Re: [RFC][PATCH 1/2] binfmt_elf: FatELF support in the binary loader.

[Jeremy Fitzhardinge <jeremy@goop.org>]

On 10/19/09 21:43, Ryan C. Gordon wrote:
> I'm not sure the added flexibility is worth the extra complications. 
> FatELF solves a specific problem: merging multiple ELF targets into one 
> file, the most compelling use-case being to glue x86_64 and i686 binaries 
> together.
>
> What you're describing would definitely be the route I'd have chosen if, 
> say, a.out files were still in widespread use and actively competed with 
> ELF for mindshare.
>   

It would be nice to have something that would conceptually work for
other architectures with other executable file formats.  Granted ELF is
most common, but its far from the only one.

A generic approach would allow the last-option fallback executable to be
a shell/python/perl script which could do something useful (like display
a useful message).

>> That is, what you have here, but without hacking up binfmt_elf more.
>>     
> I like to think of it as art, like a chef carving a fine piece of meat. :)
>   

Well, its really a bit diseased, with lots of gristly bits and a few
unexpected tubes sticking out the side.

    J