cross-compiling

cross-compiling

[chuckr]

I have a Linux PDA, a Sharp Zaurus running Linux 2.4.20, and I like to 
cross-compile my applications on my main Intel Linux box sitting next to 
it, connected via usb networking.  Your configure script does nicely 
realize that it's doing cross-compiling, and it even goes an extra step, 
if I don't give it any of the --target, --build, or --host flags, it 
gets it from the fact that the gcc in the path isn't making intel code.

The problem is, in your configure script, the section that's setting 
setpgrp, it's detecting that it is doing that cross-compiling, and seems 
to be reacting by yielding ann error and dying immediately and 
unconditionally (configure line 8478, search for getpgrp, then "cross".)

This really needs fixing.  I guess I could have it wrong, but I tried 
just about any combination of flags I could imagine (including those 
that work for other packages).  If I have it wrong, if you'd please 
correct me, and also you might want to consider a line in the INSTALL 
file saying how to get that done, many folks having PDAs want to do 
cross compiling, you want to allow for that.

Thanks very much!

Re: cross-compiling

[Peter Stephenson]

chuckr wrote:
> I have a Linux PDA, a Sharp Zaurus running Linux 2.4.20, and I like to 
> cross-compile my applications on my main Intel Linux box sitting next to 
> it, connected via usb networking.  Your configure script does nicely 
> realize that it's doing cross-compiling, and it even goes an extra step, 
> if I don't give it any of the --target, --build, or --host flags, it 
> gets it from the fact that the gcc in the path isn't making intel code.

Unfortunately that's a bit of a red herring (and it should really say so
in the INSTALL file).  The configure basics support cross-compiling, but
many zsh tests actually need to run on the system in question to find
out how the features work.  There are numerous cases where simply
compiling isn't good enough to work out what's going on.

This could be fixed along the lines of making some assumptions for tests
when cross-compiling, and telling the user what might need fixing.
Possibly some places where we need to search header files for
information (e.g. signal names) need fixing too.

That's already a sizeable chunk of work and can only be done by someone
who is set up to use it.  I can't see it happening unless somebody
volunteers, I'm afraid.

-- 
Peter Stephenson <pws@csr.com>                  Software Engineer
CSR Ltd., Science Park, Milton Road,
Cambridge, CB4 0WH, UK                          Tel: +44 (0)1223 692070


**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.

This footnote also confirms that this email message has been swept by
MIMEsweeper for the presence of computer viruses.

www.mimesweeper.com
**********************************************************************