On Fri, 18 Apr 2008, Jeff Breidenbach wrote:
For packaging purposes, I'm planning to change jcc/setup.py as follows so that
autobuilders have fewer problems. It means there are a couple extra paths for
the linker / dynamic loader to search, but I don't see this as a big deal.
Q1: I notice that rpath i386/AMD64 don't really match; one talks about client,
the other talks about a server subdirectory. Is this a simple oversight?
No, there doesn't seem to be 'client' thing on 64-bit VMs.
Q2: Maybe this should be considered for the office version of setup.py
after all ?
What do you mean by 'office' version ?
I'd expect really strange bugs to crop up by throwing all these paths for
incompatible architectures into the link flags.
You're either on i386 or on amd64 but not on both.
The problem with putting anything hardcoded into setup.py as this (or what's
there now) is that it's ultimately only an example. The correct values are
specific to your environment. Hence the side way of being able to pass these
values in via environment variables instead of modifying setup.py.
Your debian build script should invoke setup.py by passing it the correct
values for these settings by setting JCC_CFLAGS, JCC_LFLAGS and JCC_INCLUDES
env vars. When there are several parameters to pass per variable, separate
them with ':' (os.pathsep). If this conflicts with arg values, you can set a
different separator with JCC_ARGSEP (new in JCC 1.9 released today).
Andi..
Jeff
===
LFLAGS = {
'darwin': ['-framework', 'JavaVM'],
'linux2': ['-L/usr/lib/jvm/java-6-sun/jre/lib/i386',
'-L/usr/lib/jvm/java-6-sun/jre/lib/amd64',
'-ljava',
('-Wl,-rpath=/usr/lib/jvm/java-6-sun/jre/lib/amd64:'
'/usr/lib/jvm/java-6-sun/jre/lib/amd64/server:'
'/usr/lib/jvm/java-6-sun/jre/lib/i386:'
'/usr/lib/jvm/java-6-sun/jre/lib/i386/client')],
'sunos5': ['-L/usr/jdk/instances/jdk1.6.0/jre/lib/i386', '-ljava',
('-R/usr/jdk/instances/jdk1.6.0/jre/lib/i386:'
'/usr/jdk/instances/jre/lib/i386/client')],
'win32': ['/LIBPATH:o:/Java/jdk1.6.0_02/lib', 'jvm.lib']
}
_______________________________________________
pylucene-dev mailing list
[email protected]
http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
_______________________________________________
pylucene-dev mailing list
[email protected]
http://lists.osafoundation.org/mailman/listinfo/pylucene-dev