logging-log4cxx-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Wilfong, Paul" <paul.wilf...@ngc.com>
Subject RE: Log4cxx - "no default charset decoder available" message
Date Tue, 27 Nov 2007 23:23:30 GMT
Once again - Thank you Curt.  You put me on track to deal with this
problem.

I ended up finding a folder on the Sun that was in a fellow user's area,
something like ".../apr-iconv/lib".  

Repeating the configure, and adding "-with-iconv=" using this folder,
did the trick.

Best - Paul Wilfong



-----Original Message-----
From: Curt Arnold [mailto:carnold@apache.org] 
Sent: Monday, November 05, 2007 1:30 PM
To: Log4CXX User
Subject: Re: Log4cxx - "no default charset decoder available" message


On Nov 5, 2007, at 1:38 PM, Wilfong, Paul wrote:

> I am trying to build log4cxx on a Sun.
>
> When I do the make, I get "#error  No default charset decoder 
> available"  which comes out of charsetdecoder.cpp.
>
>
> Details:
>
> SunOS 5.10, and g++ 3.4.3. The computer is not connected to the 
> internet.
>
> I've managed to make the apr and apr util stuff, but can't install 
> them to /usr/lib due to lack of privilege.  I use --prefix to put them

> in a different location.
>
> I have not built cppunit on this computer and hope to be able to 
> bypass it for now if possible.
>
> For the log4cxx build I am using the following commands:
>    sh ./autogen.sh
>    ./configure --with-apr=blahblah  --with-apr-util=blahblah  -- 
> disable-cppunit
>    make
>
> At this point the error is reported from the charsetdecoder.cpp 
> compile.
>
> The ./configure step reports (among many other things) the following:
>    checking for wchar_t... yes
>    checking for logchar type...  utf_8
>
>
> I've searched the message forums and have found a couple of related 
> entries, but am not sure if they apply to what I am seeing.  Any 
> assistance with this would be greatly appreciated.
>
>


log4cxx needs some mechanism to convert from the current default  
encoding to a known encoding (in your case, UTF-8).  Generally on  
Unix environments, this function is performed by the apr_xlate  
function defined in apr-util which in turn typically delegates to  
iconv.   For you to get that message, it would appear that your apr- 
util build could not find iconv and therefore set APR_HAS_XLATE to 0.

If you know that the default encoding is fixed as UTF-8, ISO-8859-1  
(aka Latin 1) or US-ASCII, you can set a processor macro to indicate  
that.  Otherwise, you need to diagnose why apr-util could not find  
iconv.  On most Unix, running "iconv -l" will list the available  
character sets and can be used as a quick check whether iconv is  
installed.



Mime
View raw message