On 5/16/07, Stefan Küng <firstname.lastname@example.org> wrote:
> Mark Phippard wrote:
> >> * it uses new/delete instead of apr pools to reserve the memory
> > I think you submitted a patch to the thread that does this.
> Yes I did. But that was some time ago:
> But then, it got rejected with an argument I simply couldn't do anything
> I can't prove rumors wrong.
> So this time I decided to not send the patch at all but keep it to
> myself. Because I still can't prove that rumor wrong. So there was no
> point in even trying.
> >> Yes, I don't define APR_HAS_XLATE. If that isn't defined but WIN32 is,
> >> then the windows APIs are used to convert from/to utf8.
> > Using the "normal" Windows build would I have to edit the generated
> > DSW files? I really need this in Subclipse and I do not need the
> > --encoding option either.
> Well, I'm using NAnt scripts to build all the libraries (that way,
> patching them is much easier to do, and I can customize the build the
> way I want and need to). I haven't really tried building with the *.dsp
> files (*.dsw are the workspace files, I think you meant the *.dsp files
> which are the project files for the builds).
> First, you need to 'patch' apr-util:
> apr-util\include\apu.hw (or apu.h if it already exists):
> there's a define in there
> #define APU_HAVE_APR_ICONV 1
> #define APU_HAVE_ICONV 0
> #define APR_HAS_XLATE (APU_HAVE_APR_ICONV || APU_HAVE_ICONV)
> set APU_HAVE_APR_ICONV to 0 instead of 1.
I followed these instructions and built with your utf.c file and
latest trunk. The Java bindings still pass their tests OK, but the
command line gets an assertion failure at line 1185 of utf.c. Even
when running svn --version.
To unsubscribe, e-mail: email@example.com
For additional commands, e-mail: firstname.lastname@example.org
Received on Thu May 17 13:33:59 2007