Re: read charset of meta tag

optimistx wrote:
Martin Honnen wrote:
...
On the other hand most browsers by now expose a property
like document.characterSet (Mozilla) or document.charset
(IE) which should give you the charset the browser has
taken from the meta or from real HTTP headers.

Is it true that these metatags have often different values
compared to what the browser actually uses?

That is certainly possible as that browser will follow the HTTP headers and take any character set declarations in the headers in preference to anything else (as it required in HTTP). And it is certainly common for the attributes of META elements to be at odds with HTTP headers (and even at odds with document mark-up; how often do you see XHTML mark-up contain a META element that attempts to assert that document to be textt/html?).

And the page author might have not a correct idea of the
character set, which the server is sending, thus believing
that putting something to metatags the server obeys him/her.

Yes, misconceptions about web technologies are rife.

The browsers in this mess try to conclude the character set
by examining the byte stream with heuristic rules,

That has been observed (particularly with IE (at least up until 6)).

getting often somewhat correct results.

And some spectacularly wrong result, hence the observation of the phenomenon.

I thought earlier that metatags are commands to the server:

For some servers they have been; the server would process the document prior to sending and base the headers used on the META elements. But that is not very common with servers.

if I put ISO-8859-10 to the tag, then the server transforms
the page to that characterset!

Eg javascript files on the server: I suspect most authors here
have not 100 % reliable and true info which characterset they
have.

Observing that some people don't know what they are doing is no reason to assume 'most here' do not.

At least I have not
:)

Get yourself a web debugging proxy (Fiddler or Charles) or some other tool that can show you the HTTP headers.

Richard.

.

Relevant Pages

  • Re: Inserted symbols and site publishing errors.
    ... The http headers being delivered by your server are: ... require UTF-8 ... >> sends an HTTP header forcing the character set to be ...
    (microsoft.public.frontpage.programming)
  • Re: How prevent users clicking on "Back" to bring back to previou pag
    ... You can't stop the user using features in their browser :-) ... set appropriate HTTP headers so that the page is expired and not cached ... store something on the server indicating that ...
    (microsoft.public.inetserver.iis)
  • Re: Caching problem
    ... You need to work out if IIS is caching the page (ie always sending out the ... same content), or the browser is, or some proxy server in between is. ... > just setting the http headers in IIS to expire content ... >>the real HTTP headers often override the meta tags. ...
    (microsoft.public.inetserver.iis)
  • Forcing retrieval of a fresh copy of a file with a link
    ... to serve a fresh copy of a file or B) the browser to "refresh" the ... as changing HTTP headers isn't really ... so adding a 'no cache' meta tag to the file itself ...
    (comp.infosystems.www.authoring.html)
  • Re: Browser Back button problem
    ... the browser / proxy server in HIGH security applications. ... However this generally can be fixed by using HTPP headers "Cache-Control: ... manages HTTP headers in .NET, but i can assure u that there is one. ...
    (microsoft.public.dotnet.framework.aspnet.security)
Loading