From the io.Writer documentation:

Write writes len(p) bytes from p to the underlying data stream. It returns the number of bytes written from p (0 <= n <= len(p)) and any error encountered that caused the write to stop early. Write must return a non-nil error if it returns n < len(p).

However, (*gohtml.Writer).Write returns the number of output bytes it writes to the underlying Writer instead of the number of input bytes it consumes.

package main

import (
	"fmt"

	"github.com/yosssi/gohtml"
)

func main() {
	h := `<h1>Pre</h1>
	
<pre>some code</pre>

<h1>Post</h1>
`
	fmt.Println(gohtml.Format(h))
}

Prints

<h1>
  Pre
</h1><pre>some code</pre>
<h1>
  Post
</h1>

I understand the need for whitespace-preservation inside pre, but the above looks wrong to me.

Lines like:

<script type="text/javascript" src="js/lib.js"></script>

or

<div class="clearfix"></div>

could be kept in a single line.

Preformatted tags like pre, code and textarea should be kept unchanged and with the original indentation.
Changes inside those tags produce changes in visualization and in values inside a form.

Hi! Thanks for creating the gohtml package. I'm using it in Orbiton for formatting HTML from within the editor.

However, JavaScript code that is embedded within <script> tags seems to have indentation removed, so that:

<script>
  if (1 + 1 == 2) {
      x = 42
  }
</script>

Is formatted like this:

<script>
  if (1 + 1 == 2) {
  x = 42
  }
</script>

It would be nice if some tags did not take their own lines, for instance, tags that are usually inline like: "a", "image", "span", "h1" and "h2" could be kept like this:

<p>
  Paragraph text with <span>an inline span element.</span>
</p>

These may not be supported for prettifying:

• Space at the end of a tag is left there: <a href="blah/" onmousedown="some_script" >
• A script tag will not have its indentions formatted.

The script tag part seems to leave the indention of:

<script>
  window.customElements.define(
		'name',
		class extends HTMLElement {
			constructor() {
				super();
				let template = document.getElementById('name');
				let templateContent = template.content;

				const shadowRoot = this.attachShadow({mode: 'open'}).appendChild(templateContent.cloneNode(true));
			}
		}
	);
</script>

So if I have the above, where most of the opening space is tabs, the prettifier seems to leave it alone. I think it would probably be safe normalize the space to the same as other HTML elements. Just makes it look strange otherwise.

However, there may be completely valid reasons for not trying to do these, as prettying HTML is a fairly complex procedure I imagine with all kinds of corner cases.

Thanks for taking the time to read this.

My use case is a program that uses gohtml to format an html file.

Unfortunately, the current behavior is that calling gohtml.Format() on previous result of gohtml.Format() creates additional indentation on each step, which makes this uses case impossible.

An example:

package main

import (
    "fmt"

    "github.com/yosssi/gohtml"
)

func main() {
    s := "<title>my title</title>"
    s = gohtml.Format(s)
    fmt.Printf("s1:\n%s\n\n", s)
    s = gohtml.Format(s)
    fmt.Printf("s2:\n%s\n\n", s)
    s = gohtml.Format(s)
    fmt.Printf("s3:\n%s\n\n", s)
}

Result:

 go run t.go
s1:
<title>
  my title
</title>

s2:
<title>
    my title
</title>

s3:
<title>
      my title
</title>

As you can see, "my title" gets indented more after each call to Format().

Desired behavior is that s2 and s3 are exactly the same as s1.

Inline tags defined in InlineTags variable are not "inlined"

<span>Hello World</span>

shows as

<span>
    Hello World
</span>

Is there anything I missed ? May be an option ?

Cheers

Jean-Luc

HTML entities (e.g. &) should be preserved in the output. This is especially important for entities that can mess up the HTML output if converted back into their corresponding character, such as <, >, and &:

gohtml.Format("<div>0 &lt; 1. great insight! &lt;/sarcasm&gt; over&amp;out.</div>")

gives:

<div>
    0 < 1. great insight! </sarcasm> over&out.
</div>

instead should be:

<div>
    0 &lt; 1. great insight! &lt;/sarcasm&gt; over&amp;out.
</div>