Unify HtmlSerializers

[cdalexndr]

Currently there are multiple public HtmlSerializer-type classes: HtmlSerializerInnerOuterText, HtmlSerializerNormalizedText, HtmlSerializerVisibleText, the deprecated HtmlSerializer.

They share some logic and it's confusing to have multiple classes.
They should be merged into a single class, with options.

[rbri]

I made the tree out of the one because the handling of all the differences in one class was too complicated.

[cdalexndr]

Here's my use case:
I need a html serializer that respects white-space (such as HtmlSerializerVisibleText) but also include invisible elements (like HtmlSerializerNormalizedText with ignoreMaskedElements_=false).

[rbri]

Ok, let me think about this, maybe there is some chance to work with a class hierarchy and some abstract methods.

[cdalexndr]

HtmlSerializerNormalizedText and HtmlSerializerVisibleText seem to be similar, just that HtmlSerializerVisibleText has more logic such as css white-space handling and only prints visible text.
Also HtmlSerializerInnerOuterText seems to be a trimmed down version of HtmlSerializerVisibleText (duplicate code).

(Excluding HtmlSerializer that isdeprecated and should be removed in future.)

[cdalexndr]

Did you worked on this issue? Can I have a go?

[rbri]

Yes you can, I can merge this at the start of the next week Am 5. November 2021 23:02:54 MEZ schrieb cdalexndr ***@***.***>:

Did you worked on this issue? Can I have a go? -- You are receiving this because you commented. Reply to this email directly or view it on GitHub: #399 (comment)

-- Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet.

[cdalexndr]

Why are there two browser-like serialization?

• DomNode.asNormalizedText() using HtmlSerializerNormalizedText
• DomNode.getVisibleText() using HtmlSerializerVisibleText for selenium interop

Isn't the selenium representation sufficient?
Does it make sense to deprecate HtmlSerializerNormalizedText, and add a visible option to HtmlSerializerVisibleText to be also used to serialize all (including non-visible) text?

[rbri]

HtmlSerializer is gone

[rbri]

HtmlSerializerNormalizedText implements HtmlUnits way of serialization (more or less backward compatible to the old asText() method. I like to have this to have our own way to serialize.

HtmlSerializerVisibleText is there to try to implement the support needed to be in sync with the browsers (and all the strange things they do).

[cdalexndr]

But, isn't the browser serialization, the real world usage, so the only correct way of doing things? Why is the custom htmlunit impl needed?
It would be simpler to keep only the real-world serializer and deprecate other ones.