Bug with mime-type application/xhtml+xml in jquery 1.6.3

[pierre@…]

Hi,

I seems a regression has been added in jquery 1.6.3 for content-type "application/xhtml+xml".

Most browsers (ffx, Safari, Chrome) with such error message :

uncaught exception: [Exception... "An invalid or illegal string was specified" code: "12" nsresult: "0x8053000c (NS_ERROR_DOM_SYNTAX_ERR)" location: " Line: 1182"]

The error does not occurs with jquery 1.6.2.

Check this page as an example :

-> fails with jquery 1.6.3 while -> works (same page with with jquery 1.6.2)

Cheers

[anonymous]

Both pages appear to work in Opera.

[pierre@…]

Yes, it works with opera although there is content negotiation to work with old MSIE version.

The page is served with application/xhtml+xml for Opera too and it works.

Note that I am sure the bug is related to the mime type since page : ​http://vizionr.fr/index_1_6_3.html works in ffx, Safari and Chrome (exactly the same page as ​http://vizionr.fr/index_1_6_3.xhtml but served as text/html)

[pierre@…]

This bug is a bit annoying since ​http://ajax.googleapis.com/ajax/libs/jquery/1.6/jquery.js is now using version 1.6.3 for 1.6 branch.

[jaubourg]

It's a problem in support. Test string is not in xml anymore hence why the parser is complaining (since doc is supposed to be proper xml). Size optimization failure it seems.

I'll bring rwaldron in on this ;)

For now, Pierre, shouldn't be too much of an hassle to use 1.6.2 instead of 1.6 in that google CDN URL.

This is a blocker and could warrant a 1.6.4.

[dmethvin]

This bug is a bit annoying since ​http://ajax.googleapis.com/ajax/libs/jquery/1.6/jquery.js is now using version 1.6.3 for 1.6 branch.

Never use a URL like that in production code. Besides the situation here where the code changes and breaks your site, that file is cached for one HOUR but the 1.6.2 is cached for one YEAR.

​http://www.petefreitag.com/item/777.cfm

[dmethvin]

I agree this looks like we'll need 1.6.4 to fix. This is the ONE change that went in after rc1 and I didn't vet it carefully enough. Also, our unit tests don't run in XHTML mode; if we're going to support it we need to do that. I've opened a separate ticket for that issue.

[Rick Waldron]

I'll resolve this by the end of today.

[Rick Waldron]

Weird, it seems like links were removed from the original post when I replied from my mobile.

[Rick Waldron]

​https://github.com/jquery/jquery/pull/487

Protip: Don't listen to Mathias. Ever.

[mathias]

Replying to rwaldron:

Protip: Don't listen to Mathias. Ever.

Don’t use XHTML, ever. :P

[Rick Waldron]

Replying to mathias:

Replying to rwaldron:

Protip: Don't listen to Mathias. Ever.

Don’t use XHTML, ever. :P

+9001

[rwldrn]

Reverts broken support string. Fixes #10197

Changeset: 4bc691a1c3a67aefb867e868d5577155571fb659

[sS]

Hi,

I noticed the mistake at the same time that 1.6.4 was launched and indeed in XHTML the scripts don't work. I had to switch manually back to 1.6.3 to solve this (reading jquery from google cdn). Will this be fixed at version 1.6.5?

Thanks

[Rick Waldron]

There is no "1.6.4" yet, but 1.6.4 will have the fix for this issue.

[sS]

Thank you and sorry for the mistake, I meant: 1.6.3

[Rick Waldron]

No problem - I just wanted to make sure you knew that 1.6.4 will be the correct release

[Rick Waldron]

#10236 is a duplicate of this ticket.

[anonymous]

Replying to pierre@…:

Hi,

I seems a regression has been added in jquery 1.6.3 for content-type "application/xhtml+xml".

Most browsers (ffx, Safari, Chrome) with such error message :

uncaught exception: [Exception... "An invalid or illegal string was specified" code: "12" nsresult: "0x8053000c (NS_ERROR_DOM_SYNTAX_ERR)" location: " Line: 1182"]

The error does not occurs with jquery 1.6.2.

Check this page as an example :

-> fails with jquery 1.6.3 while -> works (same page with with jquery 1.6.2)

Cheers