Bug#446276: webkit: FTBFS: TCSystemAlloc.cpp:89: error: expected constructor, destructor, or type conversion before '(' token

Jay Berkenbilt qjb at debian.org
Sat Oct 13 16:30:36 UTC 2007

> When looking at the Makefile, I see things like this in the
> CXXFLAGS and LIBS: "### or --detect-prefix"
> This seems to come from running icu-config:
> $ icu-config --cppflags
> ### icu-config: Can't find /usr/lib32/libicuuc.so - ICU prefix is wrong.
> ###      Try the --prefix= or --exec-prefix= options
> ###      or --detect-prefix
> ### icu-config: Exitting.
> While it should return something like:
> -D_REENTRANT -I/usr/include
> This seems be a bug in libicu36-dev, so reassigning.  libicu-dev 3.8
> from experimental doesn't have this problem, even though
> /usr/lib32/libicuuc.so isn't present either.  And there is no reason it
> should be.

I recently modified the ICU packages to build 32-bit libraries, in
addition to 64-bit libraries, on amd64 and a few other 64-bit
platforms since this was needed by wine.  Although the library isn't
present on 3.8 d1 in experimental, it is present on the real 3.8
packages which I have had ready to upload for a month but have been
holding off until the boost transition finishes.

I don't know why webkit is trying to use the 32-bit libraries to begin
with, but I guess it looks like icu-config --cppflags is at fault.
I'll try to fix this.  The problem is that I don't have access to an
amd64 system.  According to db.debian.org/machines.cgi, pergolesi is
supposed to be one, but if I check there (from dchroot sid):

> pergolesi% dpkg-architecture 
> DEB_BUILD_GNU_TYPE=i486-linux-gnu
> DEB_HOST_GNU_TYPE=i486-linux-gnu

Am I missing something?  Anyway, if I can get access to an amd64
machine, I can try to fix this once and for all.

Jay Berkenbilt <qjb at debian.org>

More information about the Pkg-webkit-maintainers mailing list