Site is not parsed properly

[ivanlabsii]

This page is not parsed like in browsers: http://chandra.harvard.edu/photo/2015/ngc6388/

I believe that the cause are 'learn' tags that are not closed (which causes then their parent

tags to be removed).

[JonathanMagnan]

Hello @ivanicin ,

Thank you for reporting.

We will try to investigate this issue during the weekend and try to come back with a solution next Monday.

Best Regards,

Jonathan

[JonathanMagnan]

Hello @ivanicin ,

We started to investigate the issue.

The current problem is the "p" is not handled like a normal tag. See: https://github.com/zzzprojects/html-agility-pack/blob/master/src/HtmlAgilityPack/HtmlNode.cs#L118

Something that both me a lot is when I read the following comment:

<p>bla<p>bla will be transformed into <p>bla<p>bla and not <p>bla></p><p>bla</p> or <p>bla<p>bla</p></p>

I would have expected the following scenario <p>bla></p><p>bla</p>, certainly not the "p" remain unchanged!

A p tag cannot contain another p tag, so it should close automatically the previous one.

Anyway, since the learn tag is unclosed, and p tag doesn't work like other tags, the learn tag is closed to the closest parent which is the div element instead of the p tag.

That's what is currently happening.

We are continuing our investigation.

Best Regards,

Jonathan

[ivanlabsii]

Hi, thanks for your feedback. For now I found a workaround (detecting the problem and using web browser to get the properly parsed content in that case), which isn’t perfect - for example Apple TV has no browser and this also slows down everything. Best regards, Ivan Icin Labsii [email protected] www.labsii.com From: Jonathan Magnan Sent: Saturday, May 20, 2017 4:33 PM To: zzzprojects/html-agility-pack Cc: ivanicin ; Mention Subject: Re: [zzzprojects/html-agility-pack] Site is not parsed properly (#1) Hello @ivanicin , We started to investigate the issue. The current problem is the "p" is not handled like a normal tag. See: https://github.com/zzzprojects/html-agility-pack/blob/master/src/HtmlAgilityPack/HtmlNode.cs#L118 Something that both me a lot is when I read the following comment: <p>bla<p>bla will be transformed into <p>bla<p>bla and not <p>bla></p><p>bla</p> or <p>bla<p>bla</p></p> I would have expected the following scenario <p>bla></p><p>bla</p>, certainly not the "p" remain unchanged! A p tag cannot contain another p tag, so it should close automatically the previous one. Anyway, since the learn tag is unclosed, and p tag doesn't work like other tags, the learn tag is closed to the closest parent which is the div element instead of the p tag. That's what is currently happening. We are continuing our investigation. Best Regards, Jonathan — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[JonathanMagnan]

Hello @ivanicin ,

We didn't successfully fixed the issue yet but today we will release a new version.

In this version, an option will be added to make the p tag act like a div and span tag.

Until we re-write the parser, I believe this will be a good alternative for you.

This version should be available in a few hours.

Best Regards,

Jonathan

[JonathanMagnan]

Hello @ivanicin ,

The v1.5.0-beta3 has been released:
https://www.nuget.org/packages/HtmlAgilityPack/1.5.0-beta3

We added a temporary fix, you can read it more in detail here:
https://github.com/zzzprojects/html-agility-pack/releases/tag/v1.5.0-beta3

Could you confirm us if this temporary fix is working?

Best Regards,

Jonathan

[JonathanMagnan]

Closing comment: No confirmation, feel free to re-open if you still have the issue.

[ghost]

Hi @JonathanMagnan,

I believe I'm seeing a related issue in 1.5.1 - When encountering a P block with no content, HAP will remove the closing P. For example:
Starting HTML: <p style="font-size:1px;"></p><p>test</p>
HTML after HAP: <p style="font-size:1px;"><p>test</p>

Using HtmlAgilityPack.HtmlDocument.DisableBehavaiorTagP = true; seem to resolve this issue. Should we expect to continue to use that option in future releases?

[JonathanMagnan]

Hello @chrisnelsondotca ,

I'm currently right now looking how to fix all this kind of issue. People report them with p, a, etc.

The current library seems to doesn't work well when the HTML is badly formatted with some kind of tag.

btw, please try to ask or report a issue/question on a new thread, it makes easier on my side to follow them. By example, since this thread is closed, there is not a lot of chance that I will answer to it once it fixed.

Best Regards,

Jonathan

[JonathanMagnan]

You can easily reference an issue by simply adding the issue number...

Ex: Issue #1

[ghost]

Hi @JonathanMagnan thank you, I appreciate the response. Should I create a new issue referencing this one then?

[JonathanMagnan]

Please yes ;) So this issue could stay closed.