kotlinx.html
kotlinx.html copied to clipboard
Kotlin
head tag generated with forced meta tag
Hello,
I was using the JVM version of this library to generate an HTML file with a head tag on when I noticed that it generates a forced meta tag.
More specifically:
html {
head {
// some code
}
}
Generates to this:
<html>
<head>
<META type="Content-Type" content="text/html; UTF-8"/>
<!-- Some code -->
</head>
</html>
I've looked in various places to see if there's a way to prevent this from happening, but have had no luck.
How do you know that it is generated? Most likely it is appended by the browser
I never opened this in a browser. This is the direct output to a file, via a JVM program.
To clarify my issue, when I use the word "generated" I am using it in the context of Kotlin.html creating an HTML string and writing it to a file, ergo: "code generation".
My original issue was written at a bad time where I didn't have access to my computer, and was somewhat paraphrased.
The exact output that is "generated" looks like this:
<!DOCTYPE html>
<html>
<head>
<META http-equiv="Content-Type" content="text/html; charset=UTF-8">
<!-- Other Head Tags -->
</head>
<!-- Other HTML Tags -->
</html>
From my debugging I found that it always generates at the top of the head tag, and even generates if I leave the head block empty in kotlin:
fun example(html: HTML) = html.apply {
head {
// No code here
}
}
This will produce the following text when the html is serialized to a string and printed to an HTML file:
<!DOCTYPE html>
<html>
<head>
<META http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
</html>
As some final context, to make sure that there isn't anything possibly missing, the code that I use to write the html to a file is below:
val document: Document = // ...
val outputFile: File = // ...
outputFile.writer(Charsets.UTF_8).use { it.write(document, prettyPrint = true) }
Again, I have looked into a way to possibly prevent this, as I thought it might be my own fault, and my current fix is to post process the HTML string before writing it to a file, replacing the exact character sequence that is generated (<META http-equiv="Content-Type" content="text/html; charset=UTF-8">) with an empty string.
The code that I first encountered this behavior can be found here, although I doubt it helps, nor matters, as this has happened in completely separate testing I've done.
I'm pretty sure it's not kotlinx.html, may be it's Document writing something? E.g. similar issue is reported for Jekyll (https://github.com/jekyll/jemoji/issues/41)
For sure meta tag is generated by XML Transformer (javax.xml.transform.Transformer)
It's easy to see that DOM rendering itself looks like the following (excerpt from kotlinx.html)
fun Writer.write(element: Element, prettyPrint : Boolean = true) : Writer {
val transformer = TransformerFactory.newInstance().newTransformer()
transformer.setOutputProperty(OutputKeys.OMIT_XML_DECLARATION, "yes")
transformer.setOutputProperty(OutputKeys.METHOD, "html")
transformer.setOutputProperty(OutputKeys.ENCODING, "UTF-8")
if (prettyPrint) {
transformer.setOutputProperty("{http://xml.apache.org/xslt}indent-amount", "2")
transformer.setOutputProperty(OutputKeys.INDENT, "yes")
}
transformer.transform(DOMSource(element), StreamResult(this))
return this
}
Don't you have any dependencies that could override transformer factory or inject custom transformer via SPI? Which class of the instance returned by TransformerFactory.newInstance().newTransformer() ? I have it com.sun.org.apache.xalan.internal.xsltc.trax.TransformerImpl by default.
You also could try to rewrite the function above and configure transformer. Perhaps you could try to eliminate transformer.setOutputProperty(OutputKeys.ENCODING, "UTF-8")