Re: [Chicken-users] Compiling eggs to .c

[Norman Gray]

Christian, hello.

On 25 Oct 2016, at 13:37, Christian Kellermann wrote:

> It looks like
> >     % chicken-install -n uri-match
> >
> > should work, because that retains the intermediate files -- including 
> > the .c
> > files -- in the temporary (download) directory.  But (a) the only way 
> > I can
> > see of finding what that temporary directory is, is by scanning the 
> > output
> > chatter from chicken-install, which obviously isn't robust in a 
> > Makefile;
> > and (b) scavenging the two .c files from this directory (namely
> > uri-match{,.import}.c) seems a slightly ad-hoc/hacky way of obtaining 
> > them,
> > which suggests I'm Doing It Wrong.
[...]
> The -k option should keep intermediate files, you can pass this with
> CSC_OPTIONS.

Thanks for this.  However, my understanding is that the -n option to 
chicken-install implies -k, and additionally stops after building (which 
includes building the .c from the .scm) and before the extension is 
installed (so avoiding changing the rest of the system).

With 'chicken-install -n' I appear to have the required .c files ... 
somewhere, but the only way of finding where they are appears to be 
reading the chicken-install chatter to find the location of the temp 
directory.  I was wondering if there was a more automatable way.

> For my projects I have manually unrolled the compilation
> stuff into a shell script. Yes it is tedious and sucks but it
> works.

I was really hoping to avoid that, partly because I'm not confident I'd 
get all the build invocations right for the various extensions, so I'd 
never really know if subsequent problems were because of that or not.  
But the trial-and-error might be quicker than what I'm doing just now.

> Also for CHICKEN 5 the new build system will make things easier...

Roll on Chicken 5!

All the best,

Norman


--
Norman Gray  :  https://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK