packages (was Re: [MLton] syntax error for "_address")

John Reppy jhr@cs.uchicago.edu
Fri, 4 Nov 2005 12:07:43 -0600


On Nov 4, 2005, at 11:53 AM, Stephen Weeks wrote:

>
>> I guess the point I'm trying to make about whole-program compilation
>> is that the mlton command is this big monolithic operation that my
>> build script or makefile has to swallow whole.  It is not broken up
>> by file or library, or even by phase (compile vs. linking).
>
> Like Matthew, I see this as avoiding the middle man :-). It seems to
> me that in whatever model one has, it should always be possible to
> replace steps that do preprocessing (in my view) of source files, like
> building a .o, with identity steps that simply propagate the source
> files.  I don't see how whole-program compilation can get in the way.

In the final analysis it should be possible to work around it, but it  
does
break the natural structure of make (e.g., not being able to compile
libraries separately).

>
> Here's a more concrete question that will perhaps help us make
> progress.  Since frameworks are the preferred way of tackling this
> problem on Mac OS, would it be possible to use frameworks to package
> MLton libraries?  If not, why not?

A framework is just a directory tree that contains headers, libraries,
and resources in standard places.  They use a plist (property-list file
in XML representation) as a kind of manifest.  One could easily add an
MLB file and SML source code to a framework.

>
>> BTW, the default configuration of mlton, which assumes that it lives
>> in /usr is not a good match for MacOS X.  You should use /usr/local
>> and/or create an application bundle.
>
> Since I know nothing about the latter, I have done the former.
> Thanks.

Thanks.

	- John