[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Help: Strange 64bit issue



On Fri, Jul 11, 2008 at 02:40:17PM +0200, Andreas Tille wrote:
> On Fri, 11 Jul 2008, Manuel Prinz wrote:

> >With these fixes it still did not build on my system. I needed to change
> >the Build-Depends on lib64z1-dev into zlib1g-dev to get it to build in a
> >clean pbuilder chroot.

> Well, I guess that lib64z1-dev will not exist for amd64 and that this
> whole mess is just caused by the multiarch stuff.  It's the first time

It doesn't exist for AMD64 - it is 64 bit native so there is no need to
produce a 64 bit cross verson of anything.  There is lib32z1 on amd64, 
allowing 32 bit binaries to be built in an amd64 environment.

> that I have to deal with this and I have the impression that I try to
> add just problems with no real profit for the user of the program.
> Probably I should just exclude the -m64 switch when building for i386
> and everything will work fine.

That is almost certainly what you want to do.  If you build with -m64
you will produce an amd64 binary.  This can be run on i386 systems with
an appropriate processor, kernel and runtime environment but won't run
on systems where one or more of those isn't available and shouldn't be
the standard thing for the Debian port.

Depending on the needs of the package it may make sense to provide both
an i386 native and a cross-built amd64 binary in the i386 port.

> >I cannot reproduce this on my amd64 machine. With the change mentioned
> >above it builds fine and I'm able to run /usr/bin/maq on both lenny and
> >sid. Some output:

> I expect this in 64bit machines - but Charles had problems on hie PowerPC
> as well ...

PowerPC also supports mixed 64/32 bit environments so the situation is
similar to that on x86 and x86-64.  When running on a G5 or other 64 bit
processor with an appropriate kernel it is possible to execute 64 bit
PowerPC programs, even using the 32 bit PowerPC port.

-- 
"You grabbed my hand and we fell into it, like a daydream - or a fever."


Reply to: