This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#8432).
Discussion
----------
[BrowserKit] CookieJar remove unneeded var, Client remove unneeded else
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Commits
-------
10c6799 [BrowserKit] CookieJar remove unneeded var, Client remove unneeded else
This PR was squashed before being merged into the 2.3 branch (closes#8435).
Discussion
----------
[DI] Fixed bug requesting non existing service from dumped frozen container
When dumping a frozen container without aliases, the method `Container::has()` will throw a *"Warning: array_key_exists() expects parameter 2 to be array, null given"* warning. It's because `PhpDumper::addAliases` returns `''` if no aliases are given. This will work for a normal constructor, because it calls the `Container::__construct` method, but the frozen constructor doesn't do that. So it requires a `$this->aliases = array();` in the constructor.
This PR fixes this bug (and adds the test). Bug is introduced in #8252 and 63367a3d5615ff65251ccee4f73223107275f65f
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Commits
-------
8bb4e4d [DI] Fixed bug requesting non existing service from dumped frozen container
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#8411).
Discussion
----------
Update validators.sk.xlf
updatet pluralization translations for lenght validators
Commits
-------
00c2e5c Update validators.sk.xlf
This PR was merged into the 2.3 branch.
Discussion
----------
[Console] Avoided an unnecessary check if a stream is an interactive terminal
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
It only makes sense to do the second check, if the first one evaluates to false (so the `--no-interaction` option was not passed).
Running `posix_isatty()` raises a warning with some stream types (like in-memory stream). With this patch, warning could be avoided by passing the `--no-interaction option` (which is what we want with an in-memory stream).
The warning I'm getting is:
```
PHP Warning: posix_isatty(): could not use stream of type 'MEMORY' in vendor/symfony/console/Symfony/Component/Console/Application.php on line 866
```
Commits
-------
e46fd4e [Console] Avoided an unnecessary check.
It only makes sense to do the second check if the --no-interaction
option was not passed.
Running posix_isatty() raises a warning with some stream types (like MEMORY).
With this patch, warning could be avoided by passing the --no-interaction option.
This PR was merged into the 2.3 branch.
Discussion
----------
Added missing French validator translations
Commits
-------
0cc8872 Added missing French validator translations
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#8363).
Discussion
----------
typo first->second
Commits
-------
2d9a84b typo first->second
This PR was submitted for the 2.3 branch but it was merged into the 2.2 branch instead (closes#8310).
Discussion
----------
[FrameworkBundle] Fixed variable name used in translation cache
This simply fixes the `$catalogueXXX` variable name used in the translation cache files in case the user use locales such as `en-US`, the generated variable's name was `$catalogueEn-Us`, with this fix it will be `$catalogueEnUs`.
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #7824
| License | MIT
| Doc PR | -
Commits
-------
e50399c [FrameworkBundle] Fixed variable name used in translation cache
* 2.2:
[Form] fixed INF usage which does not work on Solaris (closes#8246)
Fix grammar
bumped Symfony version to 2.2.4
updated VERSION for 2.2.3
update CONTRIBUTORS for 2.2.3
updated CHANGELOG for 2.2.3
[Process] Disable exception on stream_select timeout
[HttpFoundation] fixed issue with session_regenerate_id (closes#7380)
[DomCrawler] added a note about the default charset
[Console] fixed regression when calling a command foo:bar if there is another one like foo:bar:baz (closes#8245)
force the Content-Type to html in the web profiler controllers
Conflicts:
src/Symfony/Component/HttpKernel/Kernel.php
This PR was merged into the 2.3 branch.
Discussion
----------
[Intl] Fixed tests failing on PHP 5.5 and removed this version from the allowed failures
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Commits
-------
69ec7c8 Removed PHP 5.5 from the allowed failures.
3492186 [Intl] Fixed tests failing on PHP 5.5
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#8337).
Discussion
----------
Fix grammar
Commits
-------
4163f71 Fix grammar
This PR was merged into the 2.3 branch.
Discussion
----------
[DependencyInjection] Replaced try/catch block with an @expectedException annotation
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This fix doesn't seem like changing anything but thanks to it, PHP 5.3.3 stops segfaulting.
Tests started failing after merging #8285. However, it's not the change which made them failing (the change is totally unrelated to the segfaulting test case). I suspect it has something to do with memory management in the early PHP 5.3 versions. We've had this kind of failures in the past.
Commits
-------
494e803 [DependencyInjection] Replaced try/catch block with an @expectedException annotation in a test.
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#8287).
Discussion
----------
[CssSelector] added a node about HTML extension in readme
It's a common mistake to use `CssSelector` with XML without knowing that `HTML` extension must be disbaled first (see #8286). This PR adds a note about that in the component's readme file.
Commits
-------
4235b26 [CssSelector] added a node about HTML extension in readme
This PR was merged into the 2.3 branch.
Discussion
----------
[Console] Fixed the table rendering with multi-byte strings.
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #8284
| License | MIT
| Doc PR | -
Commits
-------
ab1439e [Console] Fixed the table rendering with multi-byte strings.
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#8275).
Discussion
----------
Feature/fix unit tests
Fixed two cases where tests were relying on the pcntl extension
Commits
-------
39477dd Feature/fix unit tests
This PR was merged into the 2.3 branch.
Discussion
----------
[Security] fixed usage of the salt for the bcrypt encoder (refs #8210)
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #8210
| License | MIT
| Doc PR | n/a
see #8210
Commits
-------
b5ded81 [Security] fixed usage of the salt for the bcrypt encoder (refs #8210)
This PR was merged into the 2.2 branch.
Discussion
----------
[WebProfilerBundle] force the Content-Type to html in the web profiler controllers
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | [](https://travis-ci.org/lsmith77/symfony)
| Fixed tickets | -
| License | MIT
| Doc PR | -
This just forces the Content-Type to match what will be returned, otherwise if the request format happens to be something else than HTML (which can be the case when building an app that only does JSON/XML with FOSRestBundle) it can happen that the Response class automatically sets a different Content-Type.
The approach taken here matches https://github.com/nelmio/NelmioApiDocBundle/blob/master/Controller/ApiDocController.php#L24
Commits
-------
6d2135b force the Content-Type to html in the web profiler controllers
* 2.2:
Throw exception if value is passed to VALUE_NONE input, long syntax
fixed date type format pattern regex
[FrameworkBundle] tweaked previous merge (refs #8242)
do not re-register commands each time a Console\Application is run
[Process] moved env check to the Process class (refs #8227)
fix issue where $_ENV contains array vals
[DomCrawler] Fix handling file:// without a host
[Form] corrected interface bind() method defined against in deprecation notice
[Finder] Fix SplFileInfo::getContents isn't working with ssh2 protocol
Conflicts:
src/Symfony/Component/Console/Tests/Input/ArgvInputTest.php
src/Symfony/Component/DomCrawler/Link.php
src/Symfony/Component/Form/Form.php
This PR was submitted for the master branch but it was merged into the 2.2 branch instead (closes#8199).
Discussion
----------
[Console] Throw exception if value is passed to VALUE_NONE input optin, long syntax
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | yes
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #8135
| License | MIT
Input options with InputOption::VALUE_NONE accept values in both short and long syntaxes:
- When using the long syntax, no exception is thrown;
- When using short, a "The %s option does not exist" exception is thrown.
This PR only addresses the long syntax case. The short syntax case would require considerable refactoring of the parse code, which I believe should be discussed.
I included a test that illustrates the above mentioned problem for the long syntax scenario.
Commits
-------
32ea77f Throw exception if value is passed to VALUE_NONE input, long syntax
This PR was submitted for the 2.1 branch but it was merged into the 2.2 branch instead (closes#8198).
Discussion
----------
[Form] fixed date type format pattern regex
| Q | A
| ------------- | ---
| Bug fix? | [yes]
| New feature? | [no]
| BC breaks? | [no]
| Deprecations? | [no]
| Tests pass? | [yes]
| License | MIT
I don't understand what is the rationale behind the current regex, why is there mandatory in-between characters?
The current regex passes with the format option set to ```dMyyyy``` while it doesn't with ```dMy```, on the linked icu documentation it is stated that ```y``` is equivalent to ```yyyy```.
So when setting this format option to dMy, fields are rendered in a wrong order because of the fallback (year, month, day).
Commits
-------
6a91bbb [Form] fixed date type format pattern regex
