I've just encountered something that I don't quite understand. I received a document (administrative memo from my employer) containing a web address. The address is not a clickable hyperlink, it is just text.
What is interesting is that when the address is copy & pasted into a web browser address bar, it causes the web browser to attempt to contact a different web address than the cut & pasted text contains. The address text initially appears to be pasted correctly into the address bar, until I hit enter -- then instantly the text changes to something else.
Please note that this is not a matter of simple web site redirection. I know this because if I manually type in the same address (instead of copy & pasting it from the original document), the "correct" address is loaded. It is only following the copy/paste/load process that text appears to be magically changing.
I have also noticed that if I copy & paste the address first into a Notepad text file, save the text file, close, re-open, and then copy/paste to the web browser, the "correct" site then loads. Of note, when I save, Notepad warns that there are characters in Unicode format which will be lost. So I assume that there is some hidden unicode text that is being stripped out when I save as plain text.
But, in Notepad if I enable the "Show Unicode Control Characters" option, I see nothing. So what could be going on here?
To get really specific, the domain transforms like this: http://www.aaaaaaaaaa-usa.com/bbbbb/ddddddtools.html ==> www.xn--aaaaaaaaaausa-km6g.com. (The browser of course reports that it cannot find the IP address of the server)
For compatibility, domain names should be ASCII text, so there is a standard (IDN) to convert other characters to ASCII, using the two letter prefixes followed by two dashes --.
Additional, there were some phishing attack, using letter on other alphabets, that looked like latin letters, so deceiving users. So some browsers choose to display the ascii name instead of the intended name. (It changes from browser to browser, and usually only on selected similar characters).
Related
I was fooling around on my phone and decided to try putting an emoji in the url bar of google chrome. I entered in đ.com, the emoji which is equivalent to unicode U+1F600. Chrome ended up evaluating that as http://xn--e28h.com/, which took me to a "webpage unavailable" screen (ERR_NAME_NOT_RESOLVED). I looked up xn--e28h on godaddy and it was unavailable.
Here are my questions:
Why did đ turn into xn--e28h? I don't see any relation with the unicode.
Why are domains of this format unavailable on godaddy?
Bonus question: why can't we put emojis in domain names?
DNS uses a special way to encode Unicode into ASCII. The xn-- prefix says that it's an encoded name, and since the whole name in this case is one Unicode codepoint the rest just looks incomprehensible. You can start reading more about this here.
Most (if not all) top-level domains have rules on which Unicode characters they allow for names in that TLD. For example, .SE only allows those characters that are used in one of the official languages of Sweden. This is entirely a policy thing, so the "why" gets fuzzy.
See 2.
If I navigate to the following URL with a special UTF8 encoded character I get different results in web browsers:
http://example.com/lörickÚ
Firefox 37 - Shows the correct URL as above.
Chrome 42 - Shows the correct URL as above.
Edge - Shows the correct URL as above.
IE 11 - Shows percent encoded URL http://example.com/l%c3%b6rick%c3%a8/
Where can I find a list of browsers and versions that support this feature and are there any announcements of whether the new Microsoft Edge browser supports this.
This StackOverflow post highlights the above issue for those interested.
What is shown in browser address bars is not necessarily what is used internally.
If you enter http://example.com/lörickĂš in Firefox, it gets shown like that, but it actually gets percent-encoded and becomes http://example.com/l%C3%B6rick%C3%A8. This is for usability reasons (or, if IRIs are not supported, like in HTTP/1.1, for transforming an IRI into a URI), so users donât necessarily have to enter the correct URL (with percent-encoding), and donât get confused by seeing these cryptic parts.
You can easily check what really gets used by copy-pasting the URL from the address bar into a text document.
So the three browsers from your example probably use the same URI (i.e., percent-encoded), but two browsers decided to display the un-encoded variant instead.
I have searched and searched and applied the obvious fixes but it seems I have another variant of the problem. I have PHP pages and these display what song is currently playing, what songs are coming next and last recently played on my web radio station, the info comes from mysql. The characters are displayed correctly on the php pages. This is where it gets tricky, I also have HTML pages which load 2 div's from a php page so that the coming up songs also display on those HTML pages but there that's when the accents characters don't show correctly, I have the correct meta tag in the header on those pages and have also used the .htaccess file trick (although I was not sure how important the location of the line in the file was so tried various places). I even opened my .htaccess in notepad++ to change the encoding to use UTF8 but no BOM. I even added a meta tag for UTF8 in the php page header and then the characters didn't work on php either, probably you're not supposed to. As you can see I spent a lot of time. What's interesting the characters display correctly on iPad, it's on the PC browsers it doesn't work. Maybe no one ever tried this before loading divs from php into HTML and have special characters too. Sounds interesting anyway and if anyone is interested in having a think that would be great but it's not a vital problem just a nice to have fix. The server side of my stuff is hosted on a hosting site
thanks
As Stack Overflow seems to be unable to create links from URLs that have spaces in them, copy and paste this URL into your browser.
http://grooveshark.com/#!/search/song?q=we will rock you
It does not redirect you to ...song?q=we%20will%20rock%20you or anything like that.
The spaces just simply stay there. When I first saw this, it looked so foreign to me. How is this achieved?
I believe they use javascript to set the contents of the url bar. You can use something like Live HTTP Headers to confirm that the browser definitely sends a request with %20 encoded spaces.
Itâs a browser setting. The browser decodes the URL, to make it more readable for humans.
If you copy the URL from the browserâs address bar and paste it into a text document, youâll see that the space characters are percent-encoded.
See How can I see how the browser percent-encoded my URL? (which is not visible on address bar)
I am working on a requirement to display (make readable) characters from the URL.
When I use Google Chrome, it displays the parameters in Chinese - even though they are encoded to UTF-8.
When I use Mozilla Firefox, it displays the parameters in Chinese - even though they are encoded to UTF-8.
When I use Internet Explorer, it displays the parameters encoded in UTF-8.
N.B. The URL is encoded to UTF-8; I know that because when I copy the URL from the three of them and paste it to Notepad++ the three of them display the following:
/%E6%89%93%E5%BC%80%E7%9B%AE%E5%BD%95/%E7%9B%B8%E6%9C%BA/%E6%95%B0%E7%A0%81%E7%9B%B8%E6%9C%BA/%E5%B0%8F%E5%9E%8B%E6%95%B0%E7%A0%81%E7%9B%B8%E6%9C%BA/PowerShot-A480/p/1934793
Could it be that Mozilla Firefox and Google Chrome guys have this improvement that can make an encoded String readable and perhaps the IE guys do not support that? Or, is there any way to activate that with IE?
By the way... Going to View >> Encoding >> Unicode (UTF-8) takes care of the text inside of the page but does not make any difference for the text in the URL.
Any help will be greatly appreciated!
I've written a blog post about Internet Explorer not displaying the decoded version of non-ASCII characters and using IRIs to solve the problem.
As of today, we have the following situation:
HTML5 supports IRIs, i.e. URIs with Unicode character support
HTTP does not support IRIs, but all major browsers take care of converting IRIs to valid (encoded) URIs to retrieve the specified resource (page).
IE supports IRIs in the href attribute of anchor tags and properly displays them in its address bar just like when you enter your URL by hand (keyboard ;-)).
If you choose to percent-encode your IRI thus making it a URI, IE will not decode that URI back into an IRI.
So you could try the following:
Save your HTML files using UTF-8. This allows you to insert any Unicode character into it.
Do not percent-encode your URLs inside your HTML pages' links. Just use links like this: äșŠæć·Žć„ć
”äč±
A great article on the topic can also be found at the W3C: An Introduction to Multilingual Web Addresses.