Table page break bug and PDFBOX character bug

[rototor]

This is a reduced testcase which shows two bugs:

• The repeated table header is broken on page break. openhtmtopdf shares this bug with flyingsaucer.
• The arrow symbol is not correctly resolved from the unicode
  character. Instead a # character appears. There are even two # characters on places were no character is shown... If you view the html in a browser you see the right symbols. This works fine in flyingsaucer - after I manually register an other font (e.g. Noto Sans) which has this unicode symbol, as the fonts included in iText 2.1.7 don`t have that character.

Do you know how this "table repeated headers on page break" stuff works? I tried to figure it out / debug it, but I just dont really understand where the repeated table headers are created. Somehow the top-margin is not calculated correctly after the page break. If you increase the margin-top in line 126 of the html file you can see that the header somehow gets more into the view. If you can give me some pointers were to look I can try to future investigate that problem. But at the moment I really dont know were to look.

I don`t think that examples is the right location for such a test case. But its not a unit case neither. Please tell me were I should place this files.

Thanks for reviving flyingsaucer under a new name! Beside some rare bugs like these its a very useful library to create reports (using html). And also to render html snippets to images (which I also use in my projects)

[danfickle]

Thanks @rototor - I'm not sure why font fallback is not working - I did test it. However, I'll take another look tomorrow. As to the table header issue, as you've no doubt found, the layout code is highly complex, but I'll take a dive into that too. I think examples is a reasonable place for test cases as it has all the required imports.

[danfickle]

@rototor - I changes a couple of lines (and added the NotoSans-Regular.ttf file) to your test case and got the arrow printing out in a PDF.

The lines were:

src: url("NotoSans-Regular.ttf") format("truetype");
/* We don't support format at this time so remove it: */
src: url("NotoSans-Regular.ttf");

builder.withHtmlContent(html, "");
/* We need a base url to resolve fonts so: */
builder.withHtmlContent(html, RepeatedTableSample.class.getResource("/testcases/").toString());

Let me know if you still can't get the arrow printing. I'm investigating the table header issue now.

[rototor]

@danfickle After I copied the Noto-Sans Font into the testcase directory, I see the arrows. But there are still # characters in the pdf - on every page in the left top corner of the "body" area. Where every they come from.... no idea.

[danfickle]

I think the table break bug is to do with the ContentLimitContainer getting the wrong page content limits, I'm just not sure how this is possible.