Difference in Parser behaviour between 1.11.3, 1.15.4 and 1.16.1-SNAPSHOT with respect to whitespace and newline

[Reeniya]

Hi,
We are seeing a difference in parser behaviour between 1.11.3, 1.15.4 and 1.16.1-SNAPSHOT with respect to whitespace and newlines between the html tags and text. We are using xml parser and prettyprint set to true.

For example:
Input:

						<div class="freetext"
							id="text1">
							1111
							<div align = "left">
								<b>V2</b><br />
								<br />
								Xdxux<br />
								v3yfygygyg
							</div>
						</div>

with Jsoup version 1.11.3 we get the parser output has

       <div class="freetext" id="text1">
         1111 
        <div align="left"> 
         <b>V2</b>
         <br /> 
         <br /> Xdxux
         <br /> v3yfygygyg 
        </div> 
       </div> 

with Jsoup version 1.15.4 we get the parser output has

       <div class="freetext" id="text1">
        1111 
        <div align="left"><b>V2</b><br /><br />
          Xdxux<br />
          v3yfygygyg
        </div>
       </div>

We can re-create this issue using: https://try.jsoup.org/~HLY5GwlDvfC8Fn8tiGSpgCyIFFo

I observed that some fixes was done around whitespace and newline character so I consumed 1.16.1-snapshot version.

with Jsoup version 1.16.1-snapshot version we get the parser output has

       <div class="freetext" id="text1">
        1111 
        <div align="left">
         <b>V2</b>
         <br />
         <br />
          Xdxux
         <br />
          v3yfygygyg
        </div>
       </div>

We are bit confused about which is the correct behaviour as we are upgrading the jsoup version from 1.11.3 to 1.15.4
to understand the difference I have highlighted it in the image below which was captured using notepad++ to showcase the difference in whitespace between various versions of jsoup

In 1.11.3 we see only <div> tag, but with 1.15.4 newline between <div> and <b> tag is lost.

With 1.11.3 we see there is a whitespace after every <br /> and after the text "v3yfygygyg" but with 1.15.4 and 1.16.1-SNAPSHOT the whitespace is not there.

with 1.11.3 we see there is a whitespace after <div align="left"> which is not present with 1.15.4 and 1.16.1-SNAPSHOT version of jsoup

with 1.11.3 we see there is a whitespace after </div> tag which is not present with 1.15.4 and 1.16.1-SNAPSHOT version of jsoup

@jhy We are bit confused with what is the correct behaviour of the parser. Was there an issue with 1.11.3 which is fixed now or is this a new issue? Please can you let us know what should be the correct behaviour.

Thank you....

[brdeepak]

@jhy,
Please let us know how to move forward with this problem if you have any suggestions or assistance would be appreciated.

Thank you in advance

[jhy]

Hi,

The output of the pretty-printer is subject to change as we make improvements.

If the output of the printer causes a change to the way a browser renders the HTML, I would generally consider it a bug. E.g. see #1926. But there will be changes between releases. I believe the current output is better than the previous output, and so am inclined to keep it as-is.