[Top][All Lists]

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

Re: [Gcl-devel] Mach-O BFD back-end

From: Camm Maguire
Subject: Re: [Gcl-devel] Mach-O BFD back-end
Date: 29 Oct 2003 20:37:45 -0500
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2


Aurelian, this is absolutely wonderful!  Great job!

Please feel free to commit any macosx-specific changes you may need in
the bfd subdirectory.  I'll try to get to your posted patches in the
main files as soon as I can.  My apologies for the delay.

I'm just wondering now if macosx happens to use a gp pointer in its
relocation logic.  An example implementation of 
bfd_get_relocated_section_contents on such a machine would likely open
up in one step all the remaining others.  My guess though is that
macosx is closer to ppc linux instead.  

What kind of logic is needed in build_symbol_table_bfd?  Would be nice
to standardize this.  Right now the logic is determined very
experimentally via analysis of GCL compiled .o files.  

Take care, and thanks again!

Aurelien Chanudet <address@hidden> writes:

> Hi all,
> It looks like the Mac OS X port is getting into good shape. Jun Sawada
> has just reported about successfully compiling and testing ACL2/GCL
> with gcc 3.1, and I've done the same myself with gcc 3.3. Maybe we're
> no ready to ship a binary image !
> On top of that, I've nearly finished adding missing functionalities to
> the Mach-O BFD back-end. This means that I'm now able to build GCL
> using the --enable-locbfd configure switch. I haven't thoroughly
> tested the dynamic loading mechanism yet, but, from what I can see,
> everything seems to work fine already.
> In order to get a functional libbfd.a for Mach-O, some source files
> could be added to GCL's local BFD tree. Assuming a functional libbfd.a
> targeting Mach-O, all that needs to be done to build GCL is to
> slightly modify build_symbol_table_bfd. This is because, it cannot be
> inferred that an undefined symbol is imported from a library if its
> name contains the substring "@@" (which is what the existing code in
> build_symbol_table_bfd does, as I understand it) on Mac OS X.
> I'm now aiming at getting a fully reliable Mach-O BFD backend.
> Aurelien
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/gcl-devel

Camm Maguire                                            address@hidden
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah

reply via email to

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