yeah, this is an ugly one.  Basically depending on i386, sparc (and
hopefully alpha), the distfile is different.  Each distfile has different
contents.  In fact the i386 creates its own subdir under 'work', but the
sparc one doesn't.  Then its different shell scripts.  The install script
basically looks at machine arch to decide if it will even let you install.
Then after its all installed, theres another patch which must be applied
to the acroread shell script to again patch for a test which only runs the
binary if you have the right (in their eyes) machine arch.

its all pretty ugly.


maybe I should just try and work around even using their install script,
but I haven't deciphered all of what they're doing with it.

-Dan

On Wed, 14 Jul 1999, Hubert Feyrer wrote:

> On Tue, 13 Jul 1999 mcmahill@mtl.mit.edu wrote:
> > PATCHDIR=	patches.${MACHINE_ARCH}
> 
> At least for patching "C" code, the policy is to put everything into one
> source, and decide which code to actually use with cpp magic (#define,
> ...).
> 
> Knowing that you probably don't have sources for acroread, I guess that
> you want to patch an /bin/(c?)sh install-script here, which is more
> interresting. Solutions I can see here:
>  - use 'uname -sr' to decide what to do
>  - "." or sh in appropriate files provided in acroread/files
> 
> Not knowing of the exact problem, i can't guess more solutions right now.
> :)
> 
> 
>  - Hubert
> -- 
> NetBSD - Better for your uptime than Viagra
>