Shallow Rendering based on Component Factories #396
Comments
|
@FlorianRappl, it seems AngleSharp is converting all elements to lowercase by default, e.g. this raw markup string: <MyComponent />gets the I would like to keep the original casing when printing out the DOM nodes if possible, but think that it might not be how the HTML5 standard works. |
|
All DOM nodes are lowercase.
So this is nothing from AngleSharp, but rather from the HTML spec.
Correct! I think one way around is would be to keep the source references and to have a custom markup formatter, which looks for source references and prefers them. Quick example (sorry its a mess, just to get the idea out): void Main()
{
var config = Configuration.Default;
var context = BrowsingContext.New(config);
var parser = new HtmlParser(new AngleSharp.Html.Parser.HtmlParserOptions
{
IsKeepingSourceReferences = true,
}, context);
var document = parser.ParseDocument("<FooBar>Test</FooBar>");
document.ToHtml(new MyMarkupFormatter()).Dump();
}
class MyMarkupFormatter : AngleSharp.Html.HtmlMarkupFormatter
{
private string GetOriginalName(AngleSharp.Dom.IElement element)
{
if (element.SourceReference != null && element.SourceReference.Position.Index != -1)
{
element.Owner.Source.Index = element.SourceReference.Position.Index + 1;
var originalName = element.Owner.Source.ReadCharacters(element.LocalName.Length);
return originalName;
}
return null;
}
public override string OpenTag(AngleSharp.Dom.IElement element, bool selfClosing)
{
var content = base.OpenTag(element, selfClosing);
var originalName = GetOriginalName(element);
if (originalName != null)
{
return $"<{originalName}{content.Substring(originalName.Length + 1)}";
}
return content;
}
public override string CloseTag(AngleSharp.Dom.IElement element, bool selfClosing)
{
var content = base.CloseTag(element, selfClosing);
var originalName = GetOriginalName(element);
if (content.Length > 0 && originalName != null)
{
return $"</{originalName}{content.Substring(originalName.Length + 2)}";
}
return content;
}
}Result: <html><head></head><body><FooBar>Test</FooBar></body></html>Potentially, we could introduce a special parser option to keep the original names. As usual, my recommendation would be to stick to the standards, as deviating (e.g., via enabling a flag) could cause unwanted side-effects. |
|
Thanks @FlorianRappl. Yeah, I dont really want to change the default behavior of the DOM tree, its likely that some users of bUnit depend on that. The use case is for when we print out error messages when AngleSharp Diffing detects a diff between expected and actual markup, so I think your suggestion is just right. This is what I ended up with: 2b857ed |
|
Ill keep this issue open for now to see if there is any feedback from the upcoming preview release. |
|
Reopened due to bug in implementation. |
|
I'll close this again. There is rudimentary support for shallow rendering now, documented in the next release of the docs. |
The goal is to make shallow rendering possible, with as few caveats as possible.
It should:
TestContext.RenderTreeCascadingValue<T>which are part of the render fragment being rendered.Limitations:
ShallowRenderwill only render the first top level component. Other components, including its children will be stubbed.Render,ShallowRender, andRenderComponentmethods will all throw an exception, with the message The test context has previously been used to perform a shallow render of a component or render fragment. Rendering another component or render fragment with the same test context is not currently supported, since it can result in unpredictable results.. This limitation might be removed in the future, but right now I do not see a neat way to achieve this.Replaces #17.
The text was updated successfully, but these errors were encountered: