Trailing whitespace on text-align right break

[Ben Giese]

I am experiencing the following problem:

I have a specific structure in my source html, file attached (pdf.html.txt), that somehow creates a trailing whitespace when using text-align: right and the text is using more than one line.

When I use that html in my implementation like in the attached example.java.txt, it will add a trailing whitespace. Shown in the wrongOutput.pdf.

When I copy the html into your line-wrap-pre-wrap test case html and run the testLineWrapPreWrap() in TextVisualRegressionTest.java it works perfectly fine, also attached. Is there anything I am doing wrong in the preprocessing of the html or the implementation?

I figured out, that some html is not accepted in your test cases, like a normal
tag, while I dont get an error using something like that in my implementation.

Attachments:

example.java.txt

pdf.html.txt

line-wrap-pre-wrap---actual.pdf

wrongOutput.pdf

[Ben Giese]

There is a fix for this issue, which I guess could be noted somewhere:
In your test cases you are using a different TextBreaker:

SimpleTextBreaker implements FSTextBreaker

So what I did is copying the class from the test into my project and use it like this:

`byte[] pdf;

    try {
        os = new ByteArrayOutputStream();
        builder.toStream(os);
        builder.useFastMode();
        builder.useSVGDrawer(new BatikSVGDrawer());
        builder.useUnicodeLineBreaker(new SimpleTextBreaker());
        builder.run();
        pdf = os.toByteArray();
        os.close();
        os = null;`

It is important to have the SimpleTextBreaker after the SVGDrawer, otherwise its not rendering imported PDFs.

[Ben Giese]

Thank you for the effort and fixing this issue!
Any plans for a new release with the fix any time soon?
The solution with the SimpleTextBreaker won't work for japanese, as there are no whitespaces and any text is rendered in one line. But with your fix I can use the standard linebreaker again, which will solve the japanese problem as well.