simplify API for creating PDF

[asolntsev]

now user needs to write less code to generate PDF from HTML

This commits adds 2 new apis:

1. class Html2Pdf (a single method for generating PDF from HTML)
2. renderer.createPDF(doc, os) (instead of old sequence setDocument, layout, createPDF)

The work continues...

[asolntsev]

@andreasrosdal Yes, generally we could add some validation/html cleanup.
But using regular expressions doesn't seem to be a good idea for this purpose: https://medium.com/thecyberfibre/stop-parsing-x-html-with-regular-expression-2cf13215b411

[andreasrosdal]

How about using Jsoup to fix invalid XHTML?
https://jsoup.org/

In general I agree that regular expressions should not be used to fix bad HTML. However, we could use something, and at least regular expressions is one of the possible alternatives. Jsoup is probably better.

[andreasrosdal]

https://github.com/jhy/jsoup/

[asolntsev]

@andreasrosdal Yes, JSoup sounds good.
Could you pelase share some example of such "invalid XHTML" needing a cleanup? I am trying to understand what problem we want to solve.

[andreasrosdal]

Here are some examples of html changes had to be made to make the html XHTML valid for flying saucer for PDF export:

[asolntsev]

@andreasrosdal Thank you for the samples. Some of these could be really replaced automatically (e.g. <link> -> <link></link>), but some others are actually invalid, and FS should throw exception in these cases (e.g. <<i class="">).

[andreasrosdal]

I wish that FS would handle these cases, and not throw exception, in the same way that Firefox and Chrome is fault tolerant for invalid HTML in many cases.