#199 "whitespace: pre" CSS property not taken into account

[Arkanosis]

Hi,

When the CompactHtmlSerializer does its job, it attempts to handle pre tags and spaces in a smart way. However, in real HTML found on the dirty web, the way to interpret spaces can also be governed by the “whitespace” CSS property (in particular “whitespace: pre”). In such cases, the CompactHtmlSerializer breaks the HTML by making the hopeful but wrong assumption that it's self-contained.

The issue is at least present from 2.9 to 2.21 (I haven't checked earlier versions) and I'm pretty sure it's the result to the following code:

87                         if (childrenIt.hasNext()) {
88                             if ( !Utils.isWhitespaceString(childrenIt.next()) ) {
89                                 writer.write("\n"); // <- oops, significant in a "whitespace: pre" context
90                             }
91                             childrenIt.previous();
92                         }

The SimpleHtmlSerializer does not exhibit this issue.

Thanks,

[Scott Wilson]

Thanks Arkanosis,

Its definitely a problem if we're introducing characters into the output that are going to break the content. I'll try a few test cases with that line removed and see what it does to the output.

[Scott Wilson]

Commenting out that statement doesn't seem to cause any issues, so I've committed the change to trunk.