I have an ExtJs grid that has a "childtap" event attached to it. (cmd 7.3 modern)
When I click an entry, it shows the location I clicked, using
onGridChildtap: function(list, location, eOpts) {
console.log(location);
}
and this is the contents of the location attribute:
However, let's say I try to hide the panel that contains the grid, then show it back again.
Then, without changing anything on the grid, the "location" content becomes the following:
Note that many object keys are missing, like column for example. This seems to happen at random, and I can not replicate it doing something specific. Sometimes it also get's back to normal by itself.
Is this a bug? Has someone ever come accross this unexpected behaviour?
Related
I'm using ExtJS 5.1.3, I have a grid which is loaded from a store which has a model. The grid is set to use plugin roweditor, so I edit a cell and give it a new value, at this point the red tick is shown that the cell has been changed.
I have a Save button which when clicked gets the store.getModifiedRecords() and passes these off to a ajax request, upon success of this request a few things happen and the last action I do is to load the grid store again which then populates the grid again with the latest version of the data, this is fine and seems to be working as expected.
As this is a multi page application I also have a check when a user navigates away from this page, this is to catch any unsaved grid changes, so basically I get any form from the page and verify the isDirty() value, this is where I am finding my issue, the roweditor is being returned as dirty, this is because some columns have an editor and ExtJS uses form validation on these fields,
I can't understand why the store loading again has not cleared any dirty fields associated with the grid columns? I've tried a number of things such as clearing the store prior to ajax request along with refreshing the grid view, I've tried committing the store changes prior to doing the ajax request but each time I try navigate away from the page after a grid save I pick up the roweditor as having dirty fields :( any help is greatly appreciated.
EDIT: managed to replicate on a simple fiddle
https://fiddle.sencha.com/#view/editor&fiddle/1rmf
The fiddle is basic, to replicate follow these steps;
edit first row age, change age to 13
click Save (i'm forcing the store to load data which has the change we've made)
click 'Check roweditor is Dirty() value' button to see the value of the roweditor isDirty() function, this will return true
if you look at the button handler, you can drill into forms[0].items.items[2] and see that this field has dirty: true which is why isDirty() is returning true.
SOLUTION
As explained in accepted answer, the roweditor is not affected by the store edit/cancel or load in my case. What I did when clicking on 'Save' was to get the grid, then the editor and it's form and called reset() on this so effectively sync everything again.
grid.editingPlugin.getEditor().form.reset();
you can also get access to plugins via grid.getPlugins() which returns an arrary
updated fiddle to show it working
https://fiddle.sencha.com/#view/editor&fiddle/1rmr
During the editing process grid will eventually call loadRecord on the editor's form. However the editor's form is not cleared upon editing success or canceling. That is why your check for dirtyness returns false.
Grid reloading the data is not destroying the editors. It is an optimisation. Editors are created only once and they are destroyed along with the grid.
I'll try to answer regarding to an experience of mine with an all ExtJS desktop application.
By the way looking quickly over your description, you may have to call the Store.sync() method that refreshes your store.
Looking more deeply, there are many way to make CRUD using ExtJs.
I've been made using the "instance" of store but at certain point I had to change it to static calls like MyApp.store.Model.save() etc. That makes you have only one instance of the store avoiding dirty data.
Here's my project folder if you need
https://github.com/guilhermeribeirodev/grizzlyboilerplate/tree/master/src/main/webapp/js/MyApp
I'm using a property grid with editors set in sourceConfig. I want only some rows to be editable and not all. Returning false in beforeedit disables all. The reason is, I have a button in a grid. When I click the button, the grid turns to a textfield! Any way to do this?
The beforeedit event should provide you with the editor and the editing context. Lets say your callback function looks like this:
function(editor,context) { ... }
Using the context you will get the record which get edited by accessing context.record while the editor can provide you with the editor form from where you have access to all rendered fields within that form. To get the form you have to get the editor first and after that you can fetch the form
var form = editor.getEditor().getForm()
This way you can be sure that the editor has been set up. To get a field within that form simply call
form.findField('fieldname') // fieldname is the dataIndex of the column
You can now do nearly anything based on your conditions.
In addition the record is also loaded into this form an can be accessed by calling form.getRecord()
I am trying to build an app using Extjs 4.1. In general: it is a viewport with a tree panel on the west and a center panel that it is actually a tab-holder panel. When a user clicks on a tree node a tab is populating the center view. I had set an attribute in the tree panel that after selecting a node it gets deselected (deselectAll). The reason for this is that the user can open many tabs from different places (e.g. within each tab). But, when I set the above attribute it is producing an error (the “data” is undefined). The data that is undefined is the data related to the tree node. So, the question concerning selection model:
How can I address this problem (a solution may be to select the fist node, but I don’t want it)?
As for the history utility, I need to implement browser back button. Especially I want to disable browser’s refresh button. If user opens let’s say 15 tabs and accidentally click on browser refresh or “F5” he/she will lose everything. I had tried many things but with no luck. I am unable to understand “Ext history util”. So,
Is there any good example out there?
Could anybody guide me on how to do it?
Notice that the application is built respecting the new “MVC” architecture.
Stopping the refresh event is pretty easy - providing that your page has the focus. E.g.:
Ext.getDoc().on('keypress', function(event) {
if (event.getCharCode() == event.F5) {
event.stopEvent();
console.log('you stopped the refresh event');
}
});
If your page doesn't have the focus then there is nothing that can be done -- but most of the time your page loses the focus when a different browser tab is opened so F5 would be refreshing that one instead anyway.
I'm still trying to wrap my wits around the first part of your question, but I thought I would share that.
EDIT
Instead of trying to stop the default browser behavior (something which you can only do on a very limited basis), you should look into state management. ExtJS supports it on many components. Here is the ExtJS example.
Also here is the example of controlling browser history.
Since the ComboBox and FilteringSelect use a 'dijitPopup' whose DOM element gets inserted just before the closing body tag (presumably to help with ensuring it appears above everything else z-index-wise) this means that if the ComboBox is contained in an element that scrolls independent of the window itself and the user opens the dropdown and then scrolls the window (or whatever containing element) using the scroll wheel, that the menu part doesn't move with the control itself.
Is there a straightforward way to ensure that the menu part of the view remains positioned correctly relative to the control itself rather than simply assuming that its starting position is ok?
EDIT: appears to be a known issue (http://bugs.dojotoolkit.org/ticket/5777). I understand why they put the dijit popup just before the closing body tag for z-index stacking and overflow clipping issues, but it seems like it's maybe not the ideal way to do things given the bug in question here and things like:
You can restrict the Dijit theme to only small portions of a page; you
do this by applying the CSS class of the theme to a block-level
element, such as a div. However, keep in mind that any popup-based
widget (or widgets that use popups, such as dijit.form.ComboButton,
dijit.form.DropDownButton, and dijit.form.Select) create and place the
DOM structure for the popup as a direct child of the body
element—which means that your theme will not be applied to the popup.
~ from http://dojotoolkit.org/documentation/tutorials/1.6/themes_buttons_textboxes/
Not sure if this is the very best solution, but here's what I came up with:
Since the widget may be programmatically added/removed, and to avoid coupling a solution with some particular surrounding markup that we can't always count on in all cases, what I did was to hook the _showResultList and _hideResultList methods of ComboBox and when the popup opens, traverse up the DOM till we reach the <html> tag, adding onscroll listeners on each ancestor.
The handler for the onscroll event is simply:
var myPos = dojo.position(this.domNode, true);
this._popupWidget.domNode.parentNode.style.top = '' + (myPos.y + myPos.h) + "px";
where this is the widget in question. I scope the handler to the widget using dojo.hitch. In the close method I remove the listeners. I have to clean up the code a bit before it's presentable, but when it's finalized I'll add it to this answer.
Note: I only show updating the y position here. Part of the cleanup is to add x position updating in case someone scrolls horizontally.
Though its old I just faced this same problem and it looks like a Dojo issue and the fix is available here https://bugs.dojotoolkit.org/changeset/30911/legacy
This is a bit of a strange one, but I've been struggling for a few hours now and I can't understand what is happening.
I was wondering if anyone else has experienced this problem, and can perhaps explain it. I'm building a simple Winforms app and trying to use many of the built in controls.
Basically, I've got a form with a user control and some data capture fields. 3 of the fields are dropdown lists and on the user control I have a bindingSource control that binds directly to a Product class.
At run time I provide an instance of the Product class to the BindingSource and the class contains a property of ProductType. For simplicity I also added a List<ProductType> ProductTypes to the Product Class which loads itself when queried, which means I can just use the same bindingSource and choose the ProductTypes Data Member as the Datasource for the dropdownlist.
Upon running the form, the list binds perfectly and I can see all the product types listed, and I can select one and tab or click to the next field. But obviously the selected value won't bind because I've not chosen any bindings-SelectedValue for the dropdown, only a datasource. As soon as I make sure that the drop down modifies the instance of the Product by binding to the Bindings-SelectedValue, and then run the form, the list still gets populated perfectly and I can tab through the controls as long as I don't make a selection from the dropdown. If I make a selection from the dropdown then the dropdown holds focus. I cannot tab out for love or money and can't even click cancel button on the form, the close button top right is the only button I can click which works and I can't click any other field or dropdown. This affects all three dropdowns as soon as a selection is made.
Anyone have any ideas what I'm missing?
I have tried changing a few things and had some success by feeding the dropdown values a string[] instead of a member of an object. That seems to work, but defeats the object of using databinding doesn't it?
Any help appreciated!
Just guessing here, because I don't have time to set up a test and confirm right now, but are you doing any validating? I seem to remember that data-bound controls won't let you leave if the contents don't validate. Even if you aren't explicitly, try setting CausesValidation to False to see if there's any sort of validation going on behind the scenes, that might at least give you a hint.
Thanks for the input on this, helped me wrap my head around this.
In my case, it turns out that an exception was being thrown in one of the EventHadlers for my ComboBox.Validating event. It was hard to track down, because the IDE didn't show me that exception. I was able to modify the Exception behavior (in the debug menu) and have it show me any InvalidOperationException that was being thrown, and then I was able to track it down.
As Tom suggested, turning off CuasesValidation was the ticket to figuring it out.
For Infragisticst Dropdowns (may not be true for other winform dropdowns): If you have "LimitToList" set to true you can be stuck in a dropdown that you can't get out of without realizing it. Use the ItemNotInList even to trigger a warning message.