This PR was submitted for the 2.8 branch but it was merged into the 3.2-dev branch instead (closes#15458).
Discussion
----------
[Filesystem] Add feature to create hardlinks for files
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Todo:
- [x] Tests
- [ ] Doc
- [x] Getting someone to test on Windows as exception might differ from symlink functionality
## Why
Symlinks are good for directories, but for files hardlinks are sometime a better match when needing to represent same file in several locations.
One use case for this feature is for multi tagging of Http Cache when running against FileSystem. Multi tagging is used in FoSHttpCache and supported when running Varnish, but not with Symfony HTTPCache, yet.. With hardlinks combined with meta information containing tags within the file, lookup and reverse lookup is thus easily possible.
## What
Introduces method `hardlink()` with similar behavior as `symlink`, difference being:
- Allowing several targets to be provided to match use case in 'why'
- Like `symlink` removes existing target if not the same, but uses `fileinode` to compare target & source
Commits
-------
8475002 [Filesystem] Add feature to create hardlinks for files
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Console] Add path argument to dump a specific option in debug:config
| 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 | n/a
This adds the ability to dump a specific bundle config option from the `debug:config` command.
For instance `debug:config StofDoctrineExtensionsBundle uploadable` gives:
I hesitated to just look for a `.` in the `name` argument rather than adding a `path` argument that doesn't include the bundle alias (that is took from the first argument of the command), let me know what you think about that.
Commits
-------
05ae01b [Console] Add path argument to dump a specific option in debug:config
This PR was merged into the 3.2-dev branch.
Discussion
----------
[ExpressionLanguage] Added a way to dump the AST
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Because it's convenient and it makes easier to inline results of sub/smaller part of the main expression.
Commits
-------
87af6e5 [ExpressionLanguage] Added a way to dump AST
This PR was merged into the 3.2-dev branch.
Discussion
----------
[Routing] Throw exception when PHP start tag is missing
| Q | A
| ------------- | ---
| Branch? | master (?)
| Bug fix? | yes
| New feature? | no
| BC breaks? | yes
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | https://github.com/symfony/symfony-docs/issues/6116 (and many more)
| License | MIT
| Doc PR | -
The Problem
---
In the documentation, we never use the PHP start tag. However, in the first tutorials, people simply copy/past the code examples, save the file and expect things to work. They seem to often forget to add the PHP start tag.
Without start tag, the annotation file loader simply skips the file without providing any reason why. As a big framework is quite overwhelming, simple things like this are completely forgotten.
The Fix
---
If a `*.php` file only consists of `T_INLINE_HTML`, it means there is no `<?php` start tag. In this case, instead of skipping the file, an exception is throwed with a possible fix.
BC Break?
---
As the file loader is only executed for `*.php` files, I think the BC break is minimal, but it is possible that people have applications with `*.php` files *without* a start tag. If this file lives in a routing loaded directory (e.g. when doing `@AppBundle/Controller`), it would now result in an exception.
I think this BC break is minimal and can be ignored. If you don't agree, we can add a little str match to check if `class ... {` exists. If that's the case, I think it's safe to say that it was meant to be a PHP file.
Bug or Feature?
---
I don't know if this is considered a bug or a feature, so I submitted it as a feature.
Commits
-------
1e765c8 Throw exception when PHP start tag is missing
This PR was squashed before being merged into the 3.2-dev branch (closes#18781).
Discussion
----------
[Console] Display errors in quiet mode
| Q | A
| ------------- | ---
| Branch? | 2.8
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #18767
| License | MIT
| Doc PR |
- map VERBOSITY_QUIET normally, rather than suppressing all output
without override
- ensure that we do write to the output if we've determined (via
verbosityLevelMap) that we should
About the second point, it seems ConsoleLogger has the same issue (https://github.com/symfony/symfony/blob/2.8/src/Symfony/Component/Console/Logger/ConsoleLogger.php#L92), but since we're not using that, I haven't checked it.
Commits
-------
278c26f [Console] Display errors in quiet mode
This PR was merged into the 3.2-dev branch.
Discussion
----------
[HttpKernel] Add convenient method ArgumentResolver:: getDefaultArgumentValueResolvers
| 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 | N/A
I realized while implementing my own `ArgumentValueResolver` that registering a new one is not easy, you have to give the whole list of resolver instead of pushing your one in the stack.
Commits
-------
6fdfc92 [HttpKernel] Add convenient method ArgumentResolver::addArgumentValueResolver
This PR was squashed before being merged into the 3.2-dev branch (closes#18977).
Discussion
----------
[PropertyAccess] Add missing arguments to PropertyAccess::createPropertyAccessor()
| 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/6640
Actually, the recommended way to use the PropertyAccessor is to use `PropertyAccess::createPropertyAccessor`.
The problem is that using this way, we can't specify that invalid indexes should throw an exception, and so when calling `PropertyAccessor::isReadable([], '[foo]')` it returns always true.
It should be possible to make the exception thrown, plus the `PropertyAccessor::$throwExceptionOnInvalidIndex` already exists but is not used in `PropertyAccess::createPropertyAccessor`.
Commits
-------
5ded804 [PropertyAccess] Add missing arguments to PropertyAccess::createPropertyAccessor()
This PR was merged into the 3.2-dev branch.
Discussion
----------
[3.2][WebProfilerBundle] Fix bundle usage in Content-Security-Policy context without unsafe-inline
| Q | A
| ------------- | ---
| Branch? | 3.2
| Bug fix? | yes
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15397
| License | MIT
| Doc PR | N/A
Hello, this PR fixes the compatibility of the WebprofilerBundle in a context where Content-Security-Policy headers are could prevent `unsafe-inline` of `script-src` or `style-src` directives.
This PR has been originally proposed in 2.8 in #18434
Commits
-------
571a1f2 [WebProfilerBundle] Fix bundle usage in Content-Security-Policy context without unsafe-inline
This PR was squashed before being merged into the 3.2-dev branch (closes#18790).
Discussion
----------
[Console] Show aliases in command description instead of in different lines in application description
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #18351
| License | MIT
| Doc PR |
This PR is a suggestion about the bug referenced in ticket 18351. The problem is that aliases create new lines with identical descriptions in the application description, it is suggested to show the aliases in the command description.
I had to modify one application fixture to meet this new behaviour and make the tests pass, but I don't think that a new application fixture is needed to test this having the application fixture 2.
Comments are welcome, this is my first PR in Symfony so if I need to meet any more requirements to make it correct to merge just tell me please.
Commits
-------
548ab0a [Console] Show aliases in command description instead of in different lines in application description
This PR was squashed before being merged into the 3.2-dev branch (closes#18992).
Discussion
----------
[Cache] AbstractAdapter: avoid an extra call to ApcuAdapter::isSupported
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Commits
-------
8f2ea52 [Cache] AbstractAdapter: avoid an extra call to ApcuAdapter::isSupported
This PR was squashed before being merged into the 3.2-dev branch (closes#18337).
Discussion
----------
[PropertyInfo] Support singular adder and remover
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | yes
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #18166
| License | MIT
| Doc PR | n/a
Fix#18166 with only a soft dependency to the PropertyAccess component. If #18260 is accepted, this PR can be reworked to use this new component, in the meantime it fixes the problem with a soft dependency.
I don't now if it should be considered as a bug fix (and then merged in 2.8) or as a new feature. It's technically a bug fix but I'm not sure that introducing a new soft dependency is acceptable if older branches. What do you think @symfony/deciders?
Commits
-------
4cbb60c [PropertyInfo] Support singular adder and remover
* 3.1:
[BrowserKit] Bump dom-crawler minimum version requirement
Make one call to "OutputInterface::write" method per table row
[HttpKernel] Fix context dependent test
[Debug] Fix context dependent test
* 3.0:
[BrowserKit] Bump dom-crawler minimum version requirement
Make one call to "OutputInterface::write" method per table row
[HttpKernel] Fix context dependent test
[Debug] Fix context dependent test
* 2.8:
[BrowserKit] Bump dom-crawler minimum version requirement
Make one call to "OutputInterface::write" method per table row
[HttpKernel] Fix context dependent test
[Debug] Fix context dependent test
* 2.7:
[BrowserKit] Bump dom-crawler minimum version requirement
Make one call to "OutputInterface::write" method per table row
[HttpKernel] Fix context dependent test
[Debug] Fix context dependent test
This PR was submitted for the 2.3 branch but it was merged into the 2.7 branch instead (closes#18980).
Discussion
----------
[BrowserKit] Bump the dom-crawler minimum version requirement
| Q | A
| ------------- | ---
| Branch? | 2.3
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #18826
| License | MIT
| Doc PR | -
The BrowserKit Client will not work with DomCrawler 2.0 if the CssSelector component is not installed.
This will require an amendment when merging into the 2.8 branch. It currently uses `~2.0,>=2.0.5|~3.0.0`, while it should be `~2.1|~3.0.0`.
Commits
-------
5579d86 [BrowserKit] Bump dom-crawler minimum version requirement
This PR was submitted for the master branch but it was merged into the 2.7 branch instead (closes#18966).
Discussion
----------
[console] Improve table rendering performance
| Q | A
| ------------- | ---
| Branch? | 2.8
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #18957
| License | MIT
| Doc PR | -
With these improvements on my application, that is rendering table with 128 rows and 5 columns I've cut off total application runtime from **0.39s** to **0.26s**.
Commits
-------
73b812e Make one call to "OutputInterface::write" method per table row
This PR was merged into the 2.7 branch.
Discussion
----------
[HttpKernel] Fix context dependent test
| Q | A
| ------------- | ---
| Branch? | 2.7
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Class `Foo` exists when the ClassLoader component is tested alongside with HttpKernel
Commits
-------
065dee8 [HttpKernel] Fix context dependent test
* 3.1:
`@throws` annotations should go after `@return`
Fix merge
updated VERSION for 2.3.42
update CONTRIBUTORS for 2.3.42
updated CHANGELOG for 2.3.42
Revert "bug #18908 [DependencyInjection] force enabling the external XML entity loaders (xabbuh)"
Partial revert of previous PR
[DependencyInjection] Skip deep reference check for 'service_container'
Catch \Throwable
[Serializer] Add missing @throws annotations
Fix for #18843
force enabling the external XML entity loaders
Removed UTC specification with timestamp
* 3.0:
`@throws` annotations should go after `@return`
Fix merge
updated VERSION for 2.3.42
update CONTRIBUTORS for 2.3.42
updated CHANGELOG for 2.3.42
Revert "bug #18908 [DependencyInjection] force enabling the external XML entity loaders (xabbuh)"
Partial revert of previous PR
[DependencyInjection] Skip deep reference check for 'service_container'
Catch \Throwable
[Serializer] Add missing @throws annotations
Fix for #18843
force enabling the external XML entity loaders
Removed UTC specification with timestamp
Conflicts:
src/Symfony/Component/Yaml/Tests/InlineTest.php
* 2.8:
`@throws` annotations should go after `@return`
Fix merge
updated VERSION for 2.3.42
update CONTRIBUTORS for 2.3.42
updated CHANGELOG for 2.3.42
Revert "bug #18908 [DependencyInjection] force enabling the external XML entity loaders (xabbuh)"
Partial revert of previous PR
[DependencyInjection] Skip deep reference check for 'service_container'
Catch \Throwable
[Serializer] Add missing @throws annotations
Fix for #18843
force enabling the external XML entity loaders
Removed UTC specification with timestamp
Conflicts:
CHANGELOG-2.3.md
src/Symfony/Bundle/FrameworkBundle/Routing/DelegatingLoader.php
src/Symfony/Bundle/TwigBundle/Extension/AssetsExtension.php
src/Symfony/Component/Config/Loader/FileLoader.php
src/Symfony/Component/DependencyInjection/Container.php
src/Symfony/Component/DependencyInjection/ContainerBuilder.php
src/Symfony/Component/Finder/Expression/Expression.php
src/Symfony/Component/Finder/Finder.php
src/Symfony/Component/HttpKernel/DependencyInjection/ContainerAwareHttpKernel.php
src/Symfony/Component/HttpKernel/HttpCache/HttpCache.php
* 2.7:
`@throws` annotations should go after `@return`
Fix merge
updated VERSION for 2.3.42
update CONTRIBUTORS for 2.3.42
updated CHANGELOG for 2.3.42
Revert "bug #18908 [DependencyInjection] force enabling the external XML entity loaders (xabbuh)"
Partial revert of previous PR
[DependencyInjection] Skip deep reference check for 'service_container'
Catch \Throwable
[Serializer] Add missing @throws annotations
Fix for #18843
force enabling the external XML entity loaders
Removed UTC specification with timestamp
Conflicts:
src/Symfony/Component/DependencyInjection/Tests/Dumper/PhpDumperTest.php
src/Symfony/Component/Finder/Finder.php
src/Symfony/Component/Security/Acl/Dbal/MutableAclProvider.php
src/Symfony/Component/Security/Acl/Domain/ObjectIdentity.php
src/Symfony/Component/Security/Acl/Model/AclInterface.php
src/Symfony/Component/Security/Acl/Model/MutableAclProviderInterface.php
src/Symfony/Component/Security/Acl/Permission/MaskBuilder.php
src/Symfony/Component/Translation/Loader/XliffFileLoader.php
src/Symfony/Component/Yaml/Tests/InlineTest.php
This PR was merged into the 2.7 branch.
Discussion
----------
`@throws` annotations should go after `@return`
| Q | A
| ------------- | ---
| Branch? | 2.7
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Commits
-------
bf3a2c0 `@throws` annotations should go after `@return`
