This is the mail archive of the
mailing list for the Java project.
Re: state of umlaut and other char stuff problem
- From: João Garcia <jgarcia at uk2 dot net>
- To: gnustuff at thisiscool dot com
- Cc: Simon Rutishauser <simon dot rutishauser at gmx dot ch>, tromey at redhat dot com, java at gcc dot gnu dot org
- Date: Tue, 29 Jul 2003 17:32:00 +0100
- Subject: Re: state of umlaut and other char stuff problem
- References: <B007B0ZN1V2YHC2UMKYS83042UPO6.3f24a127@p733>
I have been away from this mailing list for some time.
Thanks for the cc.
Mohan Embar wrote:
My instinct was to
go with Unicows in these cases, but I feel bad about neglecting
João's short-term solution.
There is nothing wrong with that since you are not interested in that
At the moment, I am quite happy with my short-term solution and with gcj
I have plans to post a patch that will make libgcj send wide-chars to
the OS or (if required) will make it use libiconv insted of "my" tables.
I think this could also be useful for posix platforms. But I have been
busy with other things and I cannot say when this patch will be ready.
For me, the unicows solution is out of the question. It cannot be
distributed along side with your application, unless you state that your
users cannot re-distribute your application's instalation file (please
read the unicows license for aditional problems...). Anyone can read
about this problem at:
And now for the funny part (and quoting):
"Download the unicows.dll file. (Note: Microsoft seems to enjoy changing
the exact location of this file. You may have to search Microsoft's
If you (or anyone else) are using unicows with gcj to solve filename
problems, you must have a patch to libgcj for it to send wide-chars to
the OS. It would be nice if you could share it with everybody else that
is only interested in NT/2000/XP platform (you DO NOT need unicows for
this platform -- it supports utf-16 by default and libgcj is only
performing an unuseful and destuctive task in this case) or with anyone
that is confortable with the unicows solution for Win9x platform.
Otherwise, I will address this problem ASAP (whatever that may be...).
That's also why I was motivated
to document the MinGW build process so that others are free
to communicate their patches in a way that others can build
gcj with them, even if I don't agree with these or have the
time to evaluate them.
It is a nice work. And it was quite useful to me, thank you.
I think it would be useful for many people if you could provide, in your
website or somewhere else accessible, a tar.gz with the directory
structure and the scripts for the build process. The user would only
have to put the FSF and MingGW sources and "relevant" patches in the
right places of the tree.
I would also suggest the reading of Ranjit's posts and website to anyone
interested on building gcj for MinGW.
But I guess the idea of building gcc/gcj will still keep many potential
GCJ/MinGW people away from using it...