Added PHPStan, apply changes for level 3 #3025

Majkl578 · 2018-02-17T01:19:37Z

Mostly annotation changes/fixes to match what the methods/properties really are used for. Added few asserts and instanceof.
Not really sure if you want to consider this as BC issue - I'd hope not, but if so we can retarget to develop instead.

Level 3, not going further as it requires actually changing code itself, that should follow later in develop with strict typing.

There is a PHP notice currently since Travis doesn't have OCI driver:

Notice: Use of undefined constant OCI_DEFAULT - assumed 'OCI_DEFAULT'

morozov

I'd say, most if not all places where false/bool is added as part of the union type are not by design and should be fixed in a BC way (if possible) or silenced instead of being documented.

morozov · 2018-02-17T02:42:01Z

lib/Doctrine/DBAL/Connection.php

@@ -402,7 +403,7 @@ private function detectDatabasePlatform()
    {
        $version = $this->getDatabasePlatformVersion();

-        if (null !== $version) {
+        if ($version !== null && $this->_driver instanceof VersionAwarePlatformDriver) {


Wouldn't it be enough to check if the driver is not null? Type hint should help static analysis w/o the run time check.

$this->_platform is typed as AbstractPlatform, but createDatabasePlatformForVersion() is only defined in some drivers implementing VersionAwarePlatformDriver.
I at least replaced this by assert() inside the if since Connection::getDatabasePlatformVersion() returns null for non-version-aware drivers.

morozov · 2018-02-17T02:42:55Z

lib/Doctrine/DBAL/Connection.php

@@ -419,7 +420,7 @@ private function detectDatabasePlatform()
     * or the underlying driver connection cannot determine the platform
     * version without having to query it (performance reasons).
     *
-     * @return string|null
+     * @return string|bool|null


I'd say it's a bug and should be fixed, not documented.

Hmm this is because SqlAnywhereConnection::getServerVersion() calls ResultStatement::fetchColumn() which returns string|bool. :| Replaced with assert there and removed here.

morozov · 2018-02-17T02:43:08Z

lib/Doctrine/DBAL/Connection.php

@@ -478,7 +479,7 @@ private function getDatabasePlatformVersion()
    /**
     * Returns the database server version if the underlying driver supports it.
     *
-     * @return string|null
+     * @return string|bool|null


Boolean is a bug.

Same as above.

lib/Doctrine/DBAL/Connections/MasterSlaveConnection.php

@@ -85,7 +86,7 @@ class MasterSlaveConnection extends Connection
    /**
     * Master and slave connection (one of the randomly picked slaves).
     *
-     * @var \Doctrine\DBAL\Driver\Connection[]
+     * @var DriverConnection[]|null[]


morozov · 2018-02-17T02:47:16Z

lib/Doctrine/DBAL/Platforms/AbstractPlatform.php

@@ -2716,7 +2716,7 @@ public function getCurrentTimestampSQL()
     *
     * @param int $level
     *
-     * @return string
+     * @return string|int


Should it be fixed or silenced instead? Doesn't look like it's by design (the SQLite platform impelementation).

I can change those ints to strings, but I really have no idea of consequences (either in DBAL or for consumers). SqlAnywherePlatoform returns ints too... Maybe better leave this to 3.x?

I'm all for leaving it for 3.x where we could fix the actual code instead of complicating the API documentation.

Please ignore instead.

morozov · 2018-02-17T02:48:38Z

lib/Doctrine/DBAL/Schema/AbstractSchemaManager.php

@@ -423,7 +423,7 @@ public function dropView($name)
     *
     * @param string $database The name of the database to create.
     *
-     * @return void
+     * @return void|bool


Looks like a bug. Mixing something with void doesn't make sense.

Yep, removed. But OracleSchemaManager::createDatabase() returns true so I removed that one. Should be documented in UPGRADING?

morozov · 2018-02-17T02:51:47Z

lib/Doctrine/DBAL/Schema/AbstractSchemaManager.php

@@ -917,7 +917,7 @@ protected function _getPortableTablesList($tables)
    /**
     * @param array $table
     *
-     * @return array
+     * @return mixed[]|string


Can elements of the array be anything else than string? string[]|string would make more sense.

No idea since $table is uselessly types as array. If you are certain, I can change it.

I'd look at it from the caller's standpoint: whatever callers can consume by design (string[] or string) should be valid return, the rest shouldn't.

Please figure out the correct type from the API standpoint and ignore the deviations.

------ ------------------------------------------------------------------------------------------------------------------------ Line lib/Doctrine/DBAL/Schema/OracleSchemaManager.php ------ ------------------------------------------------------------------------------------------------------------------------ 104 Method Doctrine\DBAL\Schema\OracleSchemaManager::_getPortableTableDefinition() should return array but returns string. ------ ------------------------------------------------------------------------------------------------------------------------ ------ ---------------------------------------------------------------------------------------------------------------------------- Line lib/Doctrine/DBAL/Schema/PostgreSqlSchemaManager.php ------ ---------------------------------------------------------------------------------------------------------------------------- 222 Method Doctrine\DBAL\Schema\PostgreSqlSchemaManager::_getPortableTableDefinition() should return array but returns string. ------ ---------------------------------------------------------------------------------------------------------------------------- ------ --------------------------------------------------------------------------------------------------------------------------- Line lib/Doctrine/DBAL/Schema/SQLServerSchemaManager.php ------ --------------------------------------------------------------------------------------------------------------------------- 210 Method Doctrine\DBAL\Schema\SQLServerSchemaManager::_getPortableTableDefinition() should return array but returns string.

Honestly, no idea. Some managers return $table as is and some return a string (schema name + table name).
This definitely looks like a bad design (git-blame indicates addition back in 2008-2011).
Help? :)

reverting and putting it into ignores for now, unlikely to be properly fixable in 2.x

morozov · 2018-02-17T02:52:38Z

lib/Doctrine/DBAL/Schema/Schema.php

@@ -73,7 +73,7 @@ class Schema extends AbstractAsset
    protected $_sequences = [];

    /**
-     * @var \Doctrine\DBAL\Schema\SchemaConfig
+     * @var \Doctrine\DBAL\Schema\SchemaConfig|bool


A nullable type would make more sense here.

Can't be done in 2.x I guess...

Let's ignore it and fix instead of introducing an intermediate incorrect type.

Please ignore.

morozov · 2018-02-17T02:53:18Z

lib/Doctrine/DBAL/Sharding/PoolingShardConnection.php

@@ -65,17 +66,17 @@
 class PoolingShardConnection extends Connection
 {
    /**
-     * @var array
+     * @var DriverConnection[]|null


The variable could be initialized with an empty array instead.

Probably ok here and for $connections, changed (close() now sets empty array instead of null).

carusogabriel · 2018-02-17T18:32:58Z

Already prepared changes for tests/ as well https://github.com/carusogabriel/dbal/tree/phpstan-tests

Majkl578 · 2018-02-17T19:15:41Z

lib/Doctrine/DBAL/Connection.php

@@ -402,7 +403,7 @@ private function detectDatabasePlatform()
    {
        $version = $this->getDatabasePlatformVersion();

-        if (null !== $version) {
+        if ($version !== null && $this->_driver instanceof VersionAwarePlatformDriver) {


$this->_platform is typed as AbstractPlatform, but createDatabasePlatformForVersion() is only defined in some drivers implementing VersionAwarePlatformDriver.
I at least replaced this by assert() inside the if since Connection::getDatabasePlatformVersion() returns null for non-version-aware drivers.

Majkl578 · 2018-02-17T19:22:48Z

lib/Doctrine/DBAL/Connection.php

@@ -419,7 +420,7 @@ private function detectDatabasePlatform()
     * or the underlying driver connection cannot determine the platform
     * version without having to query it (performance reasons).
     *
-     * @return string|null
+     * @return string|bool|null


Hmm this is because SqlAnywhereConnection::getServerVersion() calls ResultStatement::fetchColumn() which returns string|bool. :| Replaced with assert there and removed here.

Majkl578 · 2018-02-17T19:22:55Z

lib/Doctrine/DBAL/Connection.php

@@ -478,7 +479,7 @@ private function getDatabasePlatformVersion()
    /**
     * Returns the database server version if the underlying driver supports it.
     *
-     * @return string|null
+     * @return string|bool|null


Same as above.

Majkl578 · 2018-02-17T19:23:38Z

lib/Doctrine/DBAL/Connection.php

@@ -1132,7 +1133,7 @@ public function getTransactionNestingLevel()
    /**
     * Fetches the SQLSTATE associated with the last database operation.
     *
-     * @return int The last error code.
+     * @return string|bool|null The last error code.


This on the other hand is not a bug, some implementors explicitly return false (SQLSrvConnection::errorCode()).

DBAL should eliminate this discrepancy. The clients will not have to know the low level details.

Majkl578 · 2018-02-17T19:25:13Z

lib/Doctrine/DBAL/Connection.php

@@ -1163,7 +1164,7 @@ public function errorInfo()
     *
     * @param string|null $seqName Name of the sequence object from which the ID should be returned.
     *
-     * @return string A string representation of the last inserted ID.
+     * @return string|int|bool A string representation of the last inserted ID.


Same here (OCI8Connection::lastInsertId()).

lib/Doctrine/DBAL/Connections/MasterSlaveConnection.php

@@ -85,7 +86,7 @@ class MasterSlaveConnection extends Connection
    /**
     * Master and slave connection (one of the randomly picked slaves).
     *
-     * @var \Doctrine\DBAL\Driver\Connection[]
+     * @var DriverConnection[]|null[]


Majkl578 · 2018-02-17T19:30:14Z

lib/Doctrine/DBAL/Platforms/AbstractPlatform.php

@@ -2716,7 +2716,7 @@ public function getCurrentTimestampSQL()
     *
     * @param int $level
     *
-     * @return string
+     * @return string|int


I can change those ints to strings, but I really have no idea of consequences (either in DBAL or for consumers). SqlAnywherePlatoform returns ints too... Maybe better leave this to 3.x?

Majkl578 · 2018-02-17T19:30:56Z

lib/Doctrine/DBAL/Platforms/SqlitePlatform.php

@@ -899,7 +899,7 @@ public function getAlterTableSQL(TableDiff $diff)
    /**
     * @param \Doctrine\DBAL\Schema\TableDiff $diff
     *
-     * @return array|bool
+     * @return string[]|false


As much as I hate using false as type, there is no better way around here. Usually these cases are something to replace with null for 3.x.

Majkl578 · 2018-02-17T19:32:37Z

lib/Doctrine/DBAL/Schema/AbstractSchemaManager.php

@@ -423,7 +423,7 @@ public function dropView($name)
     *
     * @param string $database The name of the database to create.
     *
-     * @return void
+     * @return void|bool


Yep, removed. But OracleSchemaManager::createDatabase() returns true so I removed that one. Should be documented in UPGRADING?

Majkl578 · 2018-02-17T19:46:15Z

lib/Doctrine/DBAL/Sharding/PoolingShardConnection.php

@@ -65,17 +66,17 @@
 class PoolingShardConnection extends Connection
 {
    /**
-     * @var array
+     * @var DriverConnection[]|null


Probably ok here and for $connections, changed (close() now sets empty array instead of null).

carusogabriel · 2018-04-07T18:36:33Z

Any updates on this one? :)

Majkl578 · 2018-05-26T21:59:52Z

I'll wait for PHPStan 0.10 (to be released next month) before finishing this.

Majkl578 · 2018-06-26T17:25:09Z

@morozov @carusogabriel Updated with PHPStan 0.10 and level 3.

@morozov Please re-review and point out changes that should be explicitly ignored instead of documented in phpDoc. :) I added some new issues that appeared with 0.9->0.10 update.

I also added phpstorm-stubs as a polyfill for some fancier extensions.

carusogabriel · 2018-06-26T17:36:29Z

lib/Doctrine/DBAL/Driver/IBMDB2/DB2Connection.php

@@ -74,6 +75,7 @@ public function __construct(array $params, $username, $password, $driverOptions
     */
    public function getServerVersion()
    {
+        /** @var stdClass $serverInfo */
        $serverInfo = db2_server_info($this->_conn);


Does this function always return a stdClass? Can't we force it in PHPStan stubs?

I didn't actually check whether it is stdClass or not, it returns some object crate: http://php.net/db2_server_info

oh, and false, how unusual ;)

carusogabriel · 2018-06-26T17:37:31Z

lib/Doctrine/DBAL/Driver/OCI8/OCI8Statement.php

@@ -300,6 +301,7 @@ public function bindParam($column, &$variable, $type = ParameterType::STRING, $l
        $column = $this->_paramMap[$column] ?? $column;

        if ($type === ParameterType::LARGE_OBJECT) {
+            /** @var OCI_Lob $lob */


PHPStan stubs wrong?

It's actually OCI-Lob.

hmm, that's not a valid class name 🙈 moving to ignore rather than typehinting hack

morozov · 2018-06-26T18:09:52Z

lib/Doctrine/DBAL/Driver/Connection.php

@@ -99,7 +99,7 @@ public function rollBack();
    /**
     * Returns the error code associated with the last operation on the database handle.
     *
-     * @return string|null The error code, or null if no operation has been run on the database handle.
+     * @return string|bool|null The error code, or null if no operation has been run on the database handle.


bool|null should be eliminated, not document.

morozov · 2018-06-26T19:18:11Z

lib/Doctrine/DBAL/Driver/Mysqli/MysqliStatement.php

@@ -287,7 +288,9 @@ public function fetch($fetchMode = null, $cursorOrientation = \PDO::FETCH_ORI_NE
            return $this->fetchColumn();
        }

+        /** @var mixed[]|false $values */


Can we change the _fetch()'es phpDoc instead?

lib/Doctrine/DBAL/Schema/ForeignKeyConstraint.php

@@ -333,7 +333,7 @@ public function getOptions()
     * Returns the referential action for UPDATE operations
     * on the referenced table the foreign key constraint is associated with.
     *
-     * @return string|null
+     * @return string|false|null


lib/Doctrine/DBAL/Schema/ForeignKeyConstraint.php

@@ -344,7 +344,7 @@ public function onUpdate()
     * Returns the referential action for DELETE operations
     * on the referenced table the foreign key constraint is associated with.
     *
-     * @return string|null
+     * @return string|false|null


morozov · 2018-06-26T19:19:59Z

lib/Doctrine/DBAL/Schema/ForeignKeyConstraint.php

@@ -357,7 +357,7 @@ public function onDelete()
     *
     * @param string $event Name of the database operation/event to return the referential action for.
     *
-     * @return string|null
+     * @return string|false|null


false|null is undesirable.

adding to ignore, but it's a bug, false is explicitly returned here :/

morozov · 2018-06-26T19:21:30Z

lib/Doctrine/DBAL/Schema/Table.php

@@ -62,7 +62,7 @@ class Table extends AbstractAsset
    protected $_indexes = [];

    /**
-     * @var string
+     * @var string|bool


string|false? Or ignore and rework as string|null in 3.0?

will ignore

morozov · 2018-06-26T19:25:06Z

lib/Doctrine/DBAL/Sharding/PoolingShardConnection.php

     */
-    private $connections;
+    private $connections = [];


mixed[] $connections doesn't look good. As far as I understand, it's not "connections" but rather "connection parameters". Given it's a private property, can it be renamed accordingly?

morozov · 2018-06-26T19:26:59Z

phpstan.neon.dist

+        - '~^Method Doctrine\\DBAL\\Query\\QueryBuilder::execute\(\) should return Doctrine\\DBAL\\Driver\\Statement\|int but returns Doctrine\\DBAL\\Driver\\ResultStatement\.\z~'
+
+        # http://php.net/manual/en/pdo.sqlitecreatefunction.php
+        - '~^Call to an undefined method Doctrine\\DBAL\\Driver\\PDOConnection::sqliteCreateFunction\(\)\.\z~'


Should this be reported somewhere upstream?

morozov · 2018-06-26T19:28:05Z

phpstan.neon.dist

+        - '~^Method Doctrine\\DBAL\\Platforms\\AbstractPlatform::getListTableForeignKeysSQL\(\) invoked with \d+ parameters, 1 required\.\z~'
+
+        # this method is undeclared, relies on PDO ancestor in implementors
+        - '~^Call to an undefined method Doctrine\\DBAL\\Driver\\Statement::nextRowset\(\)\.\z~'


Can this be replaced with assert($statement instanceof PDOStatement) in the code?

morozov · 2018-06-26T19:30:16Z

phpstan.neon.dist

+    reportUnmatchedIgnoredErrors: false
+    ignoreErrors:
+        # extension not available
+        - '~^(Used )?(Function|Constant) sasql_\S+ not found\.\z~i'


Could we generate a stub in the longer term? As long as we officially support the driver.

I have no idea. PhpStorm stubs don't have sasql and I didn't find any documentation either, nor PECL extension. Working with SAP SQL Anywhere is really hard. :(

For future reference, I found a few working links at https://archive.sap.com/documents/docs/DOC-40537. I'll try to drop the Windows DLL into PHP 5 and see if I can dump the signatures. Sometime.

The link lead me to: http://dcx.sap.com/index.html#sa160/en/dbprogramming/php-api.html
This looks good enough to build the stubs. Maybe i.e. @carusogabriel may want to help here? :)

But we should probably consider dropping the platform/driver altogether.
Given that DBAL 3 will require PHP 7.2, and SQL Anywhere is barely keeping up (the link only lists PHP 5 modules)...

Nevermind, did this during Slack outage. 😛
JetBrains/phpstorm-stubs#382

morozov · 2018-06-26T19:30:30Z

phpstan.neon.dist

+        - '~^Method Doctrine\\DBAL\\Driver\\PDOConnection::\w+\(\) should return Doctrine\\DBAL\\Driver\\Statement but returns PDOStatement\.\z~'
+
+        # may not exist when pdo_sqlsrv is not loaded
+        - '~^Access to undefined constant PDO::SQLSRV_ENCODING_BINARY\.\z~'


Could we generate a stub in the longer term?

Not sure if possible for a class constant...

It's missing from the PDO stub as well as many other driver-specific symbols. I'll file a PR for JetBrains/phpstorm-stubs later.

Filed JetBrains/phpstorm-stubs#381.

Not sure if possible for a class constant…

Oh, I see what you're saying. The class stub will not be loaded for PDO if the extension itself is loaded, even if it's compiled w/o the missing constants. PDO is so nicely designed.

It's definitely missing in the stubs, but it is also missing in PHP itself unless oci extension is loaded. Who said multiple inheritance in PHP is not possible? 😁

morozov · 2018-06-26T19:33:04Z

tests/phpstan-polyfill.php

+
+(function () : void {
+    static $map = [
+        'db2' => __DIR__ . '/../vendor/jetbrains/phpstorm-stubs/ibm_db2/ibm_db2.php',


Is it really 'db2', not 'ibm_db2'?

var_dump(extension_loaded('db2')); // bool(false) var_dump(extension_loaded('ibm_db2')); // bool(true)

morozov · 2018-06-26T19:40:02Z

tests/phpstan-polyfill.php

+        'db2' => __DIR__ . '/../vendor/jetbrains/phpstorm-stubs/ibm_db2/ibm_db2.php',
+        'mysqli' => __DIR__ . '/../vendor/jetbrains/phpstorm-stubs/mysqli/mysqli.php',
+        'oci8' => __DIR__ . '/../vendor/jetbrains/phpstorm-stubs/oci8/oci8.php',
+        'sqlsrv' => __DIR__ . '/../vendor/jetbrains/phpstorm-stubs/sqlsrv/sqlsrv.php',


Given that the stub files are nicely organized, in order to avoid copy/paste and simplify maintenance, can this be reworked as:

$extensions = ['ibm_db2', ...]; foreach ($extensions as $extension) { if (extension_loaded($extension)) { continue; } require sprintf(__DIR__ . '/../vendor/jetbrains/phpstorm-stubs/%1$s/%1$s.php', $extension); }

morozov · 2018-06-27T02:17:39Z

phpstan.neon.dist

+        - '~^Method Doctrine\\DBAL\\Driver\\SQLSrv\\SQLSrvConnection::errorCode\(\) should return string\|null but returns false\.\z~'
+
+        # actually OCI-Lob - not even a valid class name...
+        - '~^Call to an undefined method object::writeTemporary\(\)\.\z~'


Hmm… all generated stubs have it as OCI_Lob (probably to make it valid), shouldn't then PHPStan just use the "fixed" names? The oci_new_descriptor() is stubbed as oci_new_descriptor() : OCI_Lob, and OCI_Lob::writeTemporary() is also stubbed, so I'd expect not seeing this error.

phpstan/phpstan#1123

Not really sure what to think about this. :)

Majkl578 · 2018-06-27T15:22:33Z

CI passed. 💚

Majkl578 · 2018-06-27T18:31:05Z

@morozov Do you want this in 2.8 or wait for release and then merge into 2.9?

morozov · 2018-06-27T19:01:52Z

@Majkl578 I understand it as a non-breaking low-risk, mostly documentation improvement. Unless it's not true, I'm fine with having it in 2.8.0.

Majkl578 · 2018-06-27T19:52:34Z

Yup, and no tests changed so it should be fine. Let's ship it. 🚢

Ocramius · 2018-06-28T06:14:26Z

👍

@morozov

Release v2.8.0 [![Build Status](https://travis-ci.org/doctrine/dbal.svg?branch=v2.8.0)](https://travis-ci.org/doctrine/dbal) This is a minor release of Doctrine DBAL that aggregates over 30 fixes and improvements developed over the last 3 months. This release includes all changes of the 2.7.x series, as well as feature additions and improvements that couldn’t land in patch releases. **Backwards Compatibility Breaks** This doesn't contain any intentional Backwards Compatibility (BC) breaks. **Dependency Changes** * The dependency on [doctrine/common](https://github.com/doctrine/common) is removed. DBAL now depends on [doctrine/cache](https://github.com/doctrine/cache) and [doctrine/event-manager](https://github.com/doctrine/event-manager) instead. Please see details in the [UPGRADE.md](UPGRADE.md) documentation. **Deprecations** * The usage of binary fields whose length exceeds the maximum field size on a given platform is deprecated. Please use binary fields of a size which fits all target platforms, or use blob explicitly instead. * The usage of DB-generated UUIDs is deprecated. Their format is inconsistent across supported platforms and therefore the feature is not portable. Use a PHP library (e.g. [ramsey/uuid](https://packagist.org/packages/ramsey/uuid)) to generate UUIDs on the application side. **New features** * Initial support of MySQL 8. * Initial support of PostgreSQL 11. * Ability to evaluate arbitrary SQL expressions via `AbstractPlatform::getDummySelectSQL()`. **Improvements and Fixes** * Improved support of binary fields on Oracle and IBM DB2. * Improved SQL Server configuration capabilities via both `sqlsrv` and `pdo_sqlsrv`. * Improved handling of `AUTOINCREMENT`ed primary keys in SQLite. * Integration tests are run against IBM DB2 on Travis CI. * Code coverage is collected for the Oracle platform on continuousphp. Total issues resolved: **33** **Deprecations:** - [3187: Deprecate usage of binary fields whose length exceeds maximum](doctrine#3187) thanks to @morozov - [3188: Deprecated usage of binary fields whose length exceeds the platform maximum](doctrine#3188) thanks to @morozov - [3192: Added more information to the deprecation notice](doctrine#3192) thanks to @morozov - [3212: Deprecated usage of DB-generated UUIDs](doctrine#3212) thanks to @morozov **New Features:** **Bug Fixes:** - [3149: Introduced binary binding type to support binary parameters on Oracle](doctrine#3149) thanks to @morozov - [3178: Fix incorrect exception thrown from SQLAnywhere16Platform](doctrine#3178) thanks to @Majkl578 - [3044: Functional test for allowing dynamic intervals in date sub/add](doctrine#3044) thanks to @AshleyDawson - [3049: Test failures caused by invalid database connection result in fatal error](doctrine#3049) thanks to @Majkl578 **Improvements:** - [3033: Added support for available DSN parameters for the PDOSqlsrv driver](doctrine#3033) thanks to @aashmelev - [3128: Add MySQL 8 reserved keywords](doctrine#3128) thanks to @mlocati - [3143: initialize sql array into platform files](doctrine#3143) thanks to @AlessandroMinoccheri - [3173: Fix composer branch aliases](doctrine#3173) thanks to @Majkl578 - [3157: When building a limit query, zero offset without a limit should be ignored](doctrine#3157) thanks to @morozov - [3109: Allow to specify arbitrary SQL expression in AbstractPlatform::getDummySelectSQL()](doctrine#3109) thanks to @morozov - [3141: allow creating PRIMARY KEY AUTOINCREMENT fields for sqlite (unit tests)](doctrine#3141) thanks to @TimoBakx - [3180: Import simplified version of Common\Util\Debug for var dumping purposes](doctrine#3180) thanks to @Majkl578 **Documentation Improvements:** - [3117: Added badges for the develop branch in README](doctrine#3117) thanks to @morozov - [3125: Upgrading docs](doctrine#3125) thanks to @jwage - [3144: added improvement type into pull request template](doctrine#3144) thanks to @AlessandroMinoccheri **Code Quality Improvements:** - [3025: Added PHPStan, apply changes for level 3](doctrine#3025) thanks to @Majkl578 - [3200: Php Inspections (EA Ultimate): minor code tweaks](doctrine#3200) thanks to @kalessil - [3204: Fix typo in AbstractPlatform](doctrine#3204) thanks to @Majkl578 - [3205: Ignore OCI-* classes in static analysis (no stubs)](doctrine#3205) thanks to @Majkl578 **Continuous Integration Improvements:** - [3102: Use newer PHPUnit to prevent crashes on failures](doctrine#3102) thanks to @Majkl578 - [3112: Removed hard-coded configuration filenames from the test runner](doctrine#3112) thanks to @morozov - [3133: Travis DB2](doctrine#3133) thanks to @Majkl578, @morozov - [3135: AppVeyor tweaks, retry coverage upload on failure](doctrine#3135) thanks to @Majkl578 - [3137: Workaround for the inability to use a post-PHPUnit script on ContinuousPHP](doctrine#3137) thanks to @morozov - [3151: MSSQL DLL 5.2.0 has been released.](doctrine#3151) thanks to @photodude - [3160: Test against Postgres 11](doctrine#3160) thanks to @Majkl578 **Dependencies** - [3193: DBAL 2.8 needs Common 2.9](doctrine#3193) thanks to @Majkl578 - [3176: Eliminate dependency on doctrine/common](doctrine#3176) thanks to @Majkl578 - [3181: Remove dependency on doctrine/common](doctrine#3181) thanks to @Majkl578 # gpg: Signature made Fri Jul 13 06:02:10 2018 # gpg: using RSA key 374EADAF543AE995 # gpg: Can't check signature: public key not found # Conflicts: # .gitignore # README.md # lib/Doctrine/DBAL/Version.php # tests/appveyor/mssql.sql2008r2sp2.sqlsrv.appveyor.xml # tests/appveyor/mssql.sql2012sp1.sqlsrv.appveyor.xml # tests/appveyor/mssql.sql2017.pdo_sqlsrv.appveyor.xml # tests/appveyor/mssql.sql2017.sqlsrv.appveyor.xml # tests/travis/mariadb.mysqli.travis.xml

morozov reviewed Feb 17, 2018

View reviewed changes

Majkl578 force-pushed the phpstan branch from cc7b303 to 54bbec6 Compare February 17, 2018 19:47

Majkl578 commented Feb 17, 2018

View reviewed changes

Majkl578 force-pushed the phpstan branch from 54bbec6 to db41b5a Compare February 17, 2018 22:26

carusogabriel mentioned this pull request Mar 20, 2018

Fixes from PHPStan l1 for tests #3063

Merged

morozov mentioned this pull request Apr 13, 2018

[DBAL-3079] Reworked the usage of PDO in PDOConnection from inheritance to composition #3080

Merged

Majkl578 mentioned this pull request Apr 20, 2018

Revert complex type specification in Connection #3113

Merged

Majkl578 added WIP Improvement labels May 26, 2018

Majkl578 self-assigned this May 26, 2018

Majkl578 force-pushed the phpstan branch from db41b5a to d535561 Compare June 26, 2018 17:22

Majkl578 requested a review from morozov June 26, 2018 17:27

carusogabriel reviewed Jun 26, 2018

View reviewed changes

Majkl578 force-pushed the phpstan branch from d535561 to 752f9f5 Compare June 26, 2018 17:46

morozov reviewed Jun 26, 2018

View reviewed changes

Majkl578 force-pushed the phpstan branch from 752f9f5 to 1acc4a5 Compare June 27, 2018 01:18

morozov reviewed Jun 27, 2018

View reviewed changes

Majkl578 force-pushed the phpstan branch from 1acc4a5 to a3265a2 Compare June 27, 2018 14:38

Add PHPStan 0.10, apply level 3

df43437

Majkl578 force-pushed the phpstan branch from a3265a2 to df43437 Compare June 27, 2018 14:42

Majkl578 added this to the 2.9.0 milestone Jun 27, 2018

morozov approved these changes Jun 27, 2018

View reviewed changes

Majkl578 modified the milestones: 2.9.0, 2.8.0 Jun 27, 2018

Ocramius merged commit 402bbcf into doctrine:master Jun 28, 2018

Majkl578 deleted the phpstan branch June 28, 2018 14:58

Majkl578 removed the WIP label Jun 28, 2018

Majkl578 mentioned this pull request Jul 2, 2018

Fix OCI-Lob and OCI-Collection class names JetBrains/phpstorm-stubs#383

Closed

Majkl578 mentioned this pull request Jul 16, 2018

Connection::executeQuery return type #3214

Closed

morozov mentioned this pull request Aug 3, 2018

Fixed and suppressed issues found by PHPStan on develop #3238

Merged

greg0ire mentioned this pull request Jun 25, 2020

Remove the suppressed errors that PHPStan no longer reports #4107

Merged

github-actions bot locked as resolved and limited conversation to collaborators Aug 11, 2022

Added PHPStan, apply changes for level 3 #3025

Added PHPStan, apply changes for level 3 #3025

Conversation

Majkl578 commented Feb 17, 2018

morozov left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

This comment was marked as resolved.

This comment was marked as resolved.

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Majkl578 Jun 27, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

carusogabriel commented Feb 17, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

This comment was marked as resolved.

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

carusogabriel commented Apr 7, 2018

Majkl578 commented May 26, 2018

Majkl578 commented Jun 26, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

carusogabriel Jun 26, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

This comment was marked as resolved.

This comment was marked as resolved.

Choose a reason for hiding this comment

Choose a reason for hiding this comment

morozov Jun 26, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

morozov Jun 27, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

morozov Jun 27, 2018 • edited Loading

morozov left a comment •

edited

Loading

Majkl578 Jun 27, 2018 •

edited

Loading

carusogabriel Jun 26, 2018 •

edited

Loading

morozov Jun 26, 2018 •

edited

Loading

morozov Jun 27, 2018 •

edited

Loading

morozov Jun 27, 2018 •

edited

Loading