This PR was merged into the 2.2 branch.
Commits
-------
35b62ac [Security] renamed Constraint namespace to Constraints for validator classes in order to be consistent with the whole current validator API.
Discussion
----------
[Security] renamed Constraint namespace to Constraints for validator classes in order to be consistent with the whole current validator API.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | yes (for people already using the ``UserPassword`` constraint class in their code)
| Deprecations? | yes
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This PR was merged into the 2.1 branch.
Commits
-------
8ca00c5 [Security] fixed session creation when none is needed (closes#6917)
Discussion
----------
[Security] fixed session creation when none is needed (closes#6917)
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #6917
| License | MIT
| Doc PR | n/a
---------------------------------------------------------------------------
by drak at 2013-02-04T16:24:49Z
That looks good. Maybe we need a test for this logic to prevent any regression in the future?
---------------------------------------------------------------------------
by bendavies at 2013-02-04T16:30:38Z
Yep, this was exactly what i tried locally, but really wasn't familiar enough with it to be confident enough to submit it as a fix.
Works for me!
---------------------------------------------------------------------------
by bendavies at 2013-02-04T17:19:32Z
A few test failures which were added by the breaking PR #2414 in the first place.
---------------------------------------------------------------------------
by fabpot at 2013-02-04T18:00:31Z
I've fixed the tests which now really test that the session is not started.
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6725).
Commits
-------
957b3af [DependencyInjection]Micro-optimization
Discussion
----------
[DependencyInjection]Micro-optimization
Bug fix: no
Feature addition: not really
Backwards compatibility break: no
Fixes the following tickets: ~
Todo: see below
License of the code: MIT
Documentation PR: ~
This adds a small micro-optimization to the container once it's dumped as PHP.
Even if the speed bump isn't that visible on the default Symfony2 Standard distribution, this doesn't break anything and it should be helpful when having a lot of parameters in the container.
I've did got better results with this applied but it would be helpful if others could confirm this to be true as well.
---------------------------------------------------------------------------
by fabpot at 2013-01-13T19:57:36Z
I'm -1 on this change. I really doubt that this has any significant perf improvement and it makes the code less clear. In fact, we rejected similar changes in the past.
---------------------------------------------------------------------------
by alexandresalome at 2013-01-13T20:32:54Z
This code is executed on compilation time, which means it's not executed in each request of production server.
---------------------------------------------------------------------------
by igorw at 2013-01-13T21:03:33Z
@alexandresalome no, this is changing the generated code that will be run on every request ;-). But I agree that it's probably not worth it. Unless there is a benchmark proving *significant* gains.
---------------------------------------------------------------------------
by dlsniper at 2013-01-13T21:05:52Z
@alexandresalome the code is execute at runtime since this changes the dumper part of it.
@fabpot I didn't knew about the changes like this but I from my recent experience with optimizing various applications this kind of change was always improving speed. Also, aside from the fact that this is a known trick, readability was taken into consideration, that's why I've added the change only in the dumper.
I've did a benchmark against a default container from Symfony2 and you can get it from here: https://gist.github.com/4526170
Feel free to close this if you want.
Note: Currently I'm trying to find and improve all the things related to speed in Symfony2, see the ticket here: symfony/symfony-standard#464 which started from this: https://groups.google.com/d/msg/symfony2/-gvguJqox3s/bjaA6Ks3EZoJ which was confirmed by the OP as well. This won't bring big performance increases, but it's a step forward imho.
Thanks!
---------------------------------------------------------------------------
by vicb at 2013-01-14T07:21:12Z
@dlsniper there seems to be a huge diff in your bench between isset and ake, any chance you have xdebug turned on ?
---------------------------------------------------------------------------
by dlsniper at 2013-01-14T08:03:35Z
@vicb no, both xhprof / xdebug where turned off. Also keep in mind that there's a ` E-5 ` so the performance impact is not that much in terms of microseconds, that's why I've said micro-optimization, but it is still there and I've had projects with more that 1.3k values in that array and the project was medium in size.
That's why I've said, someone else please test it as well and provided the benchmark, to prove that the ` isset + array_key_exists ` is faster that ` array_key_exists ` when the searched value is not null.
Also, the benchmark assumes that there are only 300 searches in the parameters array which might be more or less that usual numbers and it only retrieves the same parameter every time.
---------------------------------------------------------------------------
by vicb at 2013-01-14T08:08:57Z
@dlsniper I am not referring to absolute values but to the ratio that is high.
---------------------------------------------------------------------------
by mvrhov at 2013-01-14T08:29:12Z
isset is known to be a way faster than ake. BTW my work project which I consider relatively small has just shy of 1000 entries in parameters array
---------------------------------------------------------------------------
by vicb at 2013-01-14T09:29:59Z
@mvrhov you're right, isset is faster but [especially when xdebug is on](http://ilia.ws/archives/247-Performance-Analysis-of-isset-vs-array_key_exists.html) - 2.5x vs 10x.
I would generally 👎 micro-optims but this one is in a generated file so 👍
This PR was merged into the 2.2 branch.
Commits
-------
95f0202 [FrameworkBundle] fixed wrong indentation on route debug output
Discussion
----------
[FrameworkBundle] fixed wrong indentation on route debug output
I'm not sure this is something to fix, but with this PR the debug output for a single route is like the lower part of the image, instead of the currently displayed upper part.
![output](https://f.cloud.github.com/assets/254808/120850/f2b6e8ce-6d43-11e2-9fdd-24200eb2e93b.jpg)
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6882).
Commits
-------
b8ef96e [Finder] fixed regression in the Finder component (it was possible to use it without using exec before, closes#6357)
Discussion
----------
[Finder] fixed regression in the Finder component (it was possible to use it without using exec before, closes#6357)
| Q | A
| ------------- | ---
| Bug fix? | yes (fixed a BC break)
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #6357
| License | MIT
| Doc PR | n/a
This PR was submitted for the 2.2 branch but it was merged into the 2.0 branch instead (closes#6959).
Commits
-------
ad889c9 [DependencyInjection] fixed a circular call (closes#6864)
Discussion
----------
[DependencyInjection] fixed a circular call (closes#6864)
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #6864
| License | MIT
| Doc PR | n/a
This PR was merged into the 2.2 branch.
Commits
-------
a433ba2 [Security] [Tests] added unit tests for the UserPasswordValidator class and made the validator service for the UserPassword constraint configurable.
Discussion
----------
[Security] [Tests] added unit tests for the UserPasswordValidator class ...
...and made the validator service for the UserPassword constraint configurable.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This PR was merged into the 2.2 branch.
Commits
-------
8f8d6cf [HttpKernel] fixed regression when rendering an inline controller and passing some objects (closes#6822)
Discussion
----------
[HttpKernel] fixed regression when rendering an inline controller and passing some objects (closes#6822)
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #6822
| License | MIT
| Doc PR | n/a
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6956).
Commits
-------
15625b4 typo
Discussion
----------
typo
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6945).
Commits
-------
a2a2f54 fixed wrong indentation
Discussion
----------
fixed wrong indentation
This PR was merged into the 2.1 branch.
Commits
-------
e7624b6 [Security] Add PHPDoc to AuthenticationEvents
Discussion
----------
[Security] Added PHPDoc to AuthenticationEvents
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6944).
Commits
-------
2b850b5 [Form] [Validator] [Security] Add Arabic translations.
Discussion
----------
[Form] [Validator] [Security] Add Arabic translations.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | ~
| License | MIT
| Doc PR | ~
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6925).
Commits
-------
0586643 renamed some classes and Twig functions to more descriptive names (refs #6871)
Discussion
----------
renamed some classes and Twig functions to more descriptive names (refs #6871)
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #6871
| License | MIT
| Doc PR | symfony/symfony-docs#2205
Todo:
- [x] update the docs
In #6871, @kriswallsmith wondered if the names used for the ESI/HIncludes sub-framework were meaningful enough.
I agree that this was not the case and I propose to remove the notion of **sub-requests** in favor of **fragments**. This sub-framework is a way to render fragments of a resource (the fact that it's done via another request is an implementation detail).
With that decision, all names can be renamed and are probably more meaningful. Some examples:
* `HttpContentRenderer` -> `FragmentHandler`
* `RouterProxyListener` -> `FragmentListener`
* `router_proxy` -> `fragments` (configuration entry)
* `DefaultRenderingStrategy` -> `InlineFragmentRenderer`
---------------------------------------------------------------------------
by fabpot at 2013-01-31T09:50:14Z
I forgot to mention that this renaming will also probably help documenting the feature as understanding the notion of a fragment is probably easier.
---------------------------------------------------------------------------
by Tobion at 2013-01-31T14:18:40Z
I'd like to say again that the word `render` in `FragmentRenderingStrategy` does not fit because it's not about rendering (that's left to the templating) of a fragment.
I suggest to rename it to `FragmentInclusionStrategy` because it defines the strategy to include the fragment.
This is also consistent with every strategy there is:
- ESI: `<esi:include`
- SSI: `<!--#include file="header.shtml" -->`
- hinclude (even in the name)
- inline (similar to phps `include()`)
---------------------------------------------------------------------------
by fabpot at 2013-01-31T14:48:07Z
I've just renamed `FragmentRenderer` to `FragmentHandler` and strategies like `EsiFragmentRenderingStrategy` to `EsiFragmentRenderer` (and everything is put into a new `Fragment` sub-namespace).
---------------------------------------------------------------------------
by fabpot at 2013-01-31T21:11:44Z
I've reverted the Twig function name change as the current name is more expressive with its arguments: `render(controller(...))` or `render_esi(url(...))`.
HttpContentRenderer has been renamed to FragmentHandler.
The RendererStrategy subnamespace has been renamed to Fragment.
The strategy classes now have Fragment in their names.
ProxyRouterListener has been renamed to FragmentListener
The router_proxy configuration entry has been renamed to fragments.
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6258).
Commits
-------
d4acd6d Classcollectionloader: fix traits + enhancements
Discussion
----------
Classcollectionloader: fix traits + enhancements
Commits:
- some tweaks
- generates smaller cache files (20% decrease for the SE)
- fix traits dependency - thanks php for consistency https://bugs.php.net/bug.php?id=61554
---------------------------------------------------------------------------
by vicb at 2012-12-11T07:56:07Z
@stloyd @fabpot thanks for the useful reviews, everything should be fixed now.
---------------------------------------------------------------------------
by vicb at 2012-12-11T10:03:48Z
Should be ready.
---------------------------------------------------------------------------
by vicb at 2013-01-15T21:05:42Z
@fabpot I initially marked this PR for 2.3 but it contains a fix for traits, should you merge it ?
---------------------------------------------------------------------------
by vicb at 2013-01-23T20:09:32Z
@fabpot could this be in 2.2 ?
---------------------------------------------------------------------------
by vicb at 2013-02-01T14:07:08Z
@fabpot fixed
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6938).
Commits
-------
b279dcb [Validator] Fix a deprecated method call in the tests
Discussion
----------
[Validator] Fix a deprecated method call in the tests
BUG FIX - Travis happier ?
---------------------------------------------------------------------------
by stloyd at 2013-02-01T12:23:18Z
This should go to `2.2` I guess... =)
---------------------------------------------------------------------------
by vicb at 2013-02-01T12:25:43Z
Well there has been no official announcement for the 2.2 branch I think ?
I hope that some PR pending for master will go into 2.2 anyway ?
/cc @fabpot
---------------------------------------------------------------------------
by vicb at 2013-02-01T12:38:35Z
C'mon Travis, aren't you allowed to fail for 5.5 ?
---------------------------------------------------------------------------
by stloyd at 2013-02-01T12:43:42Z
@vicb error != failed ;-) Yet I agree, that now it's more confusing than before this new status...
---------------------------------------------------------------------------
by fabpot at 2013-02-01T13:27:11Z
I will merged it into 2.2. 2.2 is entering the RC phase, so no more modifications will be merged.
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#6632).
Commits
-------
a1c0ff5 Update `composer.json` files: - to allow versions ~2.2 (>=2.2,<3.0) of Doctrine DBAL, ORM & Common - fixed Propel1 versions difference between main and bridge files - fixed Twig versions difference between main and bridge files - to allow versions ~1.11 (>=1.11,<2.0) of Twig - fixed Locale ext-intl version to accept all, not non-existing version
Discussion
----------
Update `composer.json` files
---------------------------------------------------------------------------
by fabpot at 2013-01-09T17:55:54Z
When will Doctrine 2.4 be released? If not in the next couple of weeks, this PR cannot be merged as we need to ship Symfony 2.2 with a stable version of Doctrine.
---------------------------------------------------------------------------
by stof at 2013-01-09T18:13:59Z
@fabpot This allows to use Doctrine 2.4. It does not require to use 2.4. Actually, as Doctrine releases are BC, I would even vote for using ``~2.2`` as requirement
---------------------------------------------------------------------------
by fabpot at 2013-01-10T08:18:52Z
If the Doctrine is fine with `~2.2`, let's do that. @beberlei Can you confirm?
---------------------------------------------------------------------------
by stloyd at 2013-01-10T08:30:19Z
As Doctrine latest release is `2.3.2` maybe for Symfony `2.2` we should increase to `~2.3` not `~2.2` ?
---------------------------------------------------------------------------
by stof at 2013-01-10T09:15:00Z
@stloyd Why should we force using 2.3 when we don't require any of the 2.3 features ? Composer will use the most recent package matching the constraint anyway, so 2.2 will not be used unless someone forbids 2.3 in another package.
---------------------------------------------------------------------------
by stloyd at 2013-01-29T19:24:09Z
@fabpot ping
This PR was merged into the master branch.
Commits
-------
924c256 Remove executable bit from all PHP files
Discussion
----------
Remove executable bit from all PHP files
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? |
| Fixed tickets |
| License | MIT
| Doc PR |
Apologies if I haven't got the protocol quite right here. This is my first interaction with this project.
By the way, I've noticed a non-trivial number of pull requests to remove executable bits. Could there be an automated test to prevent files getting into the repo with them on in the first place?
This PR was merged into the master branch.
Commits
-------
d0341e8 Replaced call that triggers deprecated error
Discussion
----------
[Validator][ChoiceValidator] Replaced call that triggers deprecated error
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | ~
| License | MIT
| Doc PR | ~
This removes the deprecation warnings from the ChoiceValidator.
This PR was merged into the 2.1 branch.
Commits
-------
d6c0455 Correct comment in NativeSessionStorage regarding session.save_handler
Discussion
----------
Correct comment in NativeSessionStorage regarding session.save_handler
It's save_handler, not save_handlers.
This PR was merged into the master branch.
Commits
-------
b1d1168 Fixed the NullLogger to implement the HttpKernel interface again
Discussion
----------
Fixed the NullLogger to implement the HttpKernel interface again
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no (reverting one introduced by mistake)
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR | n/a
This BC breaking mistake has been reported in https://github.com/schmittjoh/JMSSecurityExtraBundle/pull/105#issuecomment-12908659
This PR was merged into the master branch.
Commits
-------
ed3a875 [Finder] Fixed test
Discussion
----------
[Finder] Fixed test
Fixed tests broken by PR #6911.
Iterator keys were considered as incremented index in `FilePathsIteratorTest`.