This PR was merged into the 3.2-dev branch.
Discussion
----------
[Debug] fix resource type test on HHVM
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #19079
| License | MIT
| Doc PR |
Commits
-------
2d03ee8 [Debug] fix resource type test on HHVM
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Console] Add ConsoleLogger::hasErrored()
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
So that it's easy to `exit(1)` whenever any messages have been logged at error levels.
Commits
-------
cfc578a [Console] Add ConsoleLogger::hasErrored()
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Console] Test SymfonyStyle::ask() output
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | ~
| License | MIT
| Doc PR | ~
Now that we can test an interactive command that uses SymfonyStyle (after #18999), we should test their output.
Commits
-------
e870758 [Console] Test SymfonyStyle::ask() output
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Debug] Do not quote numbers in stack trace
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | yes
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
In the debug output from the exception handler, integers and floats are quoted. This adds unnecessary noise to the error page and is just wrong.
Fixing this requires introduces two new type descriptions in the output from getTrace(), so the change is not fully backwards compatible.
Commits
-------
fb10b33 [Debug] Do not quote numbers in stack trace
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Console] progress bar fix
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #13019
| License | MIT
| Doc PR | -
This is #16490 where I've simplified the code as much as possible and added a test for the bug we're trying to fix.
The main change is the renaming of the `TerminalDimensionsProvider` to just `Terminal`. The new class can probably be useful to add more about the terminal.
Commits
-------
2f81247 switched to use COLUMNS and LINES env vars to change terminal dimensions
bf7a5c5 fixed logic
a589635 deprecated some Console Application methods
8f206c8 fixed CS, simplified code
b030c24 [Console] ProgressBar - adjust to the window width (static)
This PR was merged into the 3.2-dev branch.
Discussion
----------
Deprecate using Form::isValid() with an unsubmitted form
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | yes
| Tests pass? | yes
| Fixed tickets | #7737
| License | MIT
| Doc PR | not yet
> I'm calling out for you to decide how to resolve the inconsistency between the Form::isValid() method and the ``valid`` variable available in the template:
>
> - ``isValid()`` returns ``false`` if a form was not submitted. This way it is possible to write concise controller code:
>
> ```php
> $form = $this->createForm(...);
> $form->handleRequest($request);
> if ($form->isValid()) {
> // only executed if the form is submitted AND valid
> }
> ```
>
> - ``valid`` contains ``true`` if a form was not submitted. This way it is possible to rely on this variable for error styling of a form.
>
> ```twig
> <div{% if not form.vars.valid %} class="error"{% endif %}>
> ```
>
> We have two alternatives for resolving this problem:
>
> 1. Leave the inconsistency as is.
> 2. Make ``isValid()`` return ``true`` if a form was not submitted (consistent with ``valid``)
> 3. Revert to the 2.2 behavior of throwing an exception if ``isValid()`` is called on a non-submitted form (not consistent with ``valid``).
>
> Both 2. and 3. will require additional code in the controller:
>
> ```php
> $form = $this->createForm(...);
> $form->handleRequest($request);
> if ($form->isSubmitted() && $form->isValid()) {
> // only executed if the form is submitted AND valid
> }
> ```
>
> What do you think?
This PR implements the option 3 as it was the most chosen in #7737
Commits
-------
2c3a7cc Deprecate using Form::isValid() with an unsubmitted form
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Routing] treat fragment after resolving query string params
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
The implementation from #12979 led to a conflict with #18280 which was merged in the meantime.
Commits
-------
7475aa8 treat fragment after resolving query string params
This PR was squashed before being merged into the 3.2-dev branch (closes#19029).
Discussion
----------
[YAML] Fixed parsing problem with nested DateTime lists
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
___
Without this fix, DateTime-aware parsing of a YAML source containing nested lists of dates result in an error. Consider this:
```php
$data = ['date' => ['annivesary' => new \DateTime('now')]];
$yaml = Yaml::dump($data);
var_dump($yaml);
$parsed = Yaml::parse($yaml, Yaml::PARSE_DATETIME);
print_r($parsed);
```
Everything is fine, result is:
```
string(48) "date:
annivesary: 2016-06-11T11:26:30+02:00
"
Array
(
[date] => Array
(
[annivesary] => DateTime Object
(
[date] => 2016-06-11 11:26:30.000000
[timezone_type] => 1
[timezone] => +02:00
)
)
)
```
But making the `anniversary` a list of dates
```php
$data = ['date' => ['annivesary' => [new \DateTime('now')]]];
$yaml = Yaml::dump($data);
var_dump($yaml);
$parsed = Yaml::parse($yaml, Yaml::PARSE_DATETIME);
print_r($parsed);
```
will result in:
```
string(50) "date:
annivesary: [2016-06-11T12:00:05+02:00]
"
PHP Warning: strpos() expects parameter 1 to be string, object given in [...]\vendor\symfony\yaml\Inline.php on line 382
PHP Catchable fatal error: Object of class DateTime could not be converted to string in [...]\vendor\symfony\yaml\Inline.php on line 386
```
(I didn't capture the error messages with the most recent master branch, so line numbers differ somewhat)
Commits
-------
52384cf [YAML] Fixed parsing problem with nested DateTime lists
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Router] added appending of new optional document fragment
added a new optional parameter to the generate method for the document fragment. when specified this is appended to generated urls.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | none
| License | MIT
| Doc PR | none
Commits
-------
6d79a56 [Routing] adds _fragment special option to url generation for document fragment
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Console] Simplify simulation of user inputs in CommandTester
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | https://github.com/symfony/symfony-docs/pull/6623
After @javiereguiluz pointed it in #17470, I open this PR to simplify the simulation of user inputs for testing a Command.
It would be done by calling `CommandTester::setUserInputs()` with an array of inputs as argument, and so make the CommandTester creating an input stream from the inputs set by the developer, then call `QuestionHelper::setInputStream` and assign it to the helperSet of the command, sort as all is done automatically in one call.
Depends on #18999
Commits
-------
c7ba38a [Console] Set user inputs from CommandTester
This PR was merged into the 3.2-dev branch.
Discussion
----------
Fix the missing colon in get/setInputStream deprecation warning
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony/pull/19072#discussion_r67310243
| License | MIT
| Doc PR | ~
Commits
-------
0d6dc8e Fix the missing colon in get/setInputStream deprecation warning
* 3.1:
Tweak merge
update tests to use the new error assertion helper
[ci] Upgrade to symfony/phpunit-bridge >=3.2@dev
update tests to use the new error assertion helper
Conflicts:
src/Symfony/Component/DependencyInjection/Tests/ContainerBuilderTest.php
This PR was merged into the 3.1 branch.
Discussion
----------
[3.1] update tests to use the new error assertion helper
| Q | A
| ------------- | ---
| Branch? | 3.1
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Commits
-------
9c7512b update tests to use the new error assertion helper
* 2.8:
update tests to use the new error assertion helper
[ci] Upgrade to symfony/phpunit-bridge >=3.2@dev
Conflicts:
src/Symfony/Component/DependencyInjection/Tests/ContainerBuilderTest.php
src/Symfony/Component/Yaml/Tests/ParserTest.php
This PR was merged into the 2.8 branch.
Discussion
----------
[2.8] update tests to use the new error assertion helper
| Q | A
| ------------- | ---
| Branch? | 2.8
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Commits
-------
e5cf4d4 update tests to use the new error assertion helper
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Console] Centralize input stream in base Input class
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | yes
| Tests pass? | yes
| Fixed tickets | #10844
| License | MIT
| Doc PR | not yet
Actually we have two classes that have an `inputStream` property with its getter+setter:
the QuestionHelper ('question' helper) and the SymfonyQuestionHelper (used via SymfonyStyle, that inherits the inputStream from the QuestionHelper class).
This situation makes command testing really hard.
ATM we can't test a command that uses `SymfonyStyle::ask` without ugly hacks, and we can't find a generic way to set the inputs passed to a command from the CommandTester, as it need to be done on a specific helper that may not be the one used by the command (SymfonyStyle case).
What I propose here is to add a `stream` property (and its getter+setter) to the abstract `Symfony\Component\Console\Input\Input` class.
For now I just made the two helpers setting their `inputStream` from `Input::getStream`, as both QuestionHelper and SymfonyQuestionHelper classes have an `ask` method that takes the command Input as argument.
In a next time (4.0), we could remove the `getInputStream` and `setInputStream` methods from the QuestionHelper class (this only deprecates them in favor of Input `setStream` and `getStream`).
This would close PR #18902 (trying to make interactive command testing with SymfonyStyle easier).
This would also make PR #18710 widely better by setting the input stream in a generic way (working for both helpers without caring about if they are used and which one is used).
Please give me your opinions.
Commits
-------
54d3d63 Centralize input stream in base Input class
