This PR was merged into the 4.1-dev branch.
Discussion
----------
[Routing] Allow inline definition of requirements and defaults
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #26481
| License | MIT
| Doc PR | -
```
{bar} -- no requirement, no default value
{bar<.*>} -- with requirement, no default value
{bar?default_value} -- no requirement, with default value
{bar<.*>?default_value} -- with requirement and default value
{bar?} -- no requirement, with default value of null
{bar<.*>?} -- with requirement and default value of null
```
Details:
* Requirements and default values are not escaped in any way. This is valid -> `@Route("/foo/{bar<>>?<>}")` (requirements = `>` and default value = `<>`)
* Because of the lack of escaping, you can't use a closing brace (`}`) inside the default value (wrong -> `@Route("/foo/{bar<\d+>?aa}bb}")`) but you can use it inside requirements (correct -> `@Route("/foo/{bar<\d{3}>}")`).
* PHP constants are not supported (use the traditional `defaults` syntax for that)
* ...
Commits
-------
67559e1 [Routing] Allow inline definition of requirements and defaults
This PR was merged into the 4.1-dev branch.
Discussion
----------
[Routing] Implement i18n routing
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? |no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
This PR introduces support for I18N routing into core. This is a port from a bundle I've made recently, now merged into the default implementation. While it's ok to have this as a bundle, it was suggested by @nicolas-grekas to create a PR for this so it can be included into the core.
## New usages
### YAML
```yaml
contact:
controller: ContactController::formAction
path:
en: /send-us-an-email
nl: /stuur-ons-een-email
```
Will be effectively the same as declaring:
```yaml
contact.en:
controller: ContactController::formAction
path: /send-us-an-email
defaults:
_locale: en
contact.nl:
controller: ContactController::formAction
path: /stuur-ons-een-email
defaults:
_locale: nl
```
### Annotation usage:
```php
<?php
use Symfony\Component\Routing\Annotation\Route;
class ContactController
{
/**
* @Route({"en": "/send-us-an-email", "nl": "/stuur-ons-een-email"}, name="contact")
*/
public function formAction()
{
}
}
/**
* @Route("/contact")
*/
class PrefixedContactController
{
/**
* @Route({"en": "/send-us-an-email", "nl": "/stuur-ons-een-email"}, name="contact")
*/
public function formAction()
{
}
}
```
### Route generation
```php
<?php
/** @var UrlGeneratorInterface $urlGenerator */
$urlWithCurrentLocale = $urlGenerator->generate('contact');
$urlWithSpecifiedLocale = $urlGenerator->generate('contact', ['_locale' => 'nl']);
```
Route generation is based on your request locale. When not available it falls back on a configured default. This way of route generation means you have a "route locale switcher" out of the box, but generate the current route with another locale for most cases.
## Advantages
Having i18n routes defined like this has some advantages:
* Less error prone.
* No need to keep `requirements` or `defaults` in sync with other definitions.
* No need to `{_locale}` in the path (bad for route matching performance).
* Better developer experience.
### Next steps
I've ported all the things the bundle supported, before moving on I'd like to discuss this first in order not to waste our collective time. This initial PR should give a clear enough picture to see what/how/why this is done.
If and when accepted I/we can move forward to implement the XML loader and @nicolas-grekas mentioned there should be a `Configurator` implemented for this as well. He opted to help with this (for which I'm very thankful).
- [x] Yaml Loader
- [x] Annotation Loader
- [x] XML Loader
- [x] PHP Loader?
- [ ] Documentation
Commits
-------
4ae66dc [Routing] Handle "_canonical_route"
e32c414 [Routing] Implement i18n routing
This PR was merged into the 3.4 branch.
Discussion
----------
Make sure form errors is valid HTML
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #26461
| License | MIT
| Doc PR |
Using `<div>` and `<ul>` in a `<label>` is not valid HTML. This PR uses `<span>`s instead and some Bootstrap 4 classes to make it look just like before.
Commits
-------
5dcaaadb5e Make sure form errors is valid HTML
This PR was merged into the 4.1-dev branch.
Discussion
----------
[Console] Make pretty the `box` style table
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | /
| License | MIT
| Doc PR | /
The console component is the 2nd most popular Symfony component with 70M of downloads. I think such composant has to provide faultless tables, with perfect management of borders, especially since the new `box` style table has been merged (cf https://github.com/symfony/symfony/pull/25301).
Commits
-------
41f52b3a09 Make pretty the `box` style table
This PR was merged into the 2.7 branch.
Discussion
----------
[BrowserKit] Improves CookieJar::get
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? |no
| New feature? | no <!-- don't forget to update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | no <!-- don't forget to update UPGRADE-*.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
* change a call to `substr` + `strlen` to a single `strpos`
Commits
-------
57993143a9 [BrowserKit] Improves CookieJar::get
This PR was squashed before being merged into the 4.1-dev branch (closes#26490).
Discussion
----------
[DomCrawler] Improve the sprintf() call in selectButton()
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? |no <!-- don't forget to update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? |no <!-- don't forget to update UPGRADE-*.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | n/a <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR | n/a
Code cleanup and minor performance improvement. Similar to #26485.
Commits
-------
a00cce6 [DomCrawler] Improve the sprintf() call in selectButton()
This PR was merged into the 4.1-dev branch.
Discussion
----------
[BrowserKit] Avoid nullable values in some Client's methods
| Q | A
| ------------- | ---
| Branch? | master | Bug fix? | yes/no
| New feature? | no <!-- don't forget to update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | yes <!-- don't forget to update UPGRADE-*.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| License | MIT
<!--
Write a short README entry for your feature/bugfix here (replace this comment block.)
This will help people understand your PR and can be used as a start of the Doc PR.
Additionally:
- Bug fixes must be submitted against the lowest branch where they apply
(lowest branches are regularly merged to upper ones so they get the fixes too).
- Features and deprecations must be submitted against the master branch.
-->
I suggest the some methods in `Client` should not return `null` for simplify usage.
If you are trying to get response from client when `request` method was not called, it seems an exception should be throwed.
Commits
-------
c2c285355b Avoid nullable values in some Client's methods
This PR was squashed before being merged into the 4.1-dev branch (closes#26417).
Discussion
----------
[Console] Improved rendering of optional arguments in command synopsis
Current rendering:
```
build [options] [--] [<file>] [<output-dir>] [<packages>]...
```
Fixed (and actually correct) rendering:
```
build [options] [--] [<file> [<output-dir> [<packages>...]]]
```
Also dropped duplicating required array-type argument. There's just no need for that, it only confuses the reader.
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | yes and no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | --
| License | MIT
| Doc PR | --
Optional arguments can not be independent, in general. Neither the argument parser allows that.
Commits
-------
938012f0ea [Console] Improved rendering of optional arguments in command synopsis
This PR was squashed before being merged into the 4.1-dev branch (closes#26423).
Discussion
----------
[DomCrawler] minor performance improvement
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no <!-- don't forget to update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? |no <!-- don't forget to update UPGRADE-*.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | n/a <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR | n/a
Commits
-------
eead301 [DomCrawler] minor performance improvement
This PR was merged into the 3.4 branch.
Discussion
----------
[Fix][3.4][HttpFoundation] Fix the updating of timestamp in the MemcachedSessionHandler
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
Conditions: Symfony 3.4, PHP7 and sessions handled over memcache.
Apparently `memcached::touch()` returns `false` on a subsequent call with the same parameters. Since `updateTimestamp` is used in `AbstractSessionHandler::write()`
```
public function write($sessionId, $data)
{
if (\PHP_VERSION_ID < 70000 && $this->prefetchData) {
$readData = $this->prefetchData;
$this->prefetchData = null;
if ($readData === $data) {
return $this->updateTimestamp($sessionId, $data);
}
}
...
```
the result is that `write()` will return `false` on **any subsequent request within the same second** causing the following error:
```
HP Fatal error: Uncaught Symfony\Component\Debug\Exception\ContextErrorException: Warning: session_write_close(): Failed to write session data (user). Please verify that the current setting of session.save_path is correct (/var/lib/php/sessions) in Unknown:0
Stack trace:
#0 [internal function]: Symfony\Component\Debug\ErrorHandler->handleError(2, 'session_write_c...', 'Unknown', 0, NULL)
#1 [internal function]: session_write_close()
#2 {main}
thrown in Unknown on line 0
```
Can be reproduced on `symfony/skeleton:3.4` adding the following code to `public/index.php` and performing two consecutive requests:
```
$session = $kernel->getContainer()->get('session');
$session->set("foo", "bar");
```
Commits
-------
d007469877 fix the updating of timestamp in the MemcachedSessionHandler
This PR was merged into the 3.4 branch.
Discussion
----------
[Config] Add characters to the regex
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
Commits
-------
dff88d44be [Config] Add characters to the regex
This PR was merged into the 3.4 branch.
Discussion
----------
[Config] ReflectionClassResource check abstract class
generate Signature
update hash methods `ServiceSubscriberInterface::getSubscribedServices` and `EventSubscriberInterface::getSubscribedEvents` if the class is not abstract
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #26459, #26501
| License | MIT
Commits
-------
e85151483c [Config] ReflectionClassResource check abstract ServiceSubscriberInterface and EventSubscriberInterface
This PR was merged into the 2.7 branch.
Discussion
----------
[DomCrawler] extract(): fix a bug when the attribute list is empty
| Q | A
| ------------- | ---
| Branch? | 2.7
| Bug fix? | yes
| New feature? | no <!-- don't forget to update src/**/CHANGELOG.md files -->
| BC breaks? | no <!-- see https://symfony.com/bc -->
| Deprecations? | no <!-- don't forget to update UPGRADE-*.md files -->
| Tests pass? | yes <!-- please add some, will be required by reviewers -->
| Fixed tickets | n/a <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR | n/a
Commits
-------
028c9f2366 [DomCrawler] extract(): fix a bug when the attribute list is empty
This PR was merged into the 4.1-dev branch.
Discussion
----------
Default deleteFileAfterSend() to true
This should default to true when called since there is only one option.
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | N/A
| Fixed tickets | N/A
| License | MIT
| Doc PR |
<!--
* Defaults deleteFileAfterSend() to true
-->
Commits
-------
1de6f738ef Default deleteFileAfterSend() to true
This PR was squashed before being merged into the 3.4 branch (closes#26041).
Discussion
----------
Display the Welcome Page when there is no homepage defined
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony-docs/issues/9178
| License | MIT
| Doc PR | -
In 3.4 we added a trick to display the Welcome Page when the user browses `/` and there are no routes defined. However, when using the `website-skeleton` (which is what most newcomers use ... and they are the ones that mostly need the "Welcome Page") the premise about *"no routes are defined"* is never true and the Welcome Page is never shown (see https://github.com/symfony/symfony-docs/issues/9178 for one of the multiple error reports we've received).
So, I propose to make this change to always define the "Welcome Page" as the fallback:
* If no routes are defined for `/`, the Welcome Page is displayed.
* If there is a route defined for `/`, this code will never be executed because it's the last condition of the routing matcher.
Commits
-------
5b0d9340d7 Display the Welcome Page when there is no homepage defined
This PR was merged into the 4.1-dev branch.
Discussion
----------
[Security] AuthenticationUtils::getLastUsername() return type inconsistency
Always return `string`, never `null` according to the `@return` annotation tag.
| Q | A
| ------------- | ---
| Branch? | 2.8
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Alternatively, string return might be nullable:
```php
return null === $session ? null : $session->get(Security::LAST_USERNAME);
```
Is test needed for this change?
Commits
-------
743692c3fd AuthenticationUtils::getLastUsername()` now always returns a string.
This PR was squashed before being merged into the 4.1-dev branch (closes#26467).
Discussion
----------
Add UsernameNotFoundException declaration to refreshUser().
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | maybe
| New feature? | maybe
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Symfony\Component\Security\Core\User\UserProviderInterface::refreshUser() does not declare that implementations may throw a UsernameNotFoundException, although a) it makes sense, as the user could have been deleted since the last load, and b) the ContextListener already handles the UsernameNotFoundException. So it looks like someone thought of this, but simply forgot the annotation.
Unsure if this is a bugfix or a feature, but as it doesn't change executed code, master should be soon enough.
Commits
-------
f7a0c46338 Add UsernameNotFoundException declaration to refreshUser().
* 2.7:
Php Inspections (EA Ultimate): address some of one-time used local variables
[Intl] Load locale aliases to support alias fallbacks
[CssSelector] Fix CSS identifiers parsing - they can start with dash
This PR was squashed before being merged into the 2.7 branch (closes#26452).
Discussion
----------
[Intl] Load locale aliases to support alias fallbacks
| Q | A
| ------------- | ---
| Branch? | 2.7
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #21457
| License | MIT
| Doc PR | -
For example, `zh_TW` is an alias to `zh_Hant_TW`. Without aliases,` zh_TW` would fall back to `zh` (which is incorrect). With aliases loaded, `zh_TW` will fall back properly to `zh_Hant_TW`.
Judging by git history this has never worked.
```php
\Locale::setDefault('zh'); dump(Intl::getRegionBundle()->getCountryName('AD'));
\Locale::setDefault('zh_TW'); dump(Intl::getRegionBundle()->getCountryName('AD'));
\Locale::setDefault('zh_Hant_TW'); dump(Intl::getRegionBundle()->getCountryName('AD'));
```
Before:
```
"安道尔"
"安道尔"
"安道爾"
```
After:
```
"安道尔"
"安道爾"
"安道爾"
```
All tests are passing, including those from the `intl-data` group.
Commits
-------
1debf79430 [Intl] Load locale aliases to support alias fallbacks
* 4.0:
fix the handling of timestamp in the MongoDBSessionHandler
Improve the documentation of `Finder::exclude()`
[DI] Skip resource tracking if disabled
[WebProfilerBundle] fix wrong variable for profiler counting ajax requests
This PR was merged into the 4.1-dev branch.
Discussion
----------
Revert "feature #24763 [Process] Allow writing portable "prepared" command lines (Simperfit)"
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This reverts commit 136408937b, reversing
changes made to e043478ba5.
As discussed in #24763 and #26344
This doens't revert the possibility to use prepared command lines. They just won't be *portable* anymore, unless special care is taken by "userland".
Ie the placeholders need to be shell-dependent: use eg `echo "$FOO"` on *nix (the double quotes *are* important), and `echo !FOO!` on Windows (no double quotes there).
Commits
-------
6a98bfa8d7 Revert "feature #24763 [Process] Allow writing portable "prepared" command lines (Simperfit)"
* 4.0:
[DI] Fix missing "id" normalization when dumping the container
Add entry for `container.dumper.inline_class_loader` param at `UPGRADE-3.4.md` and `UPGRADE-4.0.md`
* 3.4:
[DI] Fix missing "id" normalization when dumping the container
Add entry for `container.dumper.inline_class_loader` param at `UPGRADE-3.4.md` and `UPGRADE-4.0.md`
This PR was squashed before being merged into the 4.1-dev branch (closes#26223).
Discussion
----------
[FrameworkBundle] Add command to delete an item from a cache pool
| 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 | TBD
Currently there is no way to clear a specific item from a cache pool (except programatically), the entire pool needs to be cleared.
Especially during development, when implementing caching, it is useful to delete a specific key to test functionality. Clearing the entire pool, means that everything will need to be cached again, adding unnecessary execution time.
I propose adding a new command, `cache:pool:delete` to delete a specific item from a cache pool
Commits
-------
fd43e81 [FrameworkBundle] Add command to delete an item from a cache pool
This PR was squashed before being merged into the 3.4 branch (closes#26327).
Discussion
----------
[Form][WCAG] Errors sign for people that do not see colors
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
According to my friend and WCAG2 expect [Sandra](https://twitter.com/sandrability):
> The form errors is correctly encoded and works great. But visually they may be hard to see for people that do not see colors very well. Try to improve errors with an icon to make it more visual clear that an error has occurred.
Commits
-------
3f8cd05 [Form][WCAG] Errors sign for people that do not see colors
* 4.0:
[Form][WCAG] Add hidden labels on date and time fields
Pass on previous exception in FatalThrowableError
[Routing] remove dead code
[Routing] fix typo
[Form][WCAG] Fixed HTML errors
fix merge
[FrameworkBundle] [Console] add a warning when command is not found
[WebProfilerBundle] limit ajax request to 100 and remove the last one
* 3.4:
[Form][WCAG] Add hidden labels on date and time fields
Pass on previous exception in FatalThrowableError
[Routing] remove dead code
[Routing] fix typo
[Form][WCAG] Fixed HTML errors
fix merge
[FrameworkBundle] [Console] add a warning when command is not found
[WebProfilerBundle] limit ajax request to 100 and remove the last one
* 4.0:
Fix typos
[Routing] remove useless failing mocks
[appveyor] Workaround GitHub disabling of low versions of TLS
Use long array syntax
[Routing] Fix GC control of PHP-DSL
[Routing] Don't throw 405 when scheme requirement doesn't match
[Routing] Revert throwing 405 on missed slash/scheme redirections
[WebProfilerBundle] fix test after ajax path updated
Fix ArrayInput::toString() for InputArgument::IS_ARRAY args
Update excluded_ajax_paths for sf4
Add missing use for RoleInterface
Add missing use of Role
[Routing] fix CS
add container.autowiring.strict_mode to 3.4 docs
Set controller without __invoke method from invokable class
[VarDumper] Fixed PHPDoc
* 3.4:
[Routing] remove useless failing mocks
[appveyor] Workaround GitHub disabling of low versions of TLS
Use long array syntax
[Routing] Fix GC control of PHP-DSL
[Routing] Don't throw 405 when scheme requirement doesn't match
[Routing] Revert throwing 405 on missed slash/scheme redirections
[WebProfilerBundle] fix test after ajax path updated
Fix ArrayInput::toString() for InputArgument::IS_ARRAY args
Update excluded_ajax_paths for sf4
Add missing use for RoleInterface
[Routing] fix CS
add container.autowiring.strict_mode to 3.4 docs
Set controller without __invoke method from invokable class
[VarDumper] Fixed PHPDoc
This PR was merged into the 3.4 branch.
Discussion
----------
[Routing] Fix GC control of PHP-DSL
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Adding fluently in a collection is broken currently.
Commits
-------
239f2e2 [Routing] Fix GC control of PHP-DSL
This PR was merged into the 4.1-dev branch.
Discussion
----------
[Routing] Redirect from trailing slash to no-slash when possible
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #26207
| License | MIT
| Doc PR | -
Implemented as suggest by @Tobion in https://github.com/symfony/symfony/pull/26059#issuecomment-365071281
When a route for `/foo` exists but the request is for `/foo/`, we now redirect.
(this complements the flipped side redirection, which already exists.)
Commits
-------
69a4e94130 [Routing] Redirect from trailing slash to no-slash when possible
This PR was merged into the 4.1-dev branch.
Discussion
----------
[FrameworkBundle] fix lowest supported Serializer version
| 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 `AbstractObjectNormalizer::setMaxDepthHandler()` method does not
exist before `symfony/serializer` 4.1.
Commits
-------
3b092a08ea fix lowest supported Serializer version
This PR was merged into the 3.4 branch.
Discussion
----------
[PropertyInfo] throw exception if docblock factory does not exist
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #26259
| License | MIT
| Doc PR |
Commits
-------
5cfceed throw exception if docblock factory does not exist
