Interface DomIntegration
DOM Integration & Third-party Components
Smart GWT provides a huge variety of pre-built components and allows you to create new components via combining and composing the existing components. However in rare cases, it can make sense to write code that works directly with raw HTML and the browser's DOM (document object model) APIs. This level of coding is also involved when integrating third-party JavaScript components into Smart GWT applications.First, a warning: when you use HTML and the DOM directly, all of Smart GWT's guarantees of cross-browser consistent behavior no longer apply. When you use Smart GWT's API, Smart GWT is automatically compensating for many, many browser bugs - not just trivial things like different property names or missing utility methods, but problems where browsers fail to fire certain events, report sizes wrong only in certain modes with certain styling, or bugs that only occur with specific timing.
Before deciding to do direct HTML coding, consider whether you really want to expose yourself to all of these possible issues. If you can achieve the same look and feel and behavior through Smart GWT's APIs, that's usually best.
Adding or modifying the DOM
The DOM structures used by Smart GWT necessarily differ by browser in order to work around each browser's specific bugs. This DOM structure is intentionally undocumented because it is subject to change without notice - in may be necessary to modify the DOM structure to work around the bugs in each new browser release.
Instead of trying to modify the Smart GWT-generated DOM, you should always add new
elements. For a new standalone component that will be based on direct use of HTML, you
usually do this by subclassing Canvas and overriding Canvas.getInnerHTML()
and returning
an HTML string representing the components you want to create.
You can use a similar approach anywhere HTML is allowed in a widget property: formatting APIs
for StaticTextItem, DetailViewer, TileGrid, and other DataBoundComponents, as well as places
such as Tab.title
or Button.title
.
Third-party components
Most third-party JavaScript components have the ability to generate their HTML content into a DOM element specified by ID, or to replace such an element with new HTML. This is true of Google Maps, CKEditor and many other libraries.
To use this form of integration, implement a Canvas.getInnerHTML()
function that returns
a DOM element with a known ID, then have the third-party library target that DOM element once
the Canvas is drawn. For example, CKEDITOR.replace() takes the ID of a <textarea>
element, and the following code would create a <textarea> and have the CKEditor replace
it with a CKEditor widget:
public class CKEditor extends Canvas { private static native void replace(String id) /*-{ $wnd.CKEDITOR.replace(id); }-*/; public CKEditor() { setRedrawOnResize(false); } @Override public String getInnerHTML() { return "<textarea id='" + getID() + "_ckEditor' style='width:100%;height:100%'></textarea>"; } @Override protected void onDraw() { CKEditor.replace(getID() + "_ckEditor"); } }
This same approach can be used when you want to insert third-party generated HTML into just a
specific part of a Smart GWT widget. For example, you might want to insert
JQuery
'sparklines', which are
essentially miniature charts, into cells of a ListGrid. You could use
a cell formatter
to
write out <div> elements with
known IDs into the cells, then target them with JQuery.
If a third party component is being embedded into a Smart GWT widget, developers should
consider which overflow setting
is most
appropriate to use.
Oftentimes you'll want your embedded component to size to the available space. In this case
you should use overflow:"hidden"
. However if the embedded component
will render at a specific size, and you want to have the component size to fit it,
overflow:"visible"
may be appropriate.
Developers embedding third party text editing components into Smart GWT widgets
should typically set Canvas.canSelectText
to true
on the Smart GWT
widget. This prevents the Smart GWT event management system potentially interfering with
any native events such as focus, selection and copy/paste behavior as the
user interacts with the embedded editor.
Resizing and Redraw
When implementing canvas.getInnerHTML()
, your getInnerHTML() function will be
called every time the component redraw()s, and the new HTML will replace the old.
Also by default, your component will redraw() if it is resized. In the example above with
CKEditor, we wouldn't want this because it would wipe out the CKEditor widget every time it was
resized, so we set Canvas.redrawOnResize
to false. In other circumstances you may want
to redraw on every resize in order to generate new HTML.
If you do not redraw HTML on resize, you either have to write the HTML in a way that makes it
flow into available space (width/height 100% may be enough here) or you need to manually
resize the DOM element when the resized event
fires.
In the latter case, you should adjust the size of the DOM element to match the
inner width
and
inner height
of the containing
Canvas. The "inner" dimensions
are the dimensions after border and margins have been subtracted, so this will work even if a
border is added to your Canvas subclass via CSS or Canvas.setBorder()
.
Other redraws
Once you have set Canvas.redrawOnResize
to false you may still see redraws from other
sources. Generally this would be from code in your application which calls
Canvas.redraw()
or Canvas.markForRedraw()
unnecessarily. To
troubleshoot, you
can enable the "redraws" log category in the Developer Console - this will log the source of
any redraws in the system.
Component-specific considerations
In addition to complete redraws, certain components will refresh areas of the DOM dynamically without a full redraw. For example if you are using a ListGrid cell formatter to write out a particular DOM structure in a grid cell, this structure may be removed from the DOM or regenerated in a number of ways including explicit calls toListGrid.refreshCell()
,
automatic cell refresh due to data changing, cells being shown or hidden due to
incremental rendering
, etc.
Developers will need to
consider how best to handle the lifecycle of the DOM elements they createHTML customization within other components may require similar consideration.
Another consideration specific to ListGrids is that fact that in some circumstances the
HTML returned by a cell formatter may be rendered outside the grid body. An example
of this is the drag tracker HTML generated when ListGrid.dragTrackerMode
is
set to "record". The inactive cell formatter
may be provided to emit alternative HTML for rendering these 'inactive' versions of
the cell content (allowing developers to, for example, exclude explicit element IDs).
Masking during drags
Third-party components that utilize iframes, browser plugins or other special elements will
"swallow" events such that Smart GWT never receives them. This is a problem whenever a
drag interaction goes over the component, including drag resizing of the component itself.
To avoid this issue, set Canvas.useDragMask
for any component that contains an iframe
or browser plugin, or that appears to be swallowing events during drag. The telltale sign
is that when the mouse goes over the plugin, the visual effect of the drag (differs by
Canvas.dragAppearance
) stops
updating or stutters.
Overflow & Auto-Sizing
Consider which Canvas.overflow
setting
to use for your custom component:
-
overflow:"hidden"
is the most common. In the context of third-party components, it means the component is prepared to render itself at any requested size (above a minimum). -
overflow:"visible"
means you want Smart GWT to attempt to automatically determine a minimum size based on the HTML content generated by the custom component -
overflow:"auto"
is similar tooverflow:"hidden"
, but means your custom component needs Smart GWT to create scrollbars whenever its HTML content does not fit in the allocated space.
overflow:"visible"
or overflow:"auto"
, if you make on-the-fly
modifications to the HTML you returned from getInnerHTML()
, there is no
cross-browser-reliable way for the Canvas to detect this and auto-size again. Instead, call
Canvas.adjustForContent()
to
trigger auto-sizing again.
zIndex
zIndex values control what component is rendered "on top" when multiple components appear in the same area of the page.
To work around various browser issues, Smart GWT components use a very high range of zIndex values. If a component creates pop-up widgets such as hovers or floating toolbars via direct HTML/DOM usage, these pop-up widgets will appear behind all Smart GWT components unless they set a very high zIndex.
For your own custom HTML/DOM components, the simplest strategy is to create pop-up widgets
based on Smart GWT components, even if they are triggered by interacting with
hand-created HTML. For example, even if you've written some kind of advanced SVG-based data
visualization component, you can still implement pop-up configuration dialogs based on the
Smart GWT Window
class; there's no reason to implement such
dialogs directly in
low-level HTML/DOM code.
If a third-party widget is creating pop-ups you don't directly control, you may be able to
configure the third-party widget to use a certain zIndex range, or you may be able to
directly reach into the widget's DOM and modify zIndexes that way. You can use
Canvas.getZIndex()
to discover the zIndex
of any Smart GWT widget you need to
appear above, then set a higher value.
Finally, as a last resort and completely unsupported approach, you can modify the zIndex range used by Smart GWT by executing the following JavaScript code via JSNI, in your onModuleLoad() function :
private native void adjustZIndexRange() /*-{ $wnd.isc.Canvas.addClassProperties({ // default zIndex for the next item to be drawn _nextZIndex:200000, // zIndex of the next item to be sent to the back _SMALL_Z_INDEX:199950, // zIndex of the next item to be brought to the front _BIG_Z_INDEX:800000 }); }-*/; public void onModuleLoad() { adjustZIndexRange(); ....... }
Tab Order
Focusable elements embedded in a Smart GWT page should be accessible via tabbing. Developers may use theTabIndexManager
to ensure custom
elements are
reachable via tab keypresses and show up at the appropriate place in the page's
tab order. For details on how to achieve this, see the overview here
.
For a more general overview of the tab order within Smart GWT,
see the tab order overview
.
Other issues
There are several other issues, listed below, for which there really is no general strategy for solving the issue, although some general pointers are provided.
Because of problems like these, it's a very very bad idea to freely intermix components from multiple component libraries. While mixing components may appear to be an appealing strategy and you may experience apparent success with early attempts, the issues below will ultimately prevent you from completing an application of sufficient quality for enterprise use.
In the following discussion, "third-party widgets" should be understood to include widgets that you write using direct DOM/HTML techniques.
- tabbing order / accessibility: a correct tabbing order that visits all
components on the page is a requirement for your application to be considered accessible, as
is ARIA markup (for more information, see
Accessibility
).
Third party widgets may or may not write out ARIA markup. This may require you to modify them or reach into their DOM to add ARIA attributes. - modality: aside from zIndex issues covered above, modality means that the tab
order should be a closed loop that reaches only active widgets, which can create additional
complexity in getting tabbing to work correctly. Also, keyboard shortcuts should be
disabled for inactive widgets; this may require calls to
EventHandler.targetIsMasked()
to make third-party widgets respect Smart GWT modality, or may require calls toCanvas.showClickMask()
to cause Smart GWT components to consider themselves inactive when a third-party widget opens a pop-up that is intended to be modal. Multi-layered modality, such as a modal window that in turn pops a modal dialog, is yet more complex. - bad CSS: some third-party widgets introduce CSS selectors that target, for example, every table cell on the entire page. This very bad practice will interfere with Smart GWT (or any other HTML on the page). This may require modifying the third-party component, or extensively modifying Smart GWT CSS to reverse any changes caused by third-party CSS. For example, it may be necessary to modify every Smart GWT CSS style that may be applied to a table cell to reverse a change in padding for all table cells that is introduced by bad third-party CSS.
- skinning: third-party widgets may lack sufficient skinning APIs to allow you to
match
look and feel to SmartGWT, which may necessitate creating a custom SmartGWT skin to match
the look and feel of third-party widgets (see
Skinning Overview
) - event interference: third-party widgets may register page-wide event handling logic that conflicts with or destroys similar event handling logic in Smart GWT. For best results, load third-party JavaScript libraries before Smart GWT since Smart GWT makes a best effort to preserve any previously installed handlers and allows such handlers to cancel native browser behaviors if they do so.
- RTL / i18n: third party widgets may not allow all user-visible messages to be replaced, a requirement for internationalization / localization, or they may not support RTL/BIDI (Right-To-Left / Bi-Directional) rendering
- avoid using third-party widgets if you can build equivalent functionality in Smart GWT
- if the third-party component is completely non-interactive, either does not require ARIA markup or already includes such markup, and there are no conflicting look and feel issues, go ahead and use it
- if you anticipate issues, consider the Feature Sponsorship Program as a means of getting new supported functionality added to Smart GWT
- search for existing posts and/or ask about the feasibility of integration on the Smart GWT Forums.
- finally, you could attempt to tackle the issues above on your own. To avoid wasting time on dead ends, we would recommend assessing the amount of work involved in fixing all problems that need to be solved before attempting actual fixes for any one issue.