I'm trying to create a DateInputFieldModel (subclass of ComposableModel).
It uses a TextInputFieldModel as inputField.
I set inputField autoAccept: true..
I don't want to use return/enter to accept.
While typing, I'm checking with acceptBlock whether the input text is interpretable as a date.
In my case, for instance one digit is already interpretable as a date.
When you type 5, this would mean the 5th of the current month of the current year.
When the input field loses focus (tab or mouse), I'd like to render the final representation of the date into the inputField, and update the date value of my DateInputFieldModel.
Any pointers on how to achieve this?
The underlying Morph gets send #keyboardFocusChange:
You probably need to add #whenFocusChanged: somewhere in the ComposableModel hierarchy and make sure it is called from the correct AbstractMorphicAdapter subclasses
In response to Stephan Eggermont's suggestion I altered
following method
MorphicTextInputFieldAdapter>>adapt: aModel
super adapt: aModel.
aModel
whenBuiltDo: [ :w |
w widget color: Smalltalk ui theme backgroundColor.
w widget widget textMorph
onAnnouncement: MorphGotFocus, MorphLostFocus
send: #announce: to: aModel ]
This works but the drilldown w widget widget textMorph is a bit awkward.
In my DateInputFieldModel I subscribed to the announcements
inputField on: MorphLostFocus send: #lostFocus to: self.
Related
I'm adding keyboard shortcuts to a Flutter web application.
I have a form within a custom FocusableActionDetector where the shortcut has form like this:
SingleActivator(LogicalKeyboardKey.digit2)
and action is like:
CustomActivateIntent: CallbackAction<CustomActivateIntent>(
onInvoke: (intent) { provider.value = "2"; },)
In the same form I have a couple of numeric TextFormFields. To permit writing the character "2" I have to put these text fields inside some new FocusableActionDetector, otherwise the previous detector catches the command and the text field loses the "2" character, and this is already quite weird... Moreover, after writing in any of the text fields the form focus detector doesn't work anymore.
I think this could be related to the focus system, which is yet not that clear to me.
Can anyone help find a clean solution?
I found a workaround: the FocusableActionDetector is now preceded by an if statement. The code looks like the following:
// I extract the form to a widget to make it clearer
var searchWidget = SearchWidget();
child: textEditingInProgress
? searchWidget
: FocusableActionDetector(
child: searchWidget,
...,
),
The textEditingInProgress bool is a field in a provider and is controlled by the FocusNodes belonging to the TextControllers.
Still this is not a perfect solution, in particular I'd like to understand why the previous approach was not working.
In a v-data-table, I have one column with a simple text field, this column containing the text field is sortable. The problem is that when I change the value in the text field, the data is immediately re-sorted and in my case the lines change and in the worst case the line change and the focused input changes too, like in this example: codepen reproduction
For reproduction:
click on the iron header to sort the iron column
change one input
see that the line changed its position and that you're not anymore in the same field
Expected behavior:
sort the iron column
change one input
see that the line didn't change its position and you can keep modify the input
click sort again to reorder the data once you're done.
Is there a way of behaving like in the expected behavior?
Thank you very much for the answers
It looks like you should use a different event on the <v-text-field> component to listen for changes instead of using v-model to bind changes automatically to props.items.iron. Text Field Event Docs here
You could do something like using the blur event so it only updates when your user focuses away from the text field:
<v-text-field
#blur="updateIron"
:rules="[max25chars]"
label="Edit"
single-line
counter
autofocus
></v-text-field>
then in your JS file you'd add a method like this:
methods: {
// ...other stuff here
updateIron(val) {
this.item.iron = val
}
}
Try experimenting with different events to update your values at the appropriate time.
So I found the answer to that, which consist of redefining the header:
I use the slot header.iron in order to redefine a sortIron method only on click : basic demo
What I'll have to do next is to redefine arrows, style, 'asc' and 'desc'
I'm still open to other ways of doing it!
But right now I can follow my wanted behavior which is :
sort the table via the iron column
change one field
press tab, on the next field enter a value
repeat previous step until you're done
finally re-sort the table only by clicking the header
Basically I want to make the text strikethrough when hovered. This doesn't work easily with
el [ mouseOver [Font.strike] ] (text "some text")
as it would with
el [ mouseOver [Background.color someColor] ] (text "some other text")
because Font.strike is a Attribute msg whereas Background.color is of type Attr decorative msg
Does anyone konw how to achieve described behavior with something like Font.strike?
I would also accept non elm-ui solutions if not possible otherwise.
Like you already pointed out, mouseOver requires a Attr decorative msg. It uses CSS for the hover (hence the limitations), which takes care of applying the style on mouse over and clears it on mouse out.
For the general case, we have to detect mouse over/out ourselves, using Element.Events. We also need to keep track of this state in our Model. Then we can apply the Font.strike attribute conditionally depending on the model.
We can listen for these events on an Element.el
Events.onMouseEnter Enter
:: Events.onMouseLeave Leave
:: style
, where style is either [ Font.strike ] or [], depending on the model.
Full code and working demo here: https://ellie-app.com/bNjP6CbGrLJa1
I'm obviously missing something and hoping someone might be able to help.
I've an Interactive Grid, and a button.
When the button is pressed the dynamic action on the button has 2 steps.
Action 1 - Execute Javascript to take a value from one of the IG cells and put it into a page item.
Action 2 - Confirm Action - Are you sure you wish to delete &P10_JOB_ID.
I've made the page item, &P10_JOB_ID, visible and I can see the value has correctly been changed to the value from the IG.
I write P10_JOB_ID into a database table - I get the correct value
But the confirm message isn't picking up the correct value from P10_JOB_ID.
Namely it uses the value in P10_JOB_ID when the page starts, but then as I move around the IG pressing the button and changing the value of P10_JOB_ID, the text in the confirm message never changes.
Can anyone suggest what I might have missed, I'm baffled.
Thanks a lot
Substitutions like &P10_JOB_ID. are made when the page is rendered, not dynamically, so reflect the value at time of page load.
You will need to use Javascript to perform the conform action, something like:
apex.page.confirm ('Are you sure you wish to delete ' + $v('P10_JOB_ID') + '?', 'DELETE');
$v is an APEX Javascript function that returns the current value of a page item.
I used 'DELETE' as an example of a request value; you may want to do something different here.
Ok - having the setting of value and confirm as 2 separate actions is what causes the problem.
As per fac586
That is the expected behaviour. Static text substitutions are performed once during page show processing. They are not evaluated dynamically in the browser at runtime as values change.
Drop the second action and extend the first to display the confirm dialog using the apex.message.confirm JS API method, accessing the item value using the $v shorthand method.
https://issues.apache.org/jira/browse/PDFBOX-2148
When there are multiple copies with the same field name, the getFullyQualifiedName for each kid in the list of PDField objects returns the name of the parent, followed by .null. So if the parent field is called Button2 and it has 4 instances the result of printing out all the names will be:
Button2.null
Button2.null
Button2.null
Button2.null
According to the comments to the question, the OP refers to PDFBox 2.0.x versions, in particular 2.0.6.
getKids()
The method getKids() is undefined for the type PDField
In PDFBox 2.0.6 there are two immediate sub-classes of PDField. Different variants of the former (1.8.x) getKids() method are implemented in there:
PDNonTerminalField - the method retrieving the kids in this class is getChildren() and returns a List<PDField>, a list of form fields.
PDTerminalField - the method retrieving the kids in this class is getWidgets and returns a List<PDAnnotationWidget>, a list of widget annotations.
name of the parent, followed by .null
When there are multiple copies with the same field name, the getFullyQualifiedName for each kid in the list of PDField objects returns the name of the parent, followed by .null
This is not the case in PDFBox 2.0.x.
In the sample document attached to the PDFBox issue PDFBOX-2148 PDFBox now correctly finds only a single field which appropriately is named "Button2". This field is a PDTerminalField and has 4 widget annotations. The class of the latter, PDAnnotationWidget, has no getFullyQualifiedName method, so there are no ".null" names.
Thus, this problem is gone.
FQN of duplicate fields
(from the OP's comment responding to "What exactly is your question?")
how to get Fully Qualified Name of duplicate fields in pdfbox
There are no duplicate fields in (valid) PDFs, for a given name there is at most a single field which may have multiple widgets. Widgets do not have individual FQNs.
Thus, what you call "duplicate fields" in your example document actually is a single field with multiple widgets; the name of that field is "Button2" and can be retrieved using getFullyQualifiedName().
which page which form field
(from the OP's comments to this answer)
but how to get current page no in pdfbox.. for example there are 3 page and in page 2 there is a form field so how can i get which page which form field ?
All PDAnnotation classes, among them PDAnnotationWidget, have a getPage() method returning a PDPage instance.
BUT: As specified in ISO 32000-1, annotations (in particular form field widgets) are not required to have a link to the page on which they are drawn (except for screen annotations associated with rendition actions).
Thus, the above mentioned method getPage() may return null (probably more often than not).
So to determine the respective pages of your widgets, you have to approach the problem the other way around: Iterate over all pages and look for the annotation widgets in the respective annotation array.
For PDFBox 1.8.x you can find example code in this stackoverflow answer. With the information given in the previous parts of this answer it should be easy to port the code to PDFBox 2.0.x.
checkbox and radio button
(also from the OP's comments to this answer)
one more issue if i am using checkbox and radio button both then field.getFieldType() output is Btn for both. how to identify it?
You can identify them by inspecting the field flags which you retrieve via fields.getFieldFlags():
If the Pushbutton flag is set (PDButton.FLAG_PUSHBUTTON), the field is a regular push button.
Otherwise, if the Radio flag is set (FLAG_RADIO), the field is a radio button.
Otherwise, the field is a check box.
Alternatively you can check the class of the field object which for Btn may be PDPushButton, PDRadioButton, or PDCheckBox.
Beware: If a check box field has multiple widgets with differently named on states, this check box field and its widgets act like a radio button group! And not only in theory, I've seen PDFs with such check box fields in the wild.
To really be sure concerning the behavior of the fields, you therefore also should compare the names of the on states of all the widgets of a given check box.