Object created by #dup puts validity error to the console.

[chengguangnan]

If you need the background, check flavorjones/loofah#79.

Here is a demo:

irb(main):039:0> Nokogiri::HTML("<i id=a></i><i id=a></i>").text
=> ""
irb(main):040:0> Nokogiri::HTML("<i id=a></i><i id=a></i>").dup.text
element i: validity error : ID a already defined
=> ""

So the object created by #dup will puts error messages to the console. How to stop this behavior?

[jirutka]

I have a very similar issue, many irrelevant “validity error : ID already defined” and don’t know how to get rid of it. 🙀

[flavorjones]

Looking into this!

[flavorjones]

Hi, thanks for reporting this. This error message is emitted by libxml2 during the xmlCopyDoc call here in xml_document.c.

Usually when creating a document (e.g., while parsing), we'll capture any errors and then attach them to the document's errors instance variable (here is an example of this error handling, also from xml_document.c).

My understanding, after a cursory inspection of libxml2, is that when copying a document, libxml2 makes a duplicate of each node in the current document before reparenting it to the new document. So, momentarily, there is a violation of HTML's id semantics, which leads to this error being printed to the console.

While I can't do anything within the confines of Nokogiri about the generation of the error message, we can certainly trap it and put it into #errors like other parsing commands. I'll try to do this today.

[flavorjones]

Whoa, scratch the third paragraph -- there's just an error building the document due to how you structured the HTML markup. In any case, we'll fix the console emission.

[flavorjones]

This will be in 1.6.5 out soon.

[subimage]

I'm still getting a ton of these errors to my console with 1.6.5 ... did the fix make it in?

element form: validity error : ID commentform already defined
element input: validity error : ID author already defined
element input: validity error : ID email already defined
element input: validity error : ID url already defined
element textarea: validity error : ID comment already defined
element input: validity error : ID submit already defined
element input: validity error : ID comment_post_ID already defined
element input: validity error : ID comment_parent already defined
element input: validity error : ID akismet_comment_nonce already defined

...etc

Looks like the commit got removed. Shouldn't this be re-opened?

[jirutka]

Me too, the problem still remains in 1.6.5.

[s3ththompson]

Problem still remains for me too. I'm getting a stream of errors

element a: validity error : ID 988b already defined
element a: validity error : ID 69ba already defined
element a: validity error : ID 8c03 already defined
element a: validity error : ID a0af already defined
element a: validity error : ID b86e already defined
element a: validity error : ID 5c30 already defined
element a: validity error : ID 9559 already defined
element a: validity error : ID aec6 already defined
element a: validity error : ID b73d already defined
element a: validity error : ID 07f7 already defined
element a: validity error : ID 911a already defined
element a: validity error : ID bd9b already defined
element a: validity error : ID 1fe6 already defined
element a: validity error : ID 4c6c already defined
element a: validity error : ID 238b already defined
element a: validity error : ID 0895 already defined
element a: validity error : ID 9073 already defined
element a: validity error : ID 5ebd already defined
element a: validity error : ID 99da already defined
element a: validity error : ID 4298 already defined
element a: validity error : ID fa29 already defined
element a: validity error : ID aa2e already defined
element a: validity error : ID 52f5 already defined
element a: validity error : ID 64be already defined
element a: validity error : ID 8c89 already defined
element a: validity error : ID 55c4 already defined
element a: validity error : ID cecd already defined
element a: validity error : ID bfd6 already defined
element a: validity error : ID 4d94 already defined
element a: validity error : ID aa93 already defined
element a: validity error : ID e424 already defined
element a: validity error : ID 6fb2 already defined
element a: validity error : ID ac41 already defined
element a: validity error : ID 7af1 already defined
element a: validity error : ID 16c4 already defined
element a: validity error : ID 0f5d already defined
element a: validity error : ID f039 already defined
element a: validity error : ID 205c already defined
element a: validity error : ID 1467 already defined
element a: validity error : ID 5f9e already defined
element a: validity error : ID 569e already defined
element a: validity error : ID c9f2 already defined
element a: validity error : ID 71e5 already defined
element a: validity error : ID bcaf already defined
element a: validity error : ID 321a already defined
element a: validity error : ID f4da already defined
element a: validity error : ID 8e4b already defined
element a: validity error : ID 2b34 already defined
element a: validity error : ID 7ca4 already defined
element a: validity error : ID 18df already defined
element a: validity error : ID 64b9 already defined
element a: validity error : ID a601 already defined
element a: validity error : ID b9b7 already defined
element a: validity error : ID 9296 already defined
element a: validity error : ID c90b already defined
element a: validity error : ID 8173 already defined
element a: validity error : ID 1805 already defined
element a: validity error : ID d165 already defined
element a: validity error : ID ae05 already defined
element a: validity error : ID 772c already defined
element a: validity error : ID 0128 already defined
element a: validity error : ID 2142 already defined

[conf]

I can still reproduce the bug on nokogiri 1.6.5. This is the minimal test case:

test = Nokogiri::HTML.parse("<span id='test'></span>"); 
test.at_css('span').clone # emits element span: validity error : ID test already defined

[wipxj3]

Having same problem as @conf. Suggestions?

[zenspider]

@flavorjones this ticket should be reopened and #1208 closed.

[flavorjones]

Nod.

[flavorjones]

So commit 1696dc8 fixes this issue for document parsing, but not for fragment parsing, which is what #1208 reports. Looking.

[flavorjones]

Fix will be in the next release, in the next week or so.

[Thibaut]

@flavorjones thank you!

[henry74]

How do I get the next version - running bundler update still gives me 1.6.5

[jirutka]

@henry74 It hasn’t been released yet, 1.6.5 is currently the last release available on RubyGems, that’s why bundler gives it. @flavorjones is currently working on it.

[flavorjones]

1.6.6.1 was just released moments ago.

[henry74]

Thanks @flavorjones. When I had looked on github, the latest version was 1.6.6 but I didn't see it on rubygems :-) Just funny timing...

[Basher52]

Hi,
I'm still getting this error, version of nokogiri (1.6.6.2, 1.5.5).
Apart from that the parsing seems to be doing what I want.
Any suggestions?

[zenspider]

@Basher52 do you have a reproduction? If so, please file a new issue with the reproduction.

[flavorjones]

What @zenspider said. This issue has been closed, let's please open a new one if you're seeing something unexpected.

[Basher52]

Thanks for the tip!

I'm not quite sure, but it could be caused by nokogumbo instead of nokogiri...
So I've put the issue on the nokogumbo github.
You can take a look at it if you want:
rubys/nokogumbo#21