[NBLUG/talk] [SoCoSA/discuss] wierd dash character

[gandalf at sonic.net]

Well Eric was the most right out of three right answers. Sorry I didn't reply sooner, but I've been gone all weekend and didn't have a chance to try stuff out. Fanime Anime Convention Rocks! Anyway, the pages were all being served UTF-8 and I checked the conf and low and behold there is a section for it. It seems that if you set a default it OVERRIDES what the individual web pages specify in META tags. It had UTF-8 as the default. I commented it out, reloaded and viola, everything seems to be working fine. Long dashes look like long dashes. Funny quotes look like funny quotes. This was a colo server at sonic running the current version of CENT-OS.

# Specify a default charset for all content served; this enables
# interpretation of all content as UTF-8 by default.  To use the
# default browser choice (ISO-8859-1), or to allow the META tags
# in HTML content to override this choice, comment out this
# directive:
#
#AddDefaultCharset UTF-8




"There probably isn't any meaning in life. Perhaps you can find something interesting to do while you are alive." - Orochimaru (Naruto)

 On Thu 22/05/08  7:08 PM , Eric Eisenhart eric at nblug.org sent:
> It's likely that before the upgrade your server was sending a  
> 
> different (correct) character set, not that it was actually converting.
> 
> 
> 
> The diamond with a question mark is the "I don't know what this  
> 
> character is" or "I don't have this character in my font" indicator.
> 
> 
> 
> If you can, try one of these against both the old server and the new  
> 
> (change URL, of course):
> 
> HEAD http://www.example.com/foo.html
> curl -I http://www.example.com/foo.html
> wget -S -O /dev/null http://www.example.com/foo.html
> 
> 
> Look at the "Content-Type" header.  It's probably something like  
> 
> "Content-Type: text/html; charset=utf-8" or "Content-Type: text/html;  
> 
> charset=iso-8859-1".
> 
> 
> 
> You can put "AddDefaultCharset ISO-8859-1" or "AddDefaultCharset  
> 
> UTF-8" into your apache configuration to switch to the other behavior,  
> 
> or you can put one of these into the  of your HTML files:
> 
> 
> charset=iso-8859-1"  />
> 
> 
> 
> 
> I believe apache will actually read that meta out and pass it in the  
> 
> HTTP headers.
> 
> 
> 
> There's lots of other charset options out there, iso-8859-1 and utf-8  
> 
> are simply the most likely (other than "us-ascii, which doesn't have  
> 
> the funky characters that seem to be causing you problems)
> 
> 
> 
> On May 22, 2008, at 6:51 PM, gandalf at son
> ic.net wrote:
> 
> 
> > I have a weird dash character that is causing
> problems on web pages.  
> > It's showing up as a diamond with a question mark
> in browsers. In a  
> > GUI text editor it shows up as a dash. In vi and
> less it shows up as  
> > . In cat and more it does not show at all. I
> can't seem to grep  
> > for it as I can't seem to figure out what it is. I
> guess I could  
> > write perl code to bring it up with the text around
> it in a reg exp  
> > and then see what the ascii value is or something.
> I'm not sure how  
> > I ended up with these. I'm suspicious that is' the
> long dash from MS- 
> > Word or something.
> 
> >
> 
> > I'm almost certain these were not visible before we
> upgraded our  
> > server. Perhaps the old apache had a built in
> conversion for them.
> >
> 
> >
> 
> > "There probably isn't any meaning in life. Perhaps
> you can find  
> > something interesting to do while you are alive." -
> Orochimaru  
> > (Naruto)
> 
> >
> 
> >
> 
> >
> 
> >
> 
> > _______________________________________________
> 
> > SoCoSA discuss mailing list
> 
> > discuss at so
> cosa.org
> > Your address: freiheit@
> socosa.org
> > http://socosa.org/mailman/listinfo/discuss
> > http://socosa.org/mailman/options/discuss/freiheit%40socosa.org
> 
> >
> 
> 
> 
> 
> 
> _______________________________________________
> 
> SoCoSA discuss mailing list
> 
> discuss at so
> cosa.org
> Your address: gandalf at son
> ic.net
> http://socosa.org/mailman/listinfo/discuss
> http://socosa.org/mailman/options/discuss/gandalf%40sonic.net
> 
> 
> 
> 
>