chicken-hackers
[Top][All Lists]
Advanced

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

Re: [Chicken-hackers] [ANN]: Port Chicken to AIX


From: Erik Falor
Subject: Re: [Chicken-hackers] [ANN]: Port Chicken to AIX
Date: Tue, 13 Aug 2013 17:27:42 -0600
User-agent: Mutt/1.5.21 (2010-09-15)

On Mon, Aug 12, 2013 at 05:10:14PM -0600, Erik Falor wrote:
> Greetings!
> 
> In an effort to use Chicken at my office, I've gotten Chicken to build
> on AIX 6.1. I can run the interpreter, the compiler, and manage eggs
> with chicken-{install,remove}.  It did not take too many changes to
> the code to make this happen.
> 
> I have more details about it on github.com, along with a list of eggs
> I've been able to build.
> 
> https://github.com/fadein/chicken-core/pull/1
> 
> I hope it is not too late to fit this in to the 4.8.2 release.  Please
> let me know if there are any further changes I ought to make to this
> patch to clean it up.

This is an update on the effort to port Chicken to AIX.  Peter Bex
and rivo403 spent some time with me in #chicken today helping me clean
up Makefile.aix.

1. I'm going to explain more later about the compiler/linker flags
   that I've added to Makefile.aix and csc.scm, but one of them,
   -bbigtoc, seems to be absolutely necessary to even build
   libchicken.so on AIX.  The AIX PowerPC ABI gives a process 64kb of
   space to map the locations of global symbols in a structure called
   the Table of Contents.  The offset of a symbol into that map is
   kept in a 16-bit register.  When you have more than 8192 objects in
   a 64-bit executable you will overflow the Table of Contents.  You
   can get around this limitation by using the -bbigtoc linker flag at
   the price of a performance penalty.

   You can read all of the gory details at the link below.  One
   relevant section is entitled "Address Maps".

   
http://pic.dhe.ibm.com/infocenter/aix/v6r1/index.jsp?topic=/com.ibm.aix.cmds/doc/aixcmds3/ld.htm

2. I discovered that I have access to a copy of the XL C compiler on
   one of the AIX machines at my office.  After I sort out GCC to a
   satisfactory degree, I'll try it out.  I expect to hit fewer snags
   with it.

3. I'm working on getting repository-path to work in
   -private-repository mode. Once this is finished, I fully expect
   `make check` to succeed.

4. Everybody has been super supportive and helpful.  I can't think of
   a better community to hang out in!

Thanks,

-- 
Erik Falor                                       http://unnovative.net
Registered Linux User #445632                  http://linuxcounter.net

Attachment: signature.asc
Description: Digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]