Hi,

thanks again for the super cool localizer

I got a problem today setting it up with 2 domains

http://www.marketingfan.com (For english)
http://www.marketingfan.com (For german)

I also got "similar entries" installed which worked like a charm so far before I used localizer with it

However, today it starts returning URLs of posts that should be posted on the .at domain only, as well on the .com domain

e.g.

the german post

http://www.marketingfan.at/allgemein/marketingfan-jetzt-auch-deutsch

has the localizer "english" link correct as http://www.marketingfan.com/general/marketingfan-site-updated-drupal5-fr...

However "similr entries" spits out

http://www.marketingfan.at/general/marketingfan-site-updated-drupal5-fre...

which will cause duplicate content issues when indexed.

I noticed, that the major flaw here is, that there are no redirects in place,
but the multilingual content is served on EACH domain.

This opens up a flaw for evil competitors to bombard your domain with links to those "unwanted" pages,
which in turn will cause a lot of duplicate content issues.

I think a general "redirect" solution will be required in localizer,
regardless of the "similar entries" malfunction today

please advise,
thanks - christoph

Comments

Christoph C. Cemper’s picture

Title: Similar Entries plugin doesn't take care of domain prefixes » Localized content is served on EACH domain ==> problem: duplicate content vuln
Christoph C. Cemper’s picture

Priority: Normal » Critical

There are more problems arising from this.

The language selector block always displays a flag for the installed languages,
but it should only display the flag for the TRANSLATED languages,
otherwise this leads (once again) to duplicate content

e.g. check this german post here

http://www.marketingfan.at/suchmaschinen/seo/text-link-aufbau/spammy-blo...

which does NOT have an english translation (so far)

but the flag "english" is displayed, and leads to an url that should not exist
(germanish url on english domain)

These problems again

* cause duplicate content issues in search engines
* cause confusion with readers finding german content on english domains

Roberto Gerola’s picture

Category: bug » support

> I got a problem today setting it up with 2 domains
> http://www.marketingfan.com (For english)
> http://www.marketingfan.com (For german)

The domains are the same ?

i reassign the category to support request, because the lack of support for a 3rd party modules
cannot be considered a bug.

> The language selector block always displays a flag for the installed languages,
> but it should only display the flag for the TRANSLATED languages,
> otherwise this leads (once again) to duplicate content

The language selector is intended to be presented on every page, not only
on content's pages, but also, admin, taxonomy, front page and on.

In every case it is possible to write a customized block selector if you want.

Thanks for your feedback.

Roberto

Christoph C. Cemper’s picture

Hi Robert,

thanks for your comprehensive feedback!
and sorry for mixing up support request (for the similar entries module)
and the bug that duplicate content is served on multiple domains.

I made a mistake in the initial post - the problem is on the domains

http://www.marketingfan.com (english)
and
http://www.marketingfan.at (german)

in addition to the problems with similar entries, the more major issue
from my point of view is that the language selector provides a lanugage selection
(in the other language) even if the content has NOT been translated.

so if you take a look at this post

http://www.marketingfan.at/suchmaschinen/seo/text-link-aufbau/spammy-blo...

there's flag presented, which links to the .com domain with the german url alias
and also the german content

That doesn't make sense from a user's view
and even more it's a problem as google detects duplicate content across the domains
marketingfan.at and marketingfan.com then

>In every case it is possible to write a customized block selector if you want.
Yes - I want :-) what do I need to do?

Goal should be that the block selector

- displays a flag for EACH translation
- if a translation is not present for a node, then the flag may not be displayed
- if lanugage locale selection is made by domain (as I use it - great feature! thanks!)
and the url alias for language 1 is called up on domain for language 2, then a 301 redirect
must be issues to redirect the user and search engine (google) to the correct language domain

the last point is especially important to make sure all plugins/add modules not (yet) compatible with
localizer don't make the dupe content problem worse

I'd love to help with implementation if you point me to the right places!

Robero, will you also come to DrupalCon2007 in Barcelona in Sept?

best,christoph

Roberto Gerola’s picture

Are you using version 2 ?
I cannot help you with that version.
I didn't write that version and it has been abandoned, afaik.

If can help you only on 1.x branch.

Clarify this point, please, than we can see how to proceed.

>Robero, will you also come to DrupalCon2007 in Barcelona in Sept?
No. I don't think.

Thanks.

Roberto

Christoph C. Cemper’s picture

Hi Roberto,

no, I'm using 5.x-1.9

Christoph C. Cemper’s picture

Version: 5.x-2.x-dev » 5.x-1.9

Hi Roberto,

no, I'm using 5.x-1.9

Roberto Gerola’s picture

Duplicate content problem ?
Is this due to "Similar entries" module ?

Similar entries
I don't know and don't use this module and localizer doesn't provide support for this, at now.

Language block
A specialized block should be written than appears only when you are viewing a node.
Also flags with all translations available for a node would be a feature to add.

Christoph C. Cemper’s picture

Robert,

let's forget the "similar entries" module - this is a second issue that has nothing to do with this problem.

The issue is, that Localizer is serving up content on domains that do not relate to their language.

So if I got two domains,

domain.de for german
domain.us for english

and I set these up in the localizer config, I except localizer to server only pages
of the appropriate language.

This means, if

domain.de/english-node is called, it should 301 redirect to domain.us/english-node

if

domain.us/german-node is called, it should 301 redirect to domain.de/german-node

to make sure only the language content written for a domain is served up on it.

christoph

Christoph C. Cemper’s picture

ADD: the problem is already caused by localizer, BECAUSE it generates links wrong

on domain.de/german-node it generates a link to domain.us/german-node which does not make sense at all...

got it now?

Roberto Gerola’s picture

> got it now?
Yes.

I'll fix this problem.

Thanks.

Christoph C. Cemper’s picture

Roberto, thanks - I'll help with testing as soon as you got it ready

Christoph C. Cemper’s picture

any news roberto?

Roberto Gerola’s picture

Christoph C. Cemper’s picture

HI Roberto,

what did you change?

it's still the same problem

when I call up an english node on the german domain, it's presented as if everthign was fine.

what I expected was that when an english node is loaded on german domain, it does a redirect to the english domain

I turned on "switch by node language" but this hasnt' changed behaviour

can you comment on what you changed in this version?

thanks
christoph

Roberto Gerola’s picture

Christoph C. Cemper’s picture

Roberto, sorry, but it appears nothing has changed...

Still the site serves the german content on the english domain,
when using the german nodeids/urls on the english domain.

Again, expected behaviour would be a redirect to the german domain

Christoph C. Cemper’s picture

roberto,

I looked into your code and I have a hard time finding even the place where the header-redirect is made for the particulr page if the content is not in the correct language as the current domain... can you point me to the correct code lines?

Christoph C. Cemper’s picture

still no redirects and duplicate content issue in full effect on both domains mentioned :(

juanruben’s picture

StatusFileSize
new1.77 KB

Hello,
I have been working on this:

I changed the behaviour when the translation of a node is requested by domain and it doesn't exist, in that case the browser is redirected to the same page but in the right domain.

e.g. In the example you commented above,

If I try to get this page, http://www.marketingfan.at/a4u-expo-2007-shoutouts , that points to a node in english from a domain especified as german it now redirects to the same node but in the right domain http://www.marketingfan.at/a4u-expo-2007-shoutouts , because the german version is not found.

It also takes care of the links shown in the block, so only the translations available for the current node are shown.

Hey Roberto, let me know if you like these modifications to be included in the module.

A patch for revision 1.2.2.59 is attached.

Juan

Christoph C. Cemper’s picture

Thanks Juan!

notice to all - the MarketingFan installation is already patched, so what you see when clicking above link
and it's redirect to the proper .com domain is already because of that patch.

This patch also removes the wrong links in the language,
which used to show links to pages (i.e. for german) even if that node did not have a translation

Roberto Gerola’s picture

Thanks.
I'll integrate this improvement and release it in the upcoming 1.11 version.

Roberto

juanruben’s picture

StatusFileSize
new2.6 KB

Hello all,
I realized one mistake in my code. When a node has no locale set properly and no translation exists for it the page was redirecting endlessly to itself.
I changed the code so now it always checks for a valid locale before trying to redirect, also checks as a final check that the domains are actually different before redirection.

Hope it is finally fixed.
Greetings.