on 8.1.
I'm not sure whether it's a PHPUnit or PHP issue... So please see the code. Mock creation triggers the problem.

Test repo: b1rdex/php-8.1-segfault

I've opened sebastianbergmann/phpunit#4844 for PHPUnit, but the issue was closed with a piece of advice to report it to PHP. So please see the test repo for the details about the issue.

PHP Version

8.1.1

Operating System

Docker official image

SSL context options for in memory cert and pk

Description

It is dangerous to store certificates and privateKeys (without a password) on the disk. Unfortunately, this is needed to load the * .p12 file into stream_context_create.

Please, add the option to set certificates and privateKeys from string content or at least from Data URLs, instead the file path only.

// Using string content
// --------------------

openssl_pkcs12_read(file_get_contents('cert.p12'),$certificates,'pass');

$stream_context = stream_context_create(
[ 'ssl' => [ 'local_cert' => $certificates['cert'],
             'local_pk'   => $certificates['pkey']
           ]
]);

// Using Data URLs
// ---------------

openssl_pkcs12_read(file_get_contents('cert.p12'),$certificates,'pass');

$stream_context = stream_context_create(
[ 'ssl' => [ 'local_cert' => 'data:,'.$certificates['cert'],
             'local_pk'   => 'data:,'.$certificates['pkey']
           ]
]);

// Hack: temp files
// ----------------

openssl_pkcs12_read(file_get_contents('cert.p12'),$certificates,'pass');

file_put_contents('cert.temp',$certificates['cert']);
file_put_contents('pkey.temp',$certificates['pkey']);

$stream_context = stream_context_create(
[ 'ssl' => [ 'local_cert' => 'cert.temp',
             'local_pk'   => 'pkey.temp'
           ]
]);

Expected result: Reading certificates and privateKeys without temp files.
Actual result: Reading certificates and privateKeys only with temp files.

PHP Version

PHP 8.1.1

Operating System

All

Extends PDO, sqliteCreateFunction() and magic method __call()

Description

The following code:

<?php
declare(strict_types=1);

class DBStatement extends PDOStatement
{
    protected $db;

    protected function __construct(PDO $db)
    {
        $this->db = $db;
    }
}
class db extends PDO {
    public function __construct(string $dsn, string $username = null, string $password = null, array $options = [])
    {
        $options += [
            self::ATTR_DEFAULT_FETCH_MODE => self::FETCH_ASSOC,
            self::ATTR_EMULATE_PREPARES   => false,
            self::ATTR_STRINGIFY_FETCHES  => false,
            self::ATTR_ERRMODE            => self::ERRMODE_EXCEPTION,
            self::ATTR_STATEMENT_CLASS    => [DBStatement::class, [$this]],
        ];
        parent::__construct($dsn, $username, $password, $options);
        $this->sqliteCreateFunction('CONCAT', function (...$args) {return \implode('', $args);});
    }

    public function __call(string $name, array $args) /* : mixed */
    {
        // Here is the loading of a class with rarely used methods that are different in implementation for different databases.
        // In a real project, an exception is displayed: 
        // PHP OTHER_ERROR: Method 'sqliteCreateFunction' not found in DB driver. in ...\app\Core\DB\Sqlite.php:[70]
    }
}

$db = new db('sqlite:datadb', '', '', []);
var_dump($db->query('SELECT CONCAT(\'123\', \'456\', \'789\')')->fetch());
exit;

Resulted in this output:

Fatal error: Uncaught PDOException: SQLSTATE[HY000]: General error: 1 no such function: CONCAT in C:\laragon\www\1.php:33 Stack trace: #0 C:\laragon\www\1.php(33): PDO->query('SELECT CONCAT('...') #1 {main} thrown in C:\laragon\www\1.php on line 33

But I expected this output instead:

array(1) { ["CONCAT('123', '456', '789')"]=> string(9) "123456789" }

It turns out that the call to the sqliteCreateFunction() method is intercepted by __call()?

P.S. In my project, I extended PDO, extended PDOStatement and use _call () to call rare methods to change the structure of the database. Now I got such an incident when using SQLite.

PHP Version

PHP 8.1.1-Win32-vs16-x64, 8.1.0, 8.0.8, 7.4.5, 7.3.33

Operating System

Windows 7 x64

Intl Extension: MessageFormatter prepends date with space

Description

The following code:

<?php

$formatter = new \MessageFormatter('en_US', '{0, date, YYYY-ww}');
var_dump($formatter->format([new \DateTimeImmutable]));

$formatter = new \MessageFormatter('en_US', '{0}');
var_dump($formatter->format(['test']));

Resulted in this output:

string(8) " 2021-51"
string(4) "test"

But I expected this output instead:

string(8) "2021-51"
string(4) "test"

See this 3v4l online page. This applies to all messages using date.

PHP Version

7.3.13 - 7.3.33, 7.4.0 - 7.4.27, 8.0.0 - 8.0.13, 8.1.0

Operating System

Any

php_uname doesn't recognise Windows 11 and Windows Server 2019/2022

Description

The following code:

<?php
echo php_uname();
?>

Resulted in this output:

Windows NT HOMEPC 10.0 build 22000 (Windows 10) AMD64

But I expected this output instead:

Windows NT HOMEPC 10.0 build 22000 (Windows 11) AMD64

PHP Version

8.1.1

Operating System

Windows 11

Cli-server mode cannot send Access-Control-Allow-Origin header

Description

The following code:

<?php

header('Access-Control-Allow-Origin: *');
header("Access-Control-Expose-Headers: Content-Disposition");
header('Access-Control-Origin: xxx.hoo.com');
die('header test');

Resulted in this output:

I used the curl -I command to test, but did not output the first header.

But I expected this output instead:

PHP Version

PHP 7.4.3

Operating System

kde neon / Ubuntu 20.04.3 LTS

openssl_x509_checkpurpose_basic.phpt fails, because san-cert.pem expired on 2021-12-11

Description

Unit test ext/openssl/tests/openssl_x509_checkpurpose_basic.phpt fails. Diff file is:

--
int(-1)
int(-1)
int(-1)
036- bool(true)
037- bool(true)
038- bool(true)
039- bool(true)
040- bool(true)
041- bool(true)
042- bool(true)
036+ bool(false)
037+ bool(false)
038+ bool(false)
039+ bool(false)
040+ bool(false)
041+ bool(false)
042+ bool(false)
bool(false)
bool(false)
bool(false)

 int(-1)
 int(-1)
 int(-1)

057+ bool(false)
058+ bool(false)
059+ bool(false)
060+ bool(false)
061+ bool(false)
062+ bool(false)
063+ bool(false)
057- bool(true)
058- bool(true)
059- bool(true)
060- bool(true)
061- bool(true)
062- bool(true)
063- bool(true)

The failing tests are all for the certificate

...
$sert = "file://" . DIR . "/san-cert.pem";
...

and "openssl x509 -text" for this cert reveals:

$ openssl x509 -in /path/to/ext/openssl/tests/san-cert.pem -text
Certificate:
Data:
Version: 3 (0x2)
Serial Number:
86:aa:c6:d6:39:77:08:ed
Signature Algorithm: sha1WithRSAEncryption
Issuer: C=US, ST=MN, L=Minneapolis, OU=Domain Control Validated
Validity
Not Before: Sep 24 08:05:51 2013 GMT
Not After : Dec 11 08:05:51 2021 GMT
...

So this test cert expired a few days ago.

PHP Version

PHP 8.1.1, 8.0.14, 7.4.27

Operating System

Linux (several variants), Solaris Sparc

DateTimeZone::getTransitions() returns insufficient data

Description

The following code:

var_dump(
    (new \DateTime('2203-10-05 15:05:37.180157', new \DateTimeZone('America/Los_Angeles')))
    ->getTimestamp()
);

https://3v4l.org/HpVAr

produces different result on PHP 8.1 vs PHP 7.4/8.0

it is from our failing unit test, what result is correct and any plain to bugfix in either PHP 8.1 or the PHP 8.0 and lower?

PHP Version

PHP 8.1.0

Operating System

linux

string '0.0' evaluates to true

Description

Surely this is not a mistake, but I do not understand why it is so.

Consider the following code:

   if (0) echo "int 0 is true" . "\n";
   if ('0') echo "string '0' is true" . "\n";
   if (0.0) echo "float 0.0 is true" . "\n";
   if ('0.0') echo "string '0.0' is true" . "\n";

Why does the string '0.0' evaluate to true, when its logical value would be false? The string '0' evaluates to false. Why is the same not true for '0.0'?

https://3v4l.org/ZNhbM

PHP Version

PHP 7.X, 8.X

Operating System

All

enums

Description

Hello!

There seems to be no way to get the string representation of an instance of an enum, except with var_export()

enum Foo: string {
   case Bar = 'bar';
   case Baz = 'baz';
}

$bar = Foo::Bar;
$bar_string = var_export($bar, true);

Also, what if I want to find out the backing value of an enum variable?

$backing_value = (new \ReflectionEnumBackedCase($bar, ???))->getBackingValue();

Any ideas?

mysqli_sql_exception->getSqlState()

Description

Hello,

About mysqli_sql_exception:
https://www.php.net/manual/en/class.mysqli-sql-exception.php

What's the point of:

protected string $sqlstate;

if we cannot access it?

I was also trying:

class mysqli_sql_exception_x extends \mysqli_sql_exception {
	public function getSqlState() : string {
		return $this->sqlstate;
	}
}

But then I get:

Class mysqli_sql_exception_x may not inherit from final class (mysqli_sql_exception)

Why doesn't \mysqli_sql_exception contain a getSqlState() method?

I see some workarounds with using Reflection, etc, but that's obviously not ideal.
At least we can still use mysqli_sqlstate, but still...

Thank you very much.

Segfault when recursive call is made

Description

While working on a Symfony project, I made a small mistake in the code's design which lead to segfaults when running the testing suite. After debugging, I managed to narrow the issue, which was awkward. I know this kind of code shouldn't happen, but it's a simple and clear example of mistake that a newcomer might do in php. Just because it "shouldn't" doesn't mean it "doesn't".

Additional context:

scripts were executed from CLI, where the execution time is unlimited
usually, one would get a fatal error because of memory allocation exhaustion.
our CLI config has a memory_limit of 128M.. it's weird why we don't get memory allocation exhaustion
defined classes are in fact Doctrine entities
doctrine / phpunit doesn't have any impact. running the bare minimum code should do it.
executing the code in 3v4l does NOT trigger segfault.

The following code (in a Symfony 5.4 app):

<?php

error_reporting(E_ALL);
ini_set('display_errors', 'on');

class Product {
    public function __construct(
        private array $groups = [],
    ) {
    }

    public function addGroup(Group $group): void {
        $group->addProduct($this);
        $this->groups[] = $group;
    }
}

class Group {
    public function __construct(
        private array $products = [],
    ) {
    }

    public function addProduct(Product $product): void {
        $product->addGroup($this);
        $this->products[] = $product;
    }
}

$product = new Product();
$group = new Group();

$product->addGroup($group);

Resulted in this output:

$ php test.php; echo $?
139

(139 = exit code for segmentation fault)

But I expected this output instead:

Fatal error: ...something suggestive about recursion...

php -i output

PHP Version

8.0.13

Operating System

Debian Bullseye (Docker container)

Inherited final constant causes fatal error (if the same interface is inherited multiple times)

Description

The following code:

<?php

interface EntityInterface {
    final public const TEST = 'this';
}

interface KeyInterface extends EntityInterface {
}

interface StringableInterface extends EntityInterface {
}

class SomeTestClass implements KeyInterface, StringableInterface {
}

Resulted in this output:

Fatal error: EntityInterface::TEST cannot override final constant EntityInterface::TEST in /in/UiiTf on line 13

Process exited with code 255.

But I expected this output instead:

No error

PHP Version

PHP 8.1.0

Operating System

Ubuntu 20.04

Problem with mysqli_options and persistent connections in mysqli.

Description

According to the documentation https://www.php.net/manual/en/mysqli.options.php
mysqli_options() should be called after mysqli_init() and before mysqli_real_connect().

But, when using persistent connections and the connection in the pool is broken,
when creating a new connection, the MYSQLI_OPT_INT_AND_FLOAT_NATIVE option is lost.

Setting the option AFTER calling real_connect solves the problem

The following code:

<?php
\mysqli_report(MYSQLI_REPORT_ERROR);
function connect(): mysqli {
	$mysqli = mysqli_init();
	$mysqli->options(MYSQLI_OPT_INT_AND_FLOAT_NATIVE, 1);
	$mysqli->real_connect('p:127.0.0.1', 'user', 'password');

	return $mysqli;
}

$mysqli1 = connect();
var_dump($mysqli1->query("SELECT 1")->fetch_row()[0]); // int
@$mysqli1->query("KILL {$mysqli1->thread_id}");
unset($mysqli1);

$mysqli2 = connect();
var_dump($mysqli2->query("SELECT 2")->fetch_row()[0]); // string, but int is expected

// next ok
$mysqli2->options(MYSQLI_OPT_INT_AND_FLOAT_NATIVE, 1);
var_dump($mysqli2->query("SELECT 3")->fetch_row()[0]); // int

Resulted in this output:

int(1)
string(1) "2"
int(3)

But I expected this output instead:

int(1)
int(2)
int(3)

PHP Version

PHP 8.0.14, 8.1.1

Operating System

Debian 9.13, Ubuntu 21.10

Wrong error message if Enum doesn't correctly implement an interface

Description

The following code:

<?php
interface A {
    public function a(): void;
}

enum B implements A {
}

Resulted in this output:

PHP Fatal error:  Class E contains 1 abstract method and must therefore be declared abstract or implement the remaining methods (I::a) in Command line code on line 1

But I expected this output instead:

PHP Fatal error:  Enum E contains 1 abstract method and must therefore be declared abstract or implement the remaining methods (I::a) in Command line code on line 1

Patch:

diff --git a/Zend/zend_API.c b/Zend/zend_API.c
index 3ef291c315..04d5728ef1 100644
--- a/Zend/zend_API.c
+++ b/Zend/zend_API.c
@@ -4789,6 +4789,8 @@ ZEND_API ZEND_COLD const char *zend_get_object_type(const zend_class_entry *ce)
                return "trait";
        } else if (ce->ce_flags & ZEND_ACC_INTERFACE) {
                return "interface";
+       } else if (ce->ce_flags & ZEND_ACC_ENUM) {
+               return "enum";
        } else {
                return "class";
        }
diff --git a/Zend/zend_inheritance.c b/Zend/zend_inheritance.c
index 82af3d8afa..002da24383 100644
--- a/Zend/zend_inheritance.c
+++ b/Zend/zend_inheritance.c
@@ -2324,9 +2324,13 @@ void zend_verify_abstract_class(zend_class_entry *ce) /* {{{ */
        } ZEND_HASH_FOREACH_END();

        if (ai.cnt) {
+               char* kind = strdup(zend_get_object_type(ce));
+               kind[0] = toupper(kind[0]);
+
                zend_error_noreturn(E_ERROR, !is_explicit_abstract
-                       ? "Class %s contains %d abstract method%s and must therefore be declared abstract or implement the remaining methods (" MAX_ABSTRACT_INFO_FMT MAX_ABSTRACT_INFO_FMT MAX_ABSTRACT_INFO_FMT ")"
-                       : "Class %s must implement %d abstract private method%s (" MAX_ABSTRACT_INFO_FMT MAX_ABSTRACT_INFO_FMT MAX_ABSTRACT_INFO_FMT ")",
+                       ? "%s %s contains %d abstract method%s and must therefore be declared abstract or implement the remaining methods (" MAX_ABSTRACT_INFO_FMT MAX_ABSTRACT_INFO_FMT MAX_ABSTRACT_INFO_FMT ")"
+                       : "%s %s must implement %d abstract private method%s (" MAX_ABSTRACT_INFO_FMT MAX_ABSTRACT_INFO_FMT MAX_ABSTRACT_INFO_FMT ")",
+                       kind,
                        ZSTR_VAL(ce->name), ai.cnt,
                        ai.cnt > 1 ? "s" : "",
                        DISPLAY_ABSTRACT_FN(0),

PHP Version

8.1.*

Operating System

Irrelevant

Add mysqli_rollback_to

Description

In PHP 5.5 mysqli::savepoint and mysqli::release_savepoint were added, but not rollback_to. Maybe there is some database engine where savepoints have multiple purposes, but in mariadb the only use of savepoints is to call query ROLLBACK TO name manually.

For completion: existing rollback with $name is doing nothing (only adding comment) at least in mariadb/mysql.

Method signature

class mysqli { /* ... */
  public function rollback_to(string $name) : bool;
}

Sample test sequence

mysqli_query($link, 'DROP TABLE IF EXISTS test');
mysqli_query($link, 'CREATE TABLE test(id INT) ENGINE = InnoDB');

mysqli_begin_transaction($link);
mysqli_query($link, 'INSERT INTO test(id) VALUES (1)');
mysqli_savepoint($link, 'foo');
mysqli_query($link, 'INSERT INTO test(id) VALUES (2)');
mysqli_rollback_to($link, 'foo');
mysqli_commit($link);
$res = mysqli_query($link, "SELECT * FROM test");
var_dump($res->num_rows); // int(1)

move reference to pcre.h out of spl_iterators.h

Description

Currently spl_iterators.h includes php_pcre.h so that it can be reference in the spl_dual_it_object struct but only as a pointer to a .

This makes a dependency on the pcre2.h file for extensions that include the spl_iterators.h. Apparently this causes problems for people installing extensions that include spl_iteterators header, presumably due to the common swapping out of PCRE bundled vs system?

Would it be possible to make the header not depend on the php_pcre.h file, either through moving the relevant parts to a spl-private file or using an void pointer?

btw feel free to close if this is non-trivial. I'd missed the fact that as the countable_ce was moved from spl to zend, I no longer need to include the spl header anyway....

Incorrect return types for hash() and hash_hmac()

Description

I believe the stub file entries for hash() and hash_hmac() to be incorrect. Their return types are both specified as string|false, but starting with PHP 8.0 they in fact never return false, they throw a fatal error where they returned false before PHP 8.0. In the case of hash() the only two RETURN_FALSE statements (1, 2) are within if (isfilename)blocks, e.g. the code path used exclusively by hash_file(). It's similar for hash_hmac().

Does that make sense or did I miss something?

PHP Version

PHP >= 8.0

Operating System

All OSes

DOMNode Reflection does not return properties

Description

The following code:

<?php

$c = new ReflectionClass("DOMNode");
$ps = $c->getProperties(ReflectionProperty::IS_PUBLIC | ReflectionProperty::IS_PROTECTED);

var_dump(count($ps));

foreach($ps as $p) {

    var_dump($p->name);
    var_dump($p->getType());
}

Resulted in this output:

int(0)

But I expected this output instead:
a list of all properties, as documented in https://www.php.net/manual/en/class.domnode.php

see https://sandbox.onlinephpfunctions.com/

PHP Version

PHP 8.0.12

Operating System

macos

Inconsistent argument name in `hash-hmac-file` and `hash-file`

From manual page: https://php.net/function.hash-hmac-file

string $data, should be string $filename, to have the same naming scheme as hash-file().

Allow CData to be passed as valid argument to primitive ffi functions

Description

Currently one can only pass a float (C) from the PHP's float, which is a double.

When using a lot of floats with C libraries, this means that float's are constantly converted from float's to double' and double's to floats.

I hereby would like to request the option to send the CData class to FFI bound functions, to bypass this type juggling

I guess this would also mean that there should be an option to get the return as the "raw" CData value, which I have no idea how to fully implement from a user standpoint ($ffi->atan->raw(0.5) or something?)

A simple code example (that has no bearing on reality) that displays the "issue" would be the following:

<?php

$c = FFI::cdef('float atanf (float x);', 'libm.so.6');
$x = $c->atanf(0.4);
$z = FFI::new('float');
$z->cdata = $x;
$y = $c->atanf($z);
echo $y.$x."\n";

php will now simply throw a warning "PHP Warning: Object of class FFI\CData could not be converted to float", and use 1.0 instead

enums: is_case() function

Description

There is currently no way to determine if an expression or variable is of type "enum" (or I don't know how).

Since "enum" is a new data type, it would be desirable for there to be a function to know if an expression or variable is a case of an enumeration.

Such a function could be:

is_case()
is_enum()
is_enum_case()
is_enumcase()

array_column return empty array if file encoding contains BOM

Description

two identical file one with encoding UTF-8-BOM other with encoding UTF-8, both read successfully but when we try to get a column, the second one (with BOM) return empty array.

PHP Version

PHP8.0

Operating System

Ubuntu 20.04