Commits
-------
e8326aa Renamed attributes to attr to be consistent with templating.
c707467 Added support for additional attributes in Form types that list field as their parent.
Discussion
----------
[Form] Added support for additional attributes
Added support for additional attributes in Form types that list field as their parent.
This is needed particularity for html5 data and data- attributes support, unfortunately $options['data'] is already taken, so adding a general $options['attributes'] is the easiest solution without breaking BC.
Now I know that this will be tempting for some to stuck style and class attributes here also, but I'd rather not restrict the keys that are passed in.
---------------------------------------------------------------------------
by Seldaek at 2011/05/22 14:51:02 -0700
Maybe it should be called attr for consistency with the template stuff, or the other should be renamed attributes. Other than that, I'm +1, data-* attributes are awesome, and abusers will find ways to abuse things either way.
---------------------------------------------------------------------------
by mvrhov at 2011/05/22 21:48:01 -0700
Naming it attr also crossed my mind when I was signing off yesterday.
Along with the possibility to go the way xml attributes are handled when node is converted to/from array.
So every option with @ prefix would automatically become html attribute. However going the latter path, it'd be harder to implement.
---------------------------------------------------------------------------
by fabpot at 2011/06/13 07:43:52 -0700
Can you give an example of a real-world use case? Thanks.
---------------------------------------------------------------------------
by Seldaek at 2011/06/13 07:54:51 -0700
You can use `array('attr' => array('data-foo' => 'bar'))`, which will output `data-foo="bar"`, which can be read easily by jQuery for example as `$('el').data('foo')`. It's a standard compliant and elegant way to pass extra data needed by the JS code along with DOM nodes, without polluting everything with script tags and all.
---------------------------------------------------------------------------
by fabpot at 2011/06/13 08:01:08 -0700
@Seldaek: I understand that. But why not doing this in the template?
---------------------------------------------------------------------------
by Seldaek at 2011/06/13 08:04:10 -0700
Well, I agree it most likely belongs in the template, but it's kind of data stuff that is not directly impacting the display rules of the element, so in some cases having the possibility to set that from the php code might be useful. Anyway I'll let @mvrhov answer maybe he had a more concrete use case. I just think it's nice to leave the door open, but I don't really need it.
---------------------------------------------------------------------------
by mvrhov at 2011/06/13 10:27:03 -0700
A bit late to the party. Ok, here is my use-case.
I have a pretty large form where part of the data is of tabular form. The number of rows is almost every time a lot larger than the number of columns
So I can either output a lot of text inputs filled with data and make already a large form intimidating. Or I can use a grid that supports editing. So I serialize that tabular data as json and put it as a value into one hidden field. Somehow I also have to get the column definitions to that grid. I decided to I serialize it and put it inside data-* attribute. Putting it into another hidden field doesn't make sense as when data is submitted back I don't need the column definitions as only the number of rows changes.
---------------------------------------------------------------------------
by fabpot at 2011/06/13 10:44:58 -0700
@mvrhov: ok, but what prevents you from doing this in the template directly?
---------------------------------------------------------------------------
by mvrhov at 2011/06/13 11:22:53 -0700
I have to get it into the view somehow. What I'd really not like to do is, iterate through that data in a controller and then pass it as another template variable.
---------------------------------------------------------------------------
by fabpot at 2011/06/14 00:10:22 -0700
But the controller is where you prepare the data that you want to send to the view. Without any concrete example, I'm going to close this PR.
---------------------------------------------------------------------------
by mvrhov at 2011/06/14 01:21:10 -0700
IMHO this has to go out through form as this is the part of the form also I do have a very slim controllers in this case 10 LOC, where half of them is just setting up an view data..
Nonetheless I went looking again very closely at the AbstractType and I do have buildView function available which I can override and set the column data I need there, and then provide custom view for that, so at least from this part this is an non issue.
With this PR all default form attributes can be set from outside and when searching for a good use-case I have found out that @henrikbjorn has implemented this via extensions [1], maybe he has a good use-case for this.
[1] - https://github.com/Comways/ComwaysFormExtraBundle/blob/master/Form/Extension/FieldTypeExtension.php
---------------------------------------------------------------------------
by henrikbjorn at 2011/06/14 01:48:53 -0700
Convenience is the only use case
---------------------------------------------------------------------------
by stof at 2011/06/14 02:08:09 -0700
@fabpot The issue is that passing it from the controller as another template variable makes it really hard when you use the type twice with different values. Passing them from the form would be the easiest way
---------------------------------------------------------------------------
by shuogawa at 2011/06/14 19:37:50 -0700
hello. thanks for great form library.
I want to support two ways to display additional attributes for form elements.
1) control in template like
{{ form_row(form.name, { 'width': '30' }) }}
2) control from php
$builder->add('name', 'text', array('attr'=>array('width'=>'30')));
If form elements configure by end user like cms,
and form elements dynamically change.
The second method is useful.
template designer can write {{form_row(form)}}
but
template designer can not write {{form_row(form.name), { 'width': '30' }}}
Thank you
---------------------------------------------------------------------------
by fabpot at 2011/06/14 23:01:18 -0700
@shuogawa: That's what I fear. Setting the `width` or any other attribute in PHP is wrong. This belongs to the templates.
---------------------------------------------------------------------------
by stloyd at 2011/06/15 00:09:05 -0700
@fabpot Then maybe just restrict allowed tags to `data-*` and don't use `attr` but some other not confusing name ? Just like we do with i.e. `maxlength`.
---------------------------------------------------------------------------
by fabpot at 2011/06/15 04:44:25 -0700
I'm going to merge it as a "convenience" tool, but the documentation should clearly state that the only usage should be for `data-*` attributes.
Commits
-------
07fa82d [Form] Revert changes impacting perfomance and readability
b709551 [Order] Make Form::types and FormView::types use the same order (Parent > Child)
e56dad6 [Form] simplify the code
bdd755e [Form] Fix the exception message when no block is found
c68c511 [Form] Make theming form prototypes consistent (start by looking for a '_<id>_<section>' block)
9ec9960 [Form] Simplify the code
4e3e276 [Form] Make the prototype view child of the collection view
Discussion
----------
[Form] Make the prototype view child of the collection view
This PR should be a base for discussion.
The [current implementation](https://github.com/symfony/symfony/pull/1188) has some drawbacks because the prototype view is not a child of the collection view:
* The 'multipart' attribute is not propagated from the prototype to the collection,
* The prototype view do not use the theme from the collection.
Those 2 points are fixed by the proposed implementation and one more benefit is that the template markup might be easier to work with:
before:
```html
<div id="form_emails">
<div>
<label for="form_emails_0">0</label>
<input type="email" id="form_emails_0" name="form[emails][0]" value="a@b.com">
</div>
<script type="text/html" id="form_emails_prototype">
<div>
<label for="$$name$$">$$name$$</label>
<input type="email" id="$$name$$" name="$$name$$" value="" />
</div>
</script>
</div>
```
after:
```html
<div id="form_emails">
<div>
<label for="form_emails_0">0</label>
<input type="email" id="form_emails_0" name="form[emails][0]" value="a@b.com">
</div>
<script type="text/html" id="form_emails_prototype">
<div>
<label for="form_emails_$$name$$">$$name$$</label>
<input type="email" id="form_emails_$$name$$" name="form[emails][$$name$$]" value="" />
</div>
</script>
</div>
```
@kriswallsmith I'd like to get your feedback on this PR. thanks.
---------------------------------------------------------------------------
by stof at 2011/06/14 07:01:01 -0700
@fabpot any ETA about merging it ? Using the prototype currently is a pain to build the name. The change makes it far easier
---------------------------------------------------------------------------
by fabpot at 2011/06/14 07:09:46 -0700
The templates are much better but I'm a bit concerned that we need to add the logic into the Form class directly. That looks quite ugly. If there is no other way, I will merge it.
---------------------------------------------------------------------------
by vicb at 2011/06/14 07:14:32 -0700
I have found no better way... I am testing some minor tweaks I want to submit.
---------------------------------------------------------------------------
by kriswallsmith at 2011/06/14 07:34:25 -0700
I'm not happy with the code in Form.php either... would creating a PrototypeType accomplish the same thing?
---------------------------------------------------------------------------
by vicb at 2011/06/14 07:42:07 -0700
@kriswallsmith tried and dismissed, the id and name are bad & you have to go for `render_widget(form.get('proto'))` in the template. That should be fixeable but not any better.
---------------------------------------------------------------------------
by kriswallsmith at 2011/06/14 07:45:21 -0700
What do you mean the id and name are bad? If we have a distinct type for the prototype, can't we do whatever we want using buildView() and the template?
---------------------------------------------------------------------------
by vicb at 2011/06/14 07:53:31 -0700
@kriswallsmith the id would be smthg like `form_emails_$$name$$_prototype` but yes we should be able to do whatever we want but the code might end up being more complex.
I am done with the tweaks but still open to feedback on this PR.
---------------------------------------------------------------------------
by kriswallsmith at 2011/06/14 08:08:21 -0700
Yes, that is the type of name I would expect.
---------------------------------------------------------------------------
by kriswallsmith at 2011/06/14 08:08:33 -0700
Oops -- I mean id.
---------------------------------------------------------------------------
by kriswallsmith at 2011/06/14 08:09:42 -0700
Maybe I'm confused what id you're referring to. I'll try to spend some time on this today.
---------------------------------------------------------------------------
by vicb at 2011/06/14 08:23:56 -0700
That should be the id of the `<input>`, the id of the script would be `form_emails_$$name$$_prototype_prototype` (if prototype is the name of the nested node).
I am trying to setup a branch with my code (playing with git & netbeans local history)
---------------------------------------------------------------------------
by vicb at 2011/06/14 08:46:25 -0700
@kriswallsmith https://github.com/vicb/symfony/tree/kris/proto if that can help (there are still changes in Form.php)
---------------------------------------------------------------------------
by kriswallsmith at 2011/06/14 08:47:08 -0700
Thanks, I'll take a look.
---------------------------------------------------------------------------
by vicb at 2011/06/15 00:48:38 -0700
I would have expected it to be faster however `array_map` is about twice slower... reverted !
Rules are as follows:
* If multiple is true, then the empty_value is ignored
* If not, and if the field is not required, the empty_value is set to the empty string by default and displayed
* If the field is required, and if the user explicitely set the empty_value, then it is displayed
* kriswallsmith/form/collection-proto:
added script[type="text/html"] collection prototype to form themes
[Form] removed collection prototype from form tree
This reverts commit d044498cde.
The reason for reverting is that the name is actually used to customize
the template on a per field basis:
{% block _post_excerpt_widget %}
***{{ block('text_widget') }}***
{% endblock %}
Here, post is the name of the Type.
* kriswallsmith/form/lazier-csrf-token:
[Form] fixed xpath
[Form] moved csrf listener to its own class
fix issue with csrf token not present on collection fields because of resize listener
The current implementation is not ready for inclusion in 2.0. It has several
known problems (security, not possible to disable it, not "cloud-compatible",
...) and it's not a must have feature anyway.
Some references:
* Security issue in FileType: https://github.com/symfony/symfony/issues/1001
* Validation fails on file, still stored in TemporaryStorage: https://github.com/symfony/symfony/issues/908
* Add a size argument & ability to configure TemporaryStorage: https://github.com/symfony/symfony/pull/748
This feature should be reworked and discussed for inclusion in 2.1.
This fixes choices when all the keys are strings (but integers really).
That's because in PHP, if you have the following array:
array('1' => 'foo', '2' => 'bar');
PHP "converts" it automatically to the following array:
array(1 => 'foo', 2 => 'bar');
* fivestar/single-choice-expanded:
[Form] Fixed FixRadioInputListener to not ignore 0.
[Form] Fixed single expanded choice type to set checked attribute when passed boolean value
* jwage/master:
Revert back to using gmmktime and use day 15 instead of 1 to avoid edge case
Use mktime instead of gmmaketime. On the 1st of every month the following options are generated:
* Also changed the default value of the calendar paramter to \IntlDateFormatter:GREGORIAN
in DateTimeToLocalizedStringTransformer which is the same as the default value in
StubIntlDateFormatter
<select id="filter_start_date_month" name="filter[start][date][month]">
<option value="1">Dec</option>
<option value="2">Jan</option>
<option value="3">Feb</option>
<option value="4">Mar</option>
<option value="5">Apr</option>
<option value="6" selected="selected">May</option>
<option value="7">Jun</option>
<option value="8">Jul</option>
<option value="9">Aug</option>
<option value="10">Sep</option>
<option value="11">Oct</option>
<option value="12">Nov</option>
</select>
I do not totally understand the problem but using mktime instead of gmmktime fixes the issue.
* vicb/form-rendering:
[Form] The variable stack should not persist between section rendering (fixes#1157)
[Twig][Form] Tweak form extension phpDoc and code
[Form] Tweak phpDoc
[FormView] fix phpDoc
[Form] Some tweaks
* Seldaek/events:
[EventDispatcher] Removed temporary code
[FrameworkBundle] Improved code readability
[FrameworkBundle] Clarified code and fixed regression
Update Core and Security events to latest model
[EventDispatcher] Allow registration of arbitrary callbacks
[EventDispatcher] Remove useless code
[EventDispatcher] Minor memory optimization to getListeners()
[FrameworkBundle] Small optimization, remove some function calls
* CodeMeme/1058-fix-radio-input-listener:
Fix for RadioInputListener's empty value erroneously becoming extra data Refs #1058
Added test for RadioInputListener bug treating no data as extra data
* maoueh/date_type_dead_code:
[Form] Removed dead code in the buildForm method of DateType.php * With the introduction of the getAllowedOptionValues mechanics, the check is performed prior to the buildForm call. There is no more needs to check it again in DateType.
This in effect removes the direct link between event name and the method name on the handler.
Any callback can be given as a handler and the event name becomes an arbitrary string. Allowing for easier namespacing (see next commit)
* With the introduction of the getAllowedOptionValues mechanics, the check
is performed prior to the buildForm call. There is no more needs to check
it again in DateType.
* vicb/form-fixes:
[Form] Make the PropertyPathMapper class use the UnexpectedTypeException
[Form] Fix adding transformers in the FormBuilder
[Form] Fix the ReversedTransform class
If you use the MinLength validator with your entities, the ValidatorTypeGuesser gets the value, stored as "minlength". Then, the FormFactory generates a "pattern" attribute out of minlength and maxlength.
Modern browsers such as Chrome use this attribute to validate the form before submitting.
a "pattern" attribute is generated that validates the
The form component should now guarantee to always pass an UploadedFile object to your model. There you can call getOriginalName() to retrieve the original name of the uploaded file. For security reasons, the real file name is a generated hash value.
* bschussek/form:
[Form] Automatically setting "data_class" option if objects are passed at the creation of a form
[Form] Improved the way passed data is handled in FormFactory
[Form] Simplified FileType code
[HttpFoundation] TemporaryStorage automatically creates the directory if it doesn't exist yet
[Form] Changed FormBuilder::build() to FormBuilder::create(). You hvae to pass the resulting builder to FormBuilder::add() manually now
[Form] Added FieldTypeValidatorExtension and fixed FQCN of DelegatingValidator
With implementations of this interface, existing types can be amended.
The Csrf extension, for example, now contains a class FormTypeCsrfExtension
that adds CSRF capabilities to the "form" type.
To register new type extensions in the DIC, tag them with "form.type_extension"
and the name of the extended type as alias.
The extension classes are now the only constructor argument of the FormFactory class. They replace the existing "type loader" classes.
new FormFactory(array(
new CoreExtension($validator, $storage),
new CsrfExtension($csrfProvider),
new DoctrineOrmExtension($em),
));
Together with a few upcoming commits this mechanism will make
* extension of the form framework in bundles and
* usage of the forms outside of Symfony2
much easier.
The data can now be passed to all creation methods:
$form = $factory->create('form', $data);
By default, a form will receive the name of its type ("form" in above example). If you wish to pass a custom name, use createNamed():
$form = $factory->createNamed('form', 'myform', $data);
[Form] Fixed {get,set,has}Var references in templating php
[Form] Added getVars to FormView to ease usage in Twig. Also added some phpdoc and cleaned up the get method by adding a default value
[Form] Fix
[Form] Delete file generated by test
Based on dfb93b1bcebf1f34d3a880d00f36acb2bcca0f08:
[FORM] Fixed DateField Month Choices
The month choices were calculated using the current day of the month with
gmmktime rather than the 1st of the month. Additionally, this provides a
UTC timestamp which is passed to the formatter (IntlDateFormatter) which
converts the timestamp using the current timezone. This means that the UTC
timestamp for 1st March was being converted for my timezone (EST) and giving
a date of 28th February, leading to Feb appearing again in the popup form
instead of Mar.
Moving the template context creation makes sense and allows for simpler code for the end user:
Before:
return array('post' => $post, 'form' => $this->get('form.factory')->createTemplateContext($form));
After:
return array('post' => $post, 'form' => $form->getContext());
The problem was that "data_class" was used in two places: FormBuilder::build() and PropertyPathMapper.
PropertyPathMapper was already constructed during FormType::buildForm(), so any data class changes made to the FormBuilder wouldn't affect the data class of the PropertyPathMapper anymore and so lead to an inconsistent state.
You can now modify set or bound data by adding a filter for either of the following events:
* Filters::filterBoundDataFromClient
* Filters::filterBoundData
* Filters::filterSetData
Instead, hidden fields now override the "row" template to not include a label or errors.
The "rest" (former "hidden") helper has been adapted to output any fields that were not
rendered manually. It should usually be called at the end of a form.
Fields are not available in the templates anymore. Instead, all required information can be
accessed through view variables.
Example usage of helpers and variables in a form theme:
// use the label helper
{{ this.label('my label') }}
// use the label variable
{{ this.vars.label }}
{{ label }}
Example usage of helpers and variables in a normal template:
// use the label helper
{{ field.label('my label') }}
// use the label variable
{{ field.vars.label }}
This commit removes CollectionToStringTransformer. Transformers should never change the state of the outside world, otherwise hard-to-track bugs might creap in.
This functionality needs to be implemented as a custom FieldType (see EntityChoiceField).
The implication is that set<Reference>() in the object of the parent form will not be called (and thus not has to be implemented/public).
If you want to suppress this behaviour, manually set "by_reference" to false.
Separated validation of data and form had serious drawbacks. When a form had nested form whose data was not connected to the data of the root form, this data would not be validated.
The new implementation validates the whole object graph at once. Class Form has a new method validateData(), that manually passes the data to the GraphWalker of the Validator and overrides the Default group with the groups set in the form.
With the form factory there was no reasonable way to implement instantiation of custom form classes. So the implementation was changed to let the classes instantiate themselves. A FormContext instance with default settings has to be passed to the creation method. This context is by default configured in the DI container.
$context = $this->get('form.context');
// or
$context = FormContext::buildDefault();
$form = MyFormClass::create($context, 'author');
If you want to circumvent this process, you can also create a form manually. Remember that the services stored in the default context won't be available then unless you pass them explicitely.
$form = new MyFormClass('author');
A form now always has to be bound, independent of whether the request is a POST request or not. The bind() method detects itself whether the request was a post request or not and reads its data accordingly. The "old" bind()/isBound() methods were renamed to submit()/isSubmitted().
$form = new Form('author');
$form->bind($request, $author);
if ($form->isValid()) {
// isValid() implies isSubmitted(), non-submitted forms can
// never be valid
// do something with author now
}
Alternatively, you can only bind global variables, if you don't have a request object.
$form->bindGlobals($author);
Note that the $author object is in both cases optional. You can also pass no object at all and read the data using $form->getData(), but then no validation will occur. You can also prefill the form with an object during instantiation.
$form = new Form('author', array('data' => $author));
$form->bind($request);
// etc.
* Added empty_value option on CountryField, LanguageField, LocaleField, TimezoneField
* Added missing date_pattern to DateTimeField
* Made the currency option on MoneyField required.
I believe the empty_value option is just a leftover Django-style option for what value the field should have if left blank. I don't see this doing anything and find no reference to anything like it anywhere else.
The separator option functionality is currently handled as a parameter in the field render functions. I think this should be moved to an option on the field, but this reflects teh current functionality (i.e. this option is not used).
If the option is true, the password is never written into the input field's value. If it is false, it is only written into the input field's value after submitting a form with errors.
The default value for "always_empty" is true.
The semantics of property paths are now:
(1) if a property path is set, it is _always_ respected (relative to the object
of the parent field)
(2) if no property path is set, the object of the parent field is _always_ ignored
Fact (2) allows us to set data into fields that is updated independently of the parent
field (like CSRF tokens, subforms with different objects etc.)
What is missing now is support for subfields that pass the object of the parent field
through to their own subfields. This functionality would be needed for GoogleMapFields,
DateRangeFields etc., which are compositions of individual fields that update the
parent object of the FieldGroup.
There are several alternatives for the latter functionality that should be discussed
in a RFC.
The constraint "Valid" does not accept any options or groups anymore. As per
JSR303 1.0 final, section 3.5.1 "Object graph validation" (page 39),
properties annotated with valid should be cascaded independent of the current
group (i.e. always). Thus the group "*" is not necessary anymore and was
removed from the "Valid" constraint in the Form validation.xml.
Support for theming in PHP templates has been dropped.
True theming should support theme inheritance, e.g. mytheme <- table <- default.
Currently, the Templating component does not support such inheritance. As the
only purpose of the themes so far was to style field groups with tables or
divs, and because automatic rendering of field groups/forms through the render()
method is discouraged and only recommended for rapid prototyping, themes are
dropped for now.
Added ability to specify **match-all** validation group, which
constraints will runs on every specified validation group.
Added groups="*" option to `Form::data` Valid validator.
Fields can now easier support different data types in their underlying object.
These datatypes can be normalized to a single datatype using a normalization
transformer. The normalized value can then be transformed to the user's
representation with the value transformer (better name required?).
When reading the last bit of a property path mapped to a missing array index, the method would initialize the value to an empty array. This makes sense for cases where readPropertyPath would again be called recursively, but not when the value would be immediately returned (null would be preferable in that case).
For example, we have an object with a property called "options" that's an array of arbitrary key/value pairs. That "options" property (and getOptions()) maps directly to a FieldGroup within the Form for this object. That FieldGroup contains multiple TextFields for a few expected keys in the array. As-is, if those keys were not defined, the default data set for those TextFields could end up being "Array" (string representation of an empty array). If readPropertyPath instead returns null for this case, the default data would be transformed into an empty string.
It's better to be able to fetch all the visible and all the hidden fields separately for display purposes (hidden fields in <ul> tags without an <li> do not validate)
ReflectionClass doesn't list properties on stdClass objects (or objects cast
from arrays). This allows these annoymous objects to be used as field data.
Internally, ChoiceField expects both choices and preferred_choices to be a simple array, so I replaced incomplete bits of code that attempted to not modify a possible ArrayObject and instead added type checks in the configure() method (with unit tests for expected exceptions).
Property paths such as fields[group].fields[innerGroup].data were not being resolved correctly, since the second iteration of addError() (based on "group") would attempt to call get('fields') instead of get('innerGroup'). Solution is to remember to bump the propertyPath forward if we're at the fields property