Hi there,
When we try to traverse the styles present in the document and try to get the border or shading information out of the fonts, this is creating empty run property nodes (rPr) in the styles.xml. Attaching two documents -
- WithoutEmpty_rPr.docx
- WithoutEmpty_rPr_StyleTraversed.docx
Please use document 1, i.e; ‘WithoutEmpty_rPr.docx’, as a input document. Try the traversal code as following:
StyleCollection sampleStyles = doc.getStyles();
for (int i = 0; i < sampleStyles.getCount(); i++)
{
Style sampleStyle = sampleStyles.get(i);
try
{
Font sampleStyleFont = sampleStyle.getFont();
if (sampleStyleFont != null)
{
// sampleStyleFont.getBorder();
sampleStyleFont.getShading();
}
}
catch (Exception e)
{
throw new RuntimeException(e);
}
}
After running the above code, if we save the document to file system and unzip the styles.xml and try to compare it to input document’s styles.xml, we see the output as here.
At line number 834, 846, you would see the insertion of extra node, essentially the empty run property node, which is optimized by MS word.
Ideally accessors should not modify the document state but here we are seeing the violation. Please have a look into this. Thanks.