This PR was merged into the master branch.
Discussion
----------
[2.3] [WIP] Synchronized services...
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #5300, #6756
| License | MIT
| Doc PR | symfony/symfony-docs#2343
Todo:
- [x] update documentation
- [x] find a better name than contagious (synchronized)?
refs #6932, refs #5012
This PR is a proof of concept that tries to find a solution for some problems we have with scopes and services depending on scoped services (mostly the request service in Symfony).
Basically, whenever you want to inject the Request into a service, you have two possibilities:
* put your own service into the request scope (a new service will be created whenever a sub-request is run, and the service is not available outside the request scope);
* set the request service reference as non-strict (your service is always available but the request you have depends on when the service is created the first time).
This PR addresses this issue by allowing to use the second option but you service still always has the right Request service (see below for a longer explanation on how it works).
There is another issue that this PR fixes: edge cases and weird behaviors. There are several bug reports about some weird behaviors, and most of the time, this is related to the sub-requests. That's because the Request is injected into several Symfony objects without being updated correctly when leaving the request scope. Let me explain that: when a listener for instance needs the Request object, it can listen to the `kernel.request` event and store the request somewhere. So, whenever you enter a sub-request, the listener will get the new one. But when the sub-request ends, the listener has no way to know that it needs to reset the request to the master one. In practice, that's not really an issue, but let me show you an example of this issue in practice:
* You have a controller that is called with the English locale;
* The controller (probably via a template) renders a sub-request that uses the French locale;
* After the rendering, and from the controller, you try to generate a URL. Which locale the router will use? Yes, the French locale, which is wrong.
To fix these issues, this PR introduces a new notion in the DIC: synchronized services. When a service is marked as synchronized, all method calls involving this service will be called each time this service is set. When in a scope, methods are also called to restore the previous version of the service when the scope leaves.
If you have a look at the router or the locale listener, you will see that there is now a `setRequest` method that will called whenever the request service changes (because the `Container::set()` method is called or because the service is changed by a scope change).
Commits
-------
17269e1 [DependencyInjection] fixed management of scoped services with an invalid behavior set to null
bb83b3e [HttpKernel] added a safeguard for when a fragment is rendered outside the context of a master request
5d7b835 [FrameworkBundle] added some functional tests
ff9d688 fixed Request management for FragmentHandler
1b98ad3 fixed Request management for LocaleListener
a7b2b7e fixed Request management for RequestListener
0892135 [HttpKernel] ensured that the Request is null when outside of the Request scope
2ffcfb9 [FrameworkBundle] made the Request service synchronized
ec1e7ca [DependencyInjection] added a way to automatically update scoped services
This PR was merged into the master branch.
Discussion
----------
moved the request scope creation to the ContainerAwareHttpKernel class
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | symfony/symfony-docs#2343
While updating the scope documentation, I realized that the request scope was created in the FrameworkBundle while the HttpKernel that manages it was in the HttpKernel component. So, this PR makes things more consistent.
Commits
-------
cec98c1 [DependencyInjection] fixed PHP notice when the scope is not defined
550df5a moved the request scope creation to the ContainerAwareHttpKernel class
This PR was merged into the master branch.
Commits
-------
7c3179a [HttpKernel] moved the Container compilation for more flexibility
757194c [HttpKernel] made the Kernel class more flexible (closes#7396)
Discussion
----------
[HttpKernel] made the Kernel class more flexible (closes#7396)
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #7396
| License | MIT
| Doc PR | n/a
---------------------------------------------------------------------------
by jrobeson at 2013-03-20T14:04:53Z
I still (personally) think that the compile() method is still too close to the building of the container. Is there a reasonable way to move the compile() method away from the rest of the logic used to build the container?
---------------------------------------------------------------------------
by elnur at 2013-03-20T14:34:23Z
I like this one. Can't wait to see this merged.
---------------------------------------------------------------------------
by stof at 2013-03-20T14:50:22Z
@jrobeson compiling the container is part of building it (it is the step running the compiler passes)
---------------------------------------------------------------------------
by jrobeson at 2013-03-20T15:32:26Z
i read the code stof .. that is what it does :) . I just thought it'd be easier to modify the container from the kernel of the compile() call was placed after the buildContainer() call at https://github.com/symfony/symfony/blob/master/src/Symfony/Component/HttpKernel/Kernel.php#L556
This is something i had to work around when migrating to symfony. It's not a big deal to me anymore, so no worries.
---------------------------------------------------------------------------
by stof at 2013-03-20T15:36:12Z
@jrobeson the container is not fully build until you compile it. Half of the work is done by the compile() call.
---------------------------------------------------------------------------
by fabpot at 2013-03-20T15:38:24Z
I've moved the compile method as it makes things even more flexible.
---------------------------------------------------------------------------
by jrobeson at 2013-03-20T15:41:45Z
@stof : i know .. which is even better reason to move it outside that big method.
@fabpot: thanks!
* 2.2: (70 commits)
change wrapped exception message to be more usefull
updated VERSION for 2.0.23
update CONTRIBUTORS for 2.0.23
updated CHANGELOG for 2.0.23
[Form] fixed failing test
[DomCrawler] added support for query string with slash
Fixed invalid file path for hiddeninput.exe on Windows.
fix xsd definition for strict-requirements
[WebProfilerBundle] Fixed the toolbar styles to apply them in IE8
[ClassLoader] fixed heredocs handling
fixed handling of heredocs
Add a public modifier to an interface method
removing xdebug extension
[HttpRequest] fixes Request::getLanguages() bug
[HttpCache] added a test (cached content should be kept after purging)
[DoctrineBridge] Fixed non-utf-8 recognition
[Security] fixed HttpUtils class tests
replaced new occurences of 'Request::create()' with '::create()'
changed sub-requests creation to '::create()'
fixed merge issue
...
Conflicts:
src/Symfony/Bundle/FrameworkBundle/Command/TranslationUpdateCommand.php
src/Symfony/Bundle/WebProfilerBundle/Resources/views/Profiler/toolbar.html.twig
src/Symfony/Component/DomCrawler/Link.php
src/Symfony/Component/Translation/Translator.php
* 2.1:
Add a public modifier to an interface method
[HttpRequest] fixes Request::getLanguages() bug
[HttpCache] added a test (cached content should be kept after purging)
[DoctrineBridge] Fixed non-utf-8 recognition
[Security] fixed HttpUtils class tests
This PR was merged into the 2.1 branch.
Commits
-------
54609b9 [HttpCache] added a test (cached content should be kept after purging)
Discussion
----------
[HttpCache] added a test (cached content should be kept after purging)
closes#6830
This PR was merged into the master branch.
Commits
-------
a011842 [HttpKernel] fixed memory collector
def2ccb Add PHP memory_limit to WDT
Discussion
----------
[HttpKernel] fixed memory collector
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #7037
---------------------------------------------------------------------------
by vicb at 2013-03-15T20:21:55Z
OT, that's great to see many contributions for you @jfsimon, thanks !
---------------------------------------------------------------------------
by jfsimon at 2013-03-18T09:35:54Z
Thanks @vicb, this is so lovely :)
This PR was merged into the master branch.
Commits
-------
ed58e36 [HttpKernel] CLI - don't always display errors
Discussion
----------
[HttpKernel] CLI - don't always display errors
I had closed#6370 by mistake when cleaning up my branches.
---------------------------------------------------------------------------
by sstok at 2013-02-13T15:13:16Z
👍
* 2.1:
sub-requests are now created with the same class as their parent
[FrameworkBundle] removed BC break
[FrameworkBundle] changed temp kernel name in cache:clear
[DoctrineBridge] Avoids blob values to be logged by doctrine
[Security] use current request attributes to generate redirect url?
[Validator] fix showing wrong max file size for upload errors
[TwigBridge] removed double var initialization (refs #7344)
[2.1][TwigBridge] Fixes Issue #7342 in TwigBridge
[FrameworkBundle] fixed cahe:clear command's warmup
[TwigBridge] now enter/leave scope on Twig_Node_Module
[TwigBridge] fixed fixed scope & trans_default_domain node visitor
[TwigBridge] fixed non probant tests & added new one
[BrowserKit] added ability to ignored malformed set-cookie header
[Translation] removed wriong 'use'
[Translation] added xliff loader/dumper with resname support
[TwigBridge] fixes
Conflicts:
src/Symfony/Bundle/FrameworkBundle/HttpKernel.php
src/Symfony/Component/Security/Http/HttpUtils.php
src/Symfony/Component/Translation/Loader/XliffFileLoader.php
src/Symfony/Component/Translation/Tests/Loader/XliffFileLoaderTest.php
Use the same format as object methods to describe closures and collect
the file and the line where it's been declared.
These informations should be added in the views of the webprofiler.
This PR was squashed before being merged into the 2.2 branch (closes#7321).
Commits
-------
4f3771d [2.2][HttpKernel] fixed wrong option name in FragmentHandler::fixOptions
Discussion
----------
[2.2][HttpKernel] fixed wrong option name in FragmentHandler::fixOptions
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony/issues/7309
---------------------------------------------------------------------------
by Aitboudad at 2013-03-10T11:38:59Z
Hi @fabpot,
can you see why all commits has failed on Travis ?
---------------------------------------------------------------------------
by henrikbjorn at 2013-03-10T11:42:25Z
The test expects the option to be called `renderer` and not `strategy` so you need to change the tests.
---------------------------------------------------------------------------
by Aitboudad at 2013-03-10T12:19:19Z
@henrikbjorn thanks it's ok, but there is still 1 test :( in Symfony\Component\Form\Tests\Extension\Core\Type\TimeTypeTest::testSetDataDifferentTimezones
@bschussek Can you see why :).
* 2.1:
[FrameworkBundle] Fix code status in dockblock
Fixed test to use Reflection
[Finder] fixed a potential issue on Solaris where INF value is wrong (refs #7269)
Update RouteCompiler.php
[FrameworkBundle] avoids cache:clear to break if new/old folders already exist
[HttpKernel] Fixed possible profiler token collision (closes#7272, closes#7171)
[ClassLoader] tweaked test
[ClassLoader] made DebugClassLoader idempotent
[DomCrawler] Fix relative path handling in links
Conflicts:
src/Symfony/Component/DomCrawler/Link.php
src/Symfony/Component/Finder/Iterator/DepthRangeFilterIterator.php
src/Symfony/Component/Routing/RouteCompiler.php
* 2.2: (26 commits)
[FrameworkBundle] Fixes invalid serialized objects in cache
remove dead code in yaml component
Fixed typo in UPGRADE-2.2
fixed typo
RedisProfilerStorage wrong db-number/index-number selected
[DependencyInjection] added a test for the previous merge (refs #7261)
Unset loading[$id] in ContainerBuilder on exception
Default validation message translation fix.
remove() should not use deprecated getParent() so it does not trigger deprecation internally
adjust routing tests to not use prefix in addCollection
add test for uniqueness of resources
added tests for addDefaults, addRequirements, addOptions
adjust RouteCollectionTest for the addCollection change and refactor the tests to only skip the part that really needs the config component
added tests for remove() that wasnt covered yet and special route name
refactor interator test that was still assuming a tree
adjust tests to no use addPrefix with options
adjusted tests to not use RouteCollection::getPrefix
[Routing] trigger deprecation warning for deprecated features that will be removed in 2.3
[Console] fixed StringInput binding
[Console] added string input test
...
This PR was merged into the master branch.
Commits
-------
b0bca01 add status text to web profiler response panel
Discussion
----------
[2.3] add status text to web profiler response panel
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | cosmetic
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
Add the status text to the web profiler response panel
---------------------------------------------------------------------------
by Burgov at 2013-02-17T11:08:37Z
The response object allows for setting a custom status message, so displaying the predefined one could be confusing.
E.g. ```$reponse->setStatusCode(403, "You can't go here!")```
However, there is no getter for it to be read... Perhaps it should be added?
---------------------------------------------------------------------------
by fabpot at 2013-02-17T11:32:37Z
I think displaying the default text is good enough. I suppose that the intent is to easily remember what a given status code means. There is no getter for the status text as it has no real meaning in terms of HTTP.
---------------------------------------------------------------------------
by gimler at 2013-02-17T16:50:47Z
Yes this was my intention. But i think this is a good point the response should have a `getStatusText` method.
So actual to only point where the status text is use is when the response is converted to string.
@fabpot i would add the `getStatusText` method to the `response` class and modify the PR to use this text. Do you agree with that?
---------------------------------------------------------------------------
by fabpot at 2013-02-17T17:13:07Z
Ad I said in my previous comment, not having a `getStatusText()` was in fact intentional.
This PR was squashed before being merged into the 2.1 branch (closes#7177).
Commits
-------
c82c754 RedisProfilerStorage wrong db-number/index-number selected
Discussion
----------
RedisProfilerStorage wrong db-number/index-number selected
bug: in the webprofiler the wrong database (0) is selected when storage should not go to index 0. on redis connect the default behaviour is to select index/database number 0, but it is necessary to select a special index/database.
see: http://rediscookbook.org/multiple_databases.html
[HttpKernel] [Profiler] [RedisProfilerStorage] added select for a db-number/db-index to dsn-patttern
| Q | A
| ------------- | ---
| Bug fix? | [yes]
| New feature? | [no]
| BC breaks? | [no]
| Deprecations? | [no]
| Tests pass? | [yes]
| Fixed tickets |
| License | MIT
Usage with index/db-number = 7 :
```xml
<!-- config_dev.xml -->
<symfony:profiler only-exceptions="false" dsn="redis://127.0.0.1:6379/7" lifetime="3600" />
```
```yml
#config_dev.yml
framework:
...
profiler:
...
dsn: redis://127.0.0.1:6379/7
```
---------------------------------------------------------------------------
by fabpot at 2013-02-27T07:21:02Z
Can you also add some unit tests?
---------------------------------------------------------------------------
by JohnDoe007 at 2013-02-27T09:36:28Z
@fabpot the function getRedis() is currently not testable, I think. this means I have to do a bigger change to the class, add a public function initRedis() or something and move initialize code there to make the initialization process, where my change is, testable. is this ok to do it in that bugfix/pull request? or should I add this test change to master and only adding this small bugfix to the 2.1-branch?
---------------------------------------------------------------------------
by fabpot at 2013-03-06T16:47:37Z
Thanks, I've merged the fix. Can you work on the changes needed to allow the code to be testable (on the master branch)?
* 2.2:
fixed CS
Add persian translation to Components/Security
bumped Symfony version to 2.2.1-DEV-DEV
updated VERSION for 2.2.0
updated CHANGELOG for 2.2.0
* 2.2: (24 commits)
Options small typo
[Console] fixed unparsed StringInput tokens
Mask PHP_AUTH_PW header in profiler
[TwigBridge] fixed trans twig extractor
[Finder] adds adapter selection/unselection capabilities
[DomCrawler] fix handling of schemes by Link::getUri()
[Console] Fixed comment
[TwigBridge] fixed the translator extractor that were not trimming the text in trans tags (closes#7056)
Fixed handling absent href attribute in base tag
fixed paths/notPaths regex for shell adapters
fix issue 4911
Adds expandable globs support to shell adapters
[HttpFoundation] Fixed messed up headers
Fixes AppCache + ESI + Stopwatch problem
added a DebuClassLoader::findFile() method to make the wrapping less invasive
bumped Symfony version to 2.2.0-RC4-DEV
updated VERSION for 2.2.0-RC3
updated CHANGELOG for 2.2.0-RC3
fixed CHANGELOG
bumped Symfony version to 2.1.9-DEV
...
