OptionOutputAsXml adds spaces to comments when using InnerHtml

[spassarop]

Description

I have some cases with HTML comments which I need to preserve their syntax. Some HTML examples can be modified (by changing the Comment property on a HtmlCommentNode). When I set OptionOutputAsXml to true, spaces are always added on different places.

In the fiddle I provide below I have a working example of the issue. This are my basic cases:

Original HTML	HtmlDocument.DocumentNode.InnerHtml
<div>text <!-- comment --></div>	<div>text <!-- comment  --></div>
<div>text <!--[if IE]> <!--[if gte 6]> comment <[endif]--><[endif]--></div>	<div>text <!--[if IE]> <! - -[if gte 6]> comment <[endif] -->&lt;[endif]--&gt;</div>

For the first row case: A space has been added after "comment", now having two spaces.
For the second row case: Two spaces have been added on the second "<!--" and one after the first "[endif]".

Reading the code and changing options that don't conflict with my needs took me to no solution, also I don't understand why this could be normal behavior. That's why I'm creating this issue.

Fiddle

Custom Fiddle to reproduce the issue: https://dotnetfiddle.net/TkTqt0

Further technical details

• HAP version: 1.11.24.
• NET version: netcore 3.1, .NET Framework 4.7.2.

I hope there is a solution I'm not seeing or, if this is a real issue, it's easy to solve and could be included in the next release.
Thank you in advance.

[JonathanMagnan]

Hello @spassarop ,

He looked at this issue and here is the line that causes it: https://github.com/zzzprojects/html-agility-pack/blob/master/src/HtmlAgilityPack.Shared/HtmlNode.cs#L1830

It seems that comment, when OptionOutputAsXml is enabled, is rewrited.

The fix is easy but one problem I have is the Backward Compatiblity which people which currently use this library for a long time and which we will make a break for a small fix.

[spassarop]

Well, that's a shame. I wouldn't like to use my own version of the library as the update process would be tedious compared with just updating a Nugget package.

I guess my only option is doing that if I want to keep using HAP, unless you have a workaround in mind. Close the issue if that's not the case.

Thanks for your time.

[JonathanMagnan]

As said, I don't think changing the default behavior is a good thing since this library is so much used (We did it a few times in the past and every time a few days later, multiple people were reporting that their projects were no longer working).

However, I asked my developer to check a little bit more this week to check if we can add an option to better manage it.

[spassarop]

I understand, I know backward compatibility is hard to maintain and usually there is a trade-off on fixing or keeping an issue.

Again, thank you very much for your time :)

[JonathanMagnan]

Hello @spassarop ,

A new version has been released with the option OptionXmlForceOriginalComment. When enabled, it will exactly write the original comment instead of creating one.

Let me know if that's exactly what you were looking for.

Since that's a new option, this is still time to improve it without impacting anyone.

[spassarop]

Hello @JonathanMagnan,

That option actually works and yes, it was exacly what I was looking for, keeping comments unchanged.

Thanks to the team!