Bug : XWPFRun loses <w:br> information; cannot distinguish page breaks from line breaks

[SimonBach1]

Hello,

I’m running into an issue with XWPFDocument. In a .docx file:

Page breaks are represented in the XML as:

<w:r><w:br w:type="page"/></w:r>

Line breaks (manual line breaks, Shift+Enter) are represented as:

<w:r><w:br/></w:r>

However, after opening the document with XWPFDocument:

The corresponding XWPFRun for both types of breaks has run.getText() == "\n" and run.getCTR().getBrArray() is empty.

This means there is no way to distinguish a page break from a line break when using XWPFRun.

I would expect getBrArray() to contain the <w:br> elements with their type (page or default) so that page breaks can be detected programmatically.

Is this the intended behavior, or is there a way to preserve and distinguish page breaks in POI?

Thanks in advance,

[centic9-dt]

Not sure, but I think getBrArray() should not be empty for page-breaks.

Can you attach a sample file which contains an empty getBrArray() for a page-break?

[SimonBach1]

Hello,

Sure, please find attached a sample DOCX file where page break is present, but XWPFRun.getCTR().getBrArray() returns an empty array.
saut_page.docx

[centic9-dt]

When running the following:

        try (XWPFDocument sampleDoc = XWPFTestDataSamples.openSampleDocument("saut_page.docx")) {
            for (XWPFParagraph paragraph : sampleDoc.getParagraphs()) {
                for (XWPFRun run : paragraph.getRuns()) {
                    final CTBr[] brArray = run.getCTR().getBrArray();
                    if (brArray == null) {
                        System.out.println("Having no br-array");
                    } else {
                        for (CTBr ctBr : brArray) {
                            System.out.println("Having ctBr: " + ctBr.getType());
                        }
                    }
                }
            }
        }

I get

Having ctBr: page
Having ctBr: null
Having ctBr: page

So getBrArray() is always available for me and CtBr.getType() returns "Page" for some of them.

Please check your code.