This PR was merged into the 2.7 branch.
Discussion
----------
Normalize the way we check versions
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
Commits
-------
121eade Normalize the way we check versions
* 2.3:
Finnish translation fix
[CssSelector] Optimize regexs matching simple selectors
Fix the phpdoc in the CssSelector TranslatorInterface
[Console] Add clock mock to fix transient test on HHVM
[EventDispatcher] skip one lazy loading call
[EventDispatcher] fix memory leak in a getListeners
This PR was merged into the 2.7 branch.
Discussion
----------
[Console] Default to stderr for the console helpers (2.7+)
Interactive input/output and informational output such as progress should go to `stderr` if available.
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Only merge if #15794 is merged.
If someone explicitly wants to use `stdout`, they can simply pass `$output->getStream()` instead of `$output` in most use-cases.
Commits
-------
90c2a96 Default to stderr for console helpers (only merge if #15794 gets merged)
This PR was merged into the 2.7 branch.
Discussion
----------
[Config] Fix enum default value in Yaml dumper
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
The default value is not correctly included when dumping an EnumNode in Yaml. This is now fixed
Commits
-------
d135d82 [Config] Fix enum default value in Yaml dumper
This PR was submitted for the 2.8 branch but it was merged into the 2.3 branch instead (closes#15843).
Discussion
----------
Finnish translation fix
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15790
| License | MIT
| Doc PR | -
Fix for finnish translation typo
https://github.com/symfony/symfony/issues/15790
Commits
-------
84fbfb0 Finnish translation fix
This PR was merged into the 2.3 branch.
Discussion
----------
[CssSelector] Optimize regexs matching simple selectors
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
These shortcut parsers are applied first when converting a CSS selector to XPath, to be faster for simple selectors (tag matching, class matching with an optional tag, id matching with an optional tag).
None of the regexes defined here could have more chances to match more element when backtracking identifiers. So the backtracking is only slowing down the regex engine when the regex does not match (i.e. for any more complex selector for instance, or even for simple selectors without namespace of without tag name). Making quantifiers possessive solves this issue.
I also turned some capturing groups (around the namespace and the namespace delimiter) into non-capturing groups as we don't care about them in the output (they are just here to be optional).
Commits
-------
d5abe0b [CssSelector] Optimize regexs matching simple selectors
This PR was merged into the 2.3 branch.
Discussion
----------
[Console] Add clock mock to fix transient test on HHVM
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
This should fix the most frequent transient test on HHVM (ProgressBarTest::testAnsiColorsAndEmojis)
Commits
-------
549f43b [Console] Add clock mock to fix transient test on HHVM
This PR was merged into the 2.3 branch.
Discussion
----------
Fix the phpdoc in the CssSelector TranslatorInterface
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
The phpdoc of the interface does not match the behavior of the implementation. And the return type documented by the interface is impossible to return because the interface asks to apply a string prefix, and the only way is to cast the ``XPathExpr`` used internally to a string (which is what the implementation does).
This interface is an internal interface anyway.
Commits
-------
93ffa61 Fix the phpdoc in the CssSelector TranslatorInterface
This PR was merged into the 2.7 branch.
Discussion
----------
[DomCrawler] Optimize the regex used to find namespace prefixes
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
I ran Blackfire on the [Mink BrowserKitDriver](https://github.com/minkphp/MinkBrowserKitDriver) testsuite (which is a realistic usage of the Mink API btw, as the driver testsuite is basically a functional test of a fixtures website, where failures are considered to be caused by the driver misbehaving).
I discovered that 17% of the time was spent in ``preg_match_all`` to look for prefixes in the XPath namespace prefixes (Mink does not use any namespace prefix in its Xpath queries).
This optimizes the regex to use a possessive quantifier. This avoids useless backtracking when matching (due to the way the regex is written, backtracking cannot allow finding more matches here). See https://blackfire.io/profiles/compare/21aaebf3-e38f-456a-8fc9-cf7d9e2a35eb/graph for the improvement.
The optimization is applied in the 2.7 branch, because this regex does not exist in 2.3: the registration of namespaces was added in 2.4
Commits
-------
02818e4 [DomCrawler] Optimize the regex used to find namespace prefixes
This PR was submitted for the master branch but it was merged into the 2.7 branch instead (closes#15836).
Discussion
----------
[Form] fixed something: Remove whitespaces between <input> and <label>
| Q | A
| ------------- | ---
| Bug fix? | [no]
| New feature? | [no]
| BC breaks? | [?] What means BC?
| Deprecations? | [no]
| Tests pass? | [yes]
| Fixed tickets | []
| License | MIT
| Doc PR | []
Reason: With the whitespaces, there would be a "gap" between `<input>` and `<label>` which is clickable, but doesn't activate the `<input>`.
It's easy to *add* some margin with CSS, whereas it's impossible to *remove* whitespaces with CSS.
Commits
-------
1b97c17 Remove whitespaces between <input> and <label>
Reason: With the whitespaces, there would be a "gap" between <input> and <label> which is clickable, but doesn't activate the <input>.
It's easy to *add* some margin with CSS, whereas it's impossible to *remove* whitespaces with CSS.
This PR was merged into the 2.3 branch.
Discussion
----------
[Process] Stopped Autoruns when shelling on windows
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? |
| Fixed tickets |
| License | MIT
| Doc PR |
Stop autoruns when shelling on Windows.
Original PR was on https://github.com/symfony/Process/pull/9#issuecomment-140721398.
Commits
-------
bca6771 Update Process.php
This PR was merged into the 2.3 branch.
Discussion
----------
[travis] Remove PHP 7 from allowed failures
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Looking at the latests PRs, PHP 7 does not segfault anymore with our test suite.
Commits
-------
ba8366e [travis] Remove PHP 7 from allowed failures
This PR was merged into the 2.3 branch.
Discussion
----------
[Finder] Optimize the hot-path
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15824
| License | MIT
| Doc PR | -
A significant part of the perf gain in #15802 was related to filters not being applied recursively...
#15824 fixing this, performance dropped again.
This PR optimizes the hot path by replacing a regexp test by a simple `isset` when possible.
Blackfire diff after #15824 is the following:
https://blackfire.io/profiles/compare/9e489018-998d-4acb-92a0-46011828e83b/graph
`preg_match` is not called anymore, and `Symfony\Component\Finder\Iterator\RecursiveDirectoryIterator::current()` is also cut by two.
When this `isset` optimization is disabled and replaced by a concatenation of all the regexps patterns in a single bigger one, the gain is still significant but lower:
https://blackfire.io/profiles/compare/db86b80e-b63e-4fc9-9ff3-9ed32baeb948/graph
This makes me think that an other and last round of optimization is possible by merging all regexps in one in MultiplePcreFilterIterator filters. If someone wants to work on this, please do it :)
Commits
-------
f156de6 [Finder] Optimize the hot-path
* 2.3:
[Finder] Fix recursive filter iterator
Improve the structure of the Finder testsuite
Conflicts:
src/Symfony/Component/Finder/Iterator/ExcludeDirectoryFilterIterator.php
src/Symfony/Component/Finder/Tests/FinderTest.php
src/Symfony/Component/Finder/Tests/Iterator/DateRangeFilterIteratorTest.php
This PR was merged into the 2.3 branch.
Discussion
----------
[Finder] Fix recursive filter iterator
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15802
| License | MIT
| Doc PR | -
#15802 was broken because the filters where not propagated to children iterators. This fixes the issue and adds a test case for this situation. The RecursiveIterator implementation is moved from FilterIterator to ExcludeDirectoryFilterIterator. Doing so on other filters is possible but would be a new feature that is not required for fixing the performance issue we had previously.
Commits
-------
cf3019b [Finder] Fix recursive filter iterator
This PR was merged into the 2.3 branch.
Discussion
----------
Improve the structure of the Finder testsuite
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Testing against different adapters is now handled by multiple subclasses of a common test case rather than using data providers. This allows tests to be marked as skipped for unsupported adapters instead of making them disappear from the testsuite.
This will also make it much easier to mark tests for the find-based adapters as legacy ones when deprecating them.
This correspond to changes done by @nicolas-grekas in https://github.com/symfony/symfony/pull/15805, backported to 2.3 and with the data providers removed (his PR was keeping data providers for the adapter but making it return a single adapter all the time, hence the much bigger diff in my PR).
All tests in AbstractFinderTest are just moved from the FinderTest (they are all the tests running against multiple adapters). FinderTest itself runs the PhpAdapter (as this is the logic we want to keep in the finder in 3.0). I also tried to have a PhpFinderTest and keeping FinderTest only with the extra tests, but this would make things harder to merge branches between 2.8 and 3.0 in the future (once we remove other adapter tests) as we could simply keep the AbstractFinderTest in 3.0 for now to ease merging PRs adding new tests for bug fixes.
Commits
-------
20f2d03 Improve the structure of the Finder testsuite
Testing against different adapters is now handled by multiple subclasses
of a common test case rather than using data providers. This allows
tests to be marked as skipped for unsupported adapters instead of making
them disappear from the testsuite.
* 2.3:
[Finder] Handle filtering of recursive iterators and use it to skip looping over excluded directories
Exclude files based on path before applying the sorting
[Console] fix phpdoc of DialogHelper
This PR was submitted for the 2.8 branch but it was merged into the 2.3 branch instead (closes#15802).
Discussion
----------
[Finder] Handle filtering of recursive iterators and use it to skip looping over excluded directories
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #5951, #8685
| License | MIT
| Doc PR | -
By implementing RecursiveIterator in FilterIterator, we can make it able to skip children of excluded branches of recursive inner iterators.
We use it immediately for our main target: skip over children of excluded directories in the Finder.
This is a significant performance boost when iterating over big directories, thus the "bugfix" status.
Commits
-------
8c691bd [Finder] Handle filtering of recursive iterators and use it to skip looping over excluded directories
This PR was merged into the 2.3 branch.
Discussion
----------
[Finder] Exclude files based on path before applying the sorting
| Q | A
| ------------- | ---
| Bug fix? | yes (perf)
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Sorting can be slow, so it is best to applied it after all filtering rules are applied instead of using a bigger list and filtering again after that.
This impacts only people using both the sorting and the name filtering (with ``->names()`` or ``notNames()``)
Commits
-------
e5bf0ab Exclude files based on path before applying the sorting
This PR was merged into the 2.7 branch.
Discussion
----------
Automatically process extensions when they implement CompilerPassInterface
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | todo (if people are in favor of this PR)
Compiler passes are very powerfull, but also quite strange to work with. Especially when you just need a very simple compiler pass (like https://github.com/symfony-cmf/RoutingBundle/blob/master/DependencyInjection/Compiler/SetRouterPass.php). For 3 lines of code, you need to tweak your bundle class and create a new class.
When using the DI component standalone, compiler passes are even harder to work with, as DI extensions can't register them. I believe that's why libraries like Behat make their extensions compiler passes by default.
I think it would be very easy to just implement an interface and have a `compile` method for the simple compiler pass stuff. If a bundle needs multiple compiler passes or need compiler passes to be executed at other times in the compile process, a bundle can use the normal compiler passes. But if it's just one simple thing, like replacing a definition or getting services with a specific tag, I think this method will be very usefull.
Commits
-------
6c50013 Allowed extensions to inline compiler passes
Improve performance for the following scenarios:
- Example 1:
```php
$a = array(
'a' => array(
'b' => array(
'c' => 'old-value'
)
)
);
$pa->setValue($a, '[a][b][c]', 'new-value');
// The PropertyAccessor will try to set values for
// $a['a']['b']['c'], $a['a']['b'] and $a['a'],
// but in fact it may terminate the loop
// right after the value of $a[a][b][c] is set,
// because $a, $[a], $[a][b] and $[a][b][c]
// are all passed as reference - the reference chain is not broken.
```
- Example 2
```php
$b = array(
'a' => array(
'b' => array(
'c' => 'old-value'
)
)
)
$a = new Foo($b);
// In this example, the reference chain of $b is broken,
// because it's passed to $a.value as value
// But its elements are all passed as reference,
// so after setting the value for $b[a][b][c], there is no need
// to set value for $b[a][b] and $b[a]
$pa->setValue($a, 'value[a][b][c]', 'new-value');
```
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
