Codestin Search App

kbond · 2018-01-15T19:08:56Z

This is a failing test to show the issue.

Given the following inheritance structure:

Person <- Employee <- Manager

When hydrating a manager using a query builder created from the person repository, the employee's properties do not get hydrated.

I think this was caused by #6304.

kbond · 2018-01-15T19:10:37Z

+        ;
+
+        $this->assertSame('Kevin', $persistedManager->name);
+        $this->assertSame('555-5555', $persistedManager->phoneNumber);


This is the failure.

kbond · 2018-01-17T19:26:29Z

I added another test to show it working when using the HYDRATE_SIMPLEOBJECT hydrator.

lcobucci · 2018-02-18T09:21:32Z

                // the current discriminator value must be saved in order to disambiguate fields hydration,
                // should there be field name collisions
                if ($classMetadata->parentClasses && isset($this->_rsm->discriminatorColumns[$ownerMap])) {
+                    $discriminatorValues = array_map(


@Ocramius @guilhermeblanco in theory this addition should have some impact on performance, but the overall result of phpbench in my machine is slightly better than it was before:

- ⅀T: 1,914,921.000μs μSD/r 0.000μs μRSD/r: 0.000% + ⅀T: 1,879,858.000μs μSD/r 0.000μs μRSD/r: 0.000%

It's possible that my laptop is a bit crazy but... whatever.

Not sure if it's relevant for v2.x but for v3.x we could create the list of all discriminator values of the subclasses during the metadata creation.

Complete phpbench diff

\Doctrine\Performance\Hydration\SimpleQueryPartialObjectHydrationPerformanceBench - benchHydration I0 P0 [μ Mo]/r: 109,739.000 109,739.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 95,171.000 95,171.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\SimpleQueryScalarHydrationPerformanceBench - benchHydration I0 P0 [μ Mo]/r: 37,513.000 37,513.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 34,285.000 34,285.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\MixedQueryFetchJoinFullObjectHydrationPerformanceBench - benchHydration I0 P0 [μ Mo]/r: 105,009.000 105,009.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 100,475.000 100,475.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\SimpleQueryFullObjectHydrationPerformanceBench - benchHydration I0 P0 [μ Mo]/r: 337,610.000 337,610.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 329,276.000 329,276.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\SingleTableInheritanceInsertPerformanceBench - benchInsertFixContracts I0 P0 [μ Mo]/r: 8,695.000 8,695.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchInsertFixContracts I0 P0 [μ Mo]/r: 7,281.000 7,281.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchInsertFlexContracts I0 P0 [μ Mo]/r: 7,156.000 7,156.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchInsertFlexContracts I0 P0 [μ Mo]/r: 7,624.000 7,624.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchInsertUltraContracts I0 P0 [μ Mo]/r: 8,162.000 8,162.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchInsertUltraContracts I0 P0 [μ Mo]/r: 7,510.000 7,510.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchInsertAllContracts I0 P0 [μ Mo]/r: 11,867.000 11,867.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchInsertAllContracts I0 P0 [μ Mo]/r: 13,101.000 13,101.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\SingleTableInheritanceHydrationPerformanceBench - benchHydrateFixContracts I0 P0 [μ Mo]/r: 3,520.000 3,520.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydrateFixContracts I0 P0 [μ Mo]/r: 3,815.000 3,815.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchHydrateFlexContracts I0 P0 [μ Mo]/r: 4,791.000 4,791.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydrateFlexContracts I0 P0 [μ Mo]/r: 4,108.000 4,108.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchHydrateUltraContracts I0 P0 [μ Mo]/r: 3,921.000 3,921.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydrateUltraContracts I0 P0 [μ Mo]/r: 3,636.000 3,636.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchHydrateAllContracts I0 P0 [μ Mo]/r: 4,898.000 4,898.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydrateAllContracts I0 P0 [μ Mo]/r: 5,067.000 5,067.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\SimpleInsertPerformanceBench - benchHydration I0 P0 [μ Mo]/r: 494,637.000 494,637.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 497,152.000 497,152.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\SimpleQueryArrayHydrationPerformanceBench - benchHydration I0 P0 [μ Mo]/r: 53,799.000 53,799.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 50,772.000 50,772.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\MixedQueryFetchJoinPartialObjectHydrationPerformanceBench - benchHydration I0 P0 [μ Mo]/r: 56,808.000 56,808.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 55,068.000 55,068.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\MixedQueryFetchJoinArrayHydrationPerformanceBench - benchHydration I0 P0 [μ Mo]/r: 102,135.000 102,135.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 99,480.000 99,480.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\Hydration\SimpleHydrationBench - benchHydration I0 P0 [μ Mo]/r: 260,085.000 260,085.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchHydration I0 P0 [μ Mo]/r: 257,370.000 257,370.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\ChangeSet\UnitOfWorkComputeChangesBench - benchChangeSetComputation I0 P0 [μ Mo]/r: 1,536.000 1,536.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchChangeSetComputation I0 P0 [μ Mo]/r: 1,519.000 1,519.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\LazyLoading\ProxyInitializationTimeBench - benchCmsUserInitialization I0 P0 [μ Mo]/r: 69,404.000 69,404.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchCmsUserInitialization I0 P0 [μ Mo]/r: 65,095.000 65,095.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchCmsEmployeeInitializationI0 P0 [μ Mo]/r: 11,635.000 11,635.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchCmsEmployeeInitializationI0 P0 [μ Mo]/r: 11,820.000 11,820.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchInitializationOfAlreadyInitializedCmsUsersI0 P0 [μ Mo]/r: 777.000 777.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchInitializationOfAlreadyInitializedCmsUsersI0 P0 [μ Mo]/r: 822.000 822.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchInitializationOfAlreadyInitializedCmsEmployeesI0 P0 [μ Mo]/r: 625.000 625.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchInitializationOfAlreadyInitializedCmsEmployeesI0 P0 [μ Mo]/r: 651.000 651.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% \Doctrine\Performance\LazyLoading\ProxyInstantiationTimeBench - benchCmsUserInstantiation I0 P0 [μ Mo]/r: 127,576.000 127,576.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchCmsUserInstantiation I0 P0 [μ Mo]/r: 135,914.000 135,914.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% - benchCmsEmployeeInstantiation I0 P0 [μ Mo]/r: 93,023.000 93,023.000 (μs) [μSD μRSD]/r: 0.000μs 0.00% + benchCmsEmployeeInstantiation I0 P0 [μ Mo]/r: 92,846.000 92,846.000 (μs) [μSD μRSD]/r: 0.000μs 0.00%

Not sure if it's relevant for v2.x but for v3.x we could create the list of all discriminator values of the subclasses during the metadata creation.

More than that, questions to the metadata should be asked directly to the metadata, rather than done in random metadata consumers

This block is an O(1) called at hydrator initialisation, if I'm not mistaken.

Please move it to a private method

@lcobucci extremely flaky bench test results, but they seem to stay around normal values, so it is OK IMO

lcobucci · 2018-02-18T09:24:52Z

                    if (isset($cacheKeyInfo['discriminatorColumn'], $data[$cacheKeyInfo['discriminatorColumn']])
-                        // Note: loose comparison required. See https://github.com/doctrine/doctrine2/pull/6304#issuecomment-323294442
-                        && ! in_array($data[$cacheKeyInfo['discriminatorColumn']], $cacheKeyInfo['discriminatorValues'])
+                        && ! in_array((string) $data[$cacheKeyInfo['discriminatorColumn']], $cacheKeyInfo['discriminatorValues'], true)


In that comment @Ocramius says:

I checked the issue locally, and it's because the discriminator is not converted to a PHP value.

That said, trying to fix the issue leads to a massive can of worms that is being opened: let's keep the loose comparison there for now.

I've opened the can of worms and it wasn't that bad but it's possible that I've overlooked things, please use some ❤️ to review this commit

Ocramius

Besides code to be moved to a private method, looks OK 👍

Ocramius · 2018-02-19T07:52:34Z

                // the current discriminator value must be saved in order to disambiguate fields hydration,
                // should there be field name collisions
                if ($classMetadata->parentClasses && isset($this->_rsm->discriminatorColumns[$ownerMap])) {
+                    $discriminatorValues = array_map(


Not sure if it's relevant for v2.x but for v3.x we could create the list of all discriminator values of the subclasses during the metadata creation.

More than that, questions to the metadata should be asked directly to the metadata, rather than done in random metadata consumers

Ocramius · 2018-02-19T07:54:28Z

                // the current discriminator value must be saved in order to disambiguate fields hydration,
                // should there be field name collisions
                if ($classMetadata->parentClasses && isset($this->_rsm->discriminatorColumns[$ownerMap])) {
+                    $discriminatorValues = array_map(


This block is an O(1) called at hydrator initialisation, if I'm not mistaken.

Please move it to a private method

Ocramius · 2018-02-19T07:56:45Z

                // the current discriminator value must be saved in order to disambiguate fields hydration,
                // should there be field name collisions
                if ($classMetadata->parentClasses && isset($this->_rsm->discriminatorColumns[$ownerMap])) {
+                    $discriminatorValues = array_map(


@lcobucci extremely flaky bench test results, but they seem to stay around normal values, so it is OK IMO

lcobucci · 2018-02-19T10:09:19Z

@Ocramius please check if everything is good and

Hydrator was ignoring data from subclasses when using multiple inheritance levels. With this patch it will now use the discriminator values from all subclasses of the class being hydrated.

According to mapping drivers the discriminator values can always be converted to strings so it's safe to assume that we can actually do a strict comparison during hydration.

Ocramius · 2018-02-19T11:11:46Z

🐑

Ocramius · 2018-02-19T11:13:04Z

One unrelated test failed

- `em` protected property - annotation namespace change - `QueryBuilder` instantiation

…ated' Close #6988 Close #6304

Ocramius · 2018-02-19T11:13:43Z

Forward-port in master in 0822f8a

Inheritance middle-layer doesn't get hydrated

…es in `3.x` - `em` protected property - annotation namespace change - `QueryBuilder` instantiation

kbond commented Jan 15, 2018

View reviewed changes

Majkl578 added Failing Test Bug labels Jan 15, 2018

holtkamp mentioned this pull request Jan 15, 2018

Entity / property hydration issues after upgrading 2.5.14 => 2.6.0 #6937

Closed

kbond force-pushed the inheritance-issue branch from 72b7f8d to ca85fe3 Compare January 17, 2018 19:25

BenMorel mentioned this pull request Feb 12, 2018

Querying abstract class in the middle of a hierarchy returns NULL instead of objects #7059

Closed

lcobucci mentioned this pull request Feb 17, 2018

Bugfix: Hydrate subclasses of subclasses #7051

Closed

lcobucci force-pushed the inheritance-issue branch from ca85fe3 to 42f6bc6 Compare February 18, 2018 09:00

lcobucci requested review from Ocramius and guilhermeblanco February 18, 2018 09:00

lcobucci force-pushed the inheritance-issue branch from 42f6bc6 to 1a42971 Compare February 18, 2018 09:04

lcobucci reviewed Feb 18, 2018

View reviewed changes

lcobucci added this to the 2.6.1 milestone Feb 18, 2018

Ocramius approved these changes Feb 19, 2018

View reviewed changes

Ocramius assigned lcobucci Feb 19, 2018

lcobucci requested a review from Ocramius February 19, 2018 10:08

lcobucci assigned Ocramius and unassigned lcobucci Feb 19, 2018

lcobucci force-pushed the inheritance-issue branch from 77c5e98 to 1c45551 Compare February 19, 2018 11:07

kbond and others added 4 commits February 19, 2018 12:07

Inheritance middle-layer doesn't get hydrated with HYDRATE_OBJECT

15a4302

Use partial discriminator map on multi-inheritance

48ca6db

Hydrator was ignoring data from subclasses when using multiple inheritance levels. With this patch it will now use the discriminator values from all subclasses of the class being hydrated.

Remove loose comparison on discriminator values

2905b43

According to mapping drivers the discriminator values can always be converted to strings so it's safe to assume that we can actually do a strict comparison during hydration.

Extract private method to retrieve discriminator values

f2da5bc

lcobucci force-pushed the inheritance-issue branch from 1c45551 to f2da5bc Compare February 19, 2018 11:07

Ocramius merged commit a736a37 into doctrine:2.6 Feb 19, 2018

Ocramius added a commit that referenced this pull request Feb 19, 2018

#6988 #6304 adapted test case to ORM test suite changes in 3.x

e7e4247

- `em` protected property - annotation namespace change - `QueryBuilder` instantiation

Ocramius added a commit that referenced this pull request Feb 19, 2018

Merge branch 'fix/#6988-#6304-ensure-inheritance-middle-layer-is-hydr…

0822f8a

…ated' Close #6988 Close #6304

supersmile2009 pushed a commit to supersmile2009/doctrine2 that referenced this pull request Feb 21, 2018

Merge pull request doctrine#6988 from kbond/inheritance-issue

09acd55

Inheritance middle-layer doesn't get hydrated

maglnet pushed a commit to maglnet/doctrine2 that referenced this pull request Oct 10, 2018

doctrine#6988 doctrine#6304 adapted test case to ORM test suite chang…

e885a21

…es in `3.x` - `em` protected property - annotation namespace change - `QueryBuilder` instantiation

kbond deleted the inheritance-issue branch February 27, 2020 19:49

beberlei mentioned this pull request Dec 7, 2020

Subclasses of subclasses don't hydrate #7050

Closed

Conversation

kbond commented Jan 15, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

kbond commented Jan 17, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lcobucci Feb 18, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Ocramius left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lcobucci commented Feb 19, 2018

Uh oh!

Ocramius commented Feb 19, 2018

Uh oh!

Ocramius commented Feb 19, 2018

Uh oh!

Ocramius commented Feb 19, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

lcobucci Feb 18, 2018 •

edited

Loading