We recently raised a ticket: 1424874.
We would like to request this ability to be added as a feature.
An error is thrown in the browser's console when pasting text + image from Word into the Kendo UI Editor: "Not allowed to load local resource: file:///C:/Users/..."
This is due to the inability of the browser to access the image file, located on the client machine.
The explanation of the above is the following. If you test the following sample (note, that it does not use Kendo) in Chrome, and try to paste text + image in the contenteditable <div>, you will observe the same error. The Editor also renders its content in an editable div element. The browser does not recognize that the user attempts to paste an image. Instead of reading the image data and loading it as a base64 string to the src attribute of the <img> element it generates an <img> tag which points to the clipboard location of the file on the client machine. As the browser is not allowed to access such resource, it trows the error observed and the image is not rendered.
To resolve this issue, when pasting only an image (no text), the Editor has its own internal implementation. Therefore, the user is able to paste images (e.g. clicking on the image in Word->Ctrl +C->then Ctrl + V in the Editor). However, when pasting it together with text, the widget does recognize the text and does not apply the custom processing aimed to convert the image. Therefore the native browser behavior of pasting images is exhibited.
Dear staff,
I am working extensively with Kendo UI Drawing API, specifically with drawDOM and exportPDF methods. I encountered a strange issue, which leads to the generation of corrupt PDF files.
After several attempts to analyze the problem, I found a small piece of code that allows to systematically reproduce the issue. You can find the code as attached.
Opening the example page with Chrome browser I got a corrupted PDF file, while Internet Explorer seems to generate a valid PDF file starting from the same source page.
Checking with an online PDF validator, the file generated by Chrome exploits this validation error:
“The content stream contains an invalid operator.
The operator has an invalid number of operands.”
I’m not skilled with PDF internals, but comparing the valid file of IE and the invalid file of Chrome, I noticed that an “infinite” value appears in the invalid PDF.
The row in the valid PDF is “1 0 0 -1 -9494.7900391 -9965.5996094 cm“, while the same row in the invalid PDF is “1 0 0 -1 Infinity Infinity cm”.
I noticed that the presence of an “Infinite” value is the common behavior whenever the corrupt PDF problem occurs.
When using a grid with locked columns if the unlocked area is scrollable once scrolled horizontally the resize handle keeps it's original absolute positioning and doesn't respect the change in scroll.
When hovering over the edge of the column I can no longer see the resize handle. I can hover to the left where it would have been with no scrolling and then see it.
On DOM inspection I see the event to update the handle's position STILL FIRES when hovering the edge of a column header, but the position is based on scrollLeft() === 0.
I will assume the needed left position for each column is calculated on creation but is not re-calculated for horizontal scroll.
This can be reproduced in the live demos on Telerik's site.
https://demos.telerik.com/kendo-ui/grid/frozen-columns
Hi there
This is just a general question in relation to Kendo UI. Does the Kendo UI component from the DevCraft susbcription has a wrapper for expressjs/nodejs
If not, there is any plan to add this to the library
Cheers
AL
dojo: http://dojo.telerik.com/UMuRaBax/3
Steps to reproduce:
1. Run the dojo, let the comobox list load and let it select the default item ('Oulu' in this case).
2. start typing 'Madrid' in the input field of the combobox.
3. press Tab key and observe behavior.
We are currently in the process of converting our site from using RadAjax controls into Kendo UI controls. We use a lot of drop downs (single select and multi select). While working with the Kendo UI controls, we saw that there were no multi select drop down controls available. There is a multi select control, however it uses a 'token' style and is not a control that we can use in our environment. There should be a multi select drop down control similar to the control used in RadAjax framework.
The following typings are missing in @types/kendo-ui version 2019.3.0
Hi,
I'm not sure this will be a bug in your mind but I just wanted to make sure I understood how things work. We have a grid definition which includes the dataSource.schema.model definition. One of the fields (TradeDate) in the schema has a type = "date". In addition, we assign local json data to the "data" property of the dataSource. Our JSON data comes in with dates in the ISO format (a string with the "T" in it). When the grid is initially created everything works fine. We notice that the TradeDate field in the dataSource.data() is converted to an actual javascript date. I assume this is happening because of the schema definition. Later we retrieve new data via an ajax call and apply it to the grid using the grid.dataSource.data(newJson) command. However this does not seem to process the TradeDate field and convert it into a real javascript date. This causes problems because we have filters applied based on the TradeDate. It works fine the first time but fails any time the data is refreshed (because the field value is a string). We tried the schema.parse function but that also is only called the first time. Ultimately we coded the grids "dataBound" event and everything worked. It gets called each time the data is refreshed. I was just wondering if I could be doing something differently so that the schema handled the conversion all the time.
self._grid = grd$.kendoGrid({
columns: [...],
dataSource: {
schema: {
model: {
id: "somekey",
fields: {
NewTradeFlag: { editable: false },
TradeDate: { editable: false, type: "date" },
...
}
}
},
data: self.ViewModel.DisplayData.TransactionData
}
...
})
Later on we do this
self._grid.dataSource.data(newJsonData);
P.S. we originally were using MVVM binding straight to the ViewModel but that didn't work either.
Thank you in advance, Jim.
See dojo at https://dojo.telerik.com/@GaloisGirl/akAlaDIP .
The first DateTimePicker is set as readonly and has also aria-readonly manually set to true. It is correctly read by my screen reader (NVDA).
The second DateTimePicker is only set as readonly. It is read by the screen reader editable. It is not marked as readonly in the accessibility tree (per Firefox Developer Edition 71).
I can't find a formal specification, but this page says an input with ARIA role="combobox" needs to have "aria-readonly" set in addition to "readonly".
I'm documenting this for the DateTimePicker , but the same problem occurs for ComboBox ( https://dojo.telerik.com/@GaloisGirl/egeLOxIt ) and possibly other widgets that use the ARIA role combobox.
I've setup a new ASP.NET Core project with Kendo UI Core. One of the requirements for this site is to load the scripts/styles from CDN except when the internet is down, in which case it should load from local files. I downloaded a copy of Kendo UI Core 2019.3.1023 and extracted it into the project. I have an ASP.NET Core Tag Helper which handles hashing the local file and using that SRI to load the CDN version.
Here's an example of that generated tag and the fallback:
<link rel="stylesheet" href="https://kendo.cdn.telerik.com/2019.3.1023/styles/kendo.common.min.css" integrity="sha384-tr4PSEV2CnVkku/o7s0ApjrlIpaxzBOMQTYc8SsPFXXh6ITUZ/SS7t8k8PVUa1Tk" crossorigin="anonymous">
<meta name="x-stylesheet-fallback-test" content="" class="k-widget"><script>!function(a,b,c,d){var e,f=document,g=f.getElementsByTagName("SCRIPT"),h=g[g.length-1].previousElementSibling,i=f.defaultView&&f.defaultView.getComputedStyle?f.defaultView.getComputedStyle(h):h.currentStyle;if(i&&i[a]!==b)for(e=0;e<c.length;e++)f.write('<link href="'+c[e]+'" '+d+"/>")}("box-sizing","content-box",["/lib/kendo-ui-core/styles/kendo.common.min.css"], "rel=\u0022stylesheet\u0022 integrity=\u0022sha384-tr4PSEV2CnVkku/o7s0ApjrlIpaxzBOMQTYc8SsPFXXh6ITUZ/SS7t8k8PVUa1Tk\u0022 crossorigin=\u0022anonymous\u0022 ");</script>
When the hash is calculated from the local file, the CDN copy does not match. So far I have only had an issue with two files: kendo.common.min.css and kendo.bootstrap-v4.min.css: both of these files have different headers from the download and CDN versions and appear to be minified differently. It would be great if future versions could be the same so that the hash for a local file matches the hash for the CDN version.
As a side note, it would be really great if you could work with Libman (the Visual Studio CDN downloader tool) to be able to download Kendo UI Core. The only version that exists currently on that system is very old.
On https://demos.telerik.com/kendo-ui?_ga=2.149841974.1082962146.1579102350-1080570292.1579102350 , all links in the "Sample Applications" section return 404 (for example https://demos.telerik.com/bootstrap/ )
Removing all URL params helps.
Any URL param ( https://demos.telerik.com/kendo-ui?x=1 ) reproduces the issue.
The event handling between DatePicker and DateTimePicker is inconsistent.
If I manually type a date into DatePicker and tab away from the input, the change event is triggered
If I do the same in DateTimePicker, no event is triggered
They should behave the same, of course DateTimePicker should also trigger whether or not the user types in a time
Hi,
I'm using the Grid version 2019.3.1023 but changelog for newer version doesn't seem to show this has been fixed.
Using Html.Kendo().Grid().Columns(columns => colums.ForeignKey(x => x.ForeignCol, data, "value", "label"));
If data contains html, it will not be escaped when the javascript for the kendoGrid component is rendered, and thus is exposed to XSS.
This should be either fixed or documentation and samples must tell explicitly we need to html encoded the data, as this is not the case for the model itself.
Hello,
The function kendo.support.mobileOS is returning false when on Chrome browser using Huawei mobile, model MRD-LX1F.
If using Firefox on same mobile, it returns true.
Android version is 9
EMUI versio 9.1.0
my contact : fuadk@gordano.com
or fuadks@gmail.com
Regards
Fuad
As a developer, I want to be able to add paragraphs with empty content and not have it stripped, so that I can add custom functionality, like icons.
I want to be able to control if the paragraph's content should be stripped or have the editor allow empty elements when they have classes.
Currently, if I add markup similar to a font awesome icon or some custom icon in a paragraph, there is a function, _fillEmptyElements, that strips the content of the paragraph:
Input:
<p> <i class="fa fa-check" style="color: green;"></i> </p>
or
<p> <span class="customicon icon-iconname"></span> </p>
On save, the paragraph's content is stripped to:
<p> </p>
As a workaround, I can add dummy text in the paragraph:
HY,
i need a or advanced and user friendly tablet edit and format (for example: cell context menu for formatting, multi cell select for formatting and merge, and so on )
( i want replace all external editor in my application for the Kendo Editor, but i need more user friendly table and image format/edit functionality, some like the other WYSIWYG editor (ChkEditor, TinyMCE,...) )
Nice feature will we the document format with text flow ( like google docs), so i can do a document in a A4 page size and flow text over the pages.
Thanks
Hi,
Every widget that supports control of visibility configuration uses "VISIBLE", but when using MVVM to control HTML objects it is: data-bind = "invisible: someThing" . This IMHO is inconsistent, counterintuitive, and adds complication. If you are controlling the visibility of a mix of Kendo and HTML objects on the same bit of data, you need to duplicate it with a negated version or resort needlessly to templating logic. Maybe I'm missing something but this seems like a no-brainer.
Thanks,
Will Dougherty
Reported for iPad Air 2, tested and reproduced on iPad Pro 2020 12.9' iOS 13.5.1.
Dojo example.
mobileOS returns false.
mobileOS returns an object with information about the mobile OS, device and browser.