Increase of 404 error after change of encoding

[RetailMeNotFr]

Hello,

We just have launch a new version of our website with a new utf-8 encoding.

Thing is, we use comma as a separator and since the new website went live, I have a massive increase of 404 error of comma-encoded URL.

Here is an example :

http://web.bons-de-reduction.com/annuaire%2C321-sticker%2Csite%2Cpromotions%2C5941.html

instead of :

http://web.bons-de-reduction.com/annuaire,321-sticker,site,promotions,5941.html

I check with Screaming Frog SEO and Xenu, I can't manage to find any encoded URL.

Is anyone have a clue on how to fix that ?

Thanks

[RetailMeNotFr]

I will take a look at it but it's not the issue that SEOMoz tell me as this format concerns only images. It's actually a little trick to do lazyloading on images.

The link you pointed out on your example is good ("/annuaire,amkashop,site,promotion...) as comma are not encoded.

And for your example I see no issue except capitalization.

I bet this is a Moz problem because when I fetch as Googlebot, I don't find encoded URL...

[BlueprintMarketing]

just wanted to give you one more thing that I think would help http://www.w3schools.com/html5/att_meta_charset.asp

I believe you should clean up your encoding and that it will not be a big deal.

Sincerely,

Tom

[BlueprintMarketing]

I thought this may help as well because you do have to clean up your source code

The online quoted-printable encoder tool first encodes the input text in either UTF-8 or ISO-8859-1. The characters are then output according to this schema:

| Character | Result | Comment |
| "=" (0x3D) | =3D | Special handling of the equal sign |
| " " (0x20) to "~" (0x7E) | Unmodified | Printable ASCII (7 bits) |
| Any other | =XX | Hexadecimal char code |

Since quoted-printable does not in itself specify the text character encoding, it is important to specify this correctly when used. The online quoted-printable decoder tool attempts to auto-detect the text encoding.

See the Wikipedia article on quoted-printable for more info.

[BlueprintMarketing]

I would use a tool similar to this http://www.percederberg.net/tools/text_converter.html

as you can see your links for your gif photos are encoded "data:image/gif;base64"

please give it a try and tell me if that helps?

Sincerely,

Thomas

[RetailMeNotFr]

Hello and thanks for your answer.

No word involved here.

We move from :

http-equiv="content-type" content="text/html; charset=iso-8859-1" />

to

charset="utf-8">

Everything is fine except for Mozbot

[BlueprintMarketing]

what you need to do is go into your site and cleanup the links that have been converted and messed up because of the change. Once you clean them you will have no problem this is what your links look like

data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH+A1BTQQAsAAAAAAEAAQAAAgJEAQA7

utf-8 is definitely the right coding it's very good you just have to go in and clean it up looking your source code

"

| {"m":2571,"a":"wrap"}" width="108" height="65" data-original="/upload/merchants_logo/108-65/amkashop.jpeg" src="data:image/gif;base64,R0lGODlhAQABAIAAAP///////yH+A1BTQQAsAAAAAAEAAQAAAgJEAQA7"> <noscript></span><img data-merchant="2571" class="merchantLogo lazy" data-out="{"m":2571,"a":"wrap"}" width="108" height="65" src="/upload/merchants_logo/108-65/amkashop.jpeg" alt="Amkashop"><span></noscript> |
|   |

<a <span="">href</a><a <span="">="</a>/annuaire,amkashop,site,promotions,2685.html" title="Amkashop">Code promo Amkashop"

|

I hope I've been of help to you.

Thomas

[BlueprintMarketing]

Did you happened to possibly write it using Microsoft Word and paste content in? Or are you speaking about a website that you converted from another encoding to Unicode 8?

sincerely,

Thomas