[Simh] Readline support gone?

Kevin Handy kth at srv.net
Wed Sep 17 20:37:57 EDT 2008


Philipp Hachtmann wrote:
>   
>> In my opinion, worse than poor performance, is that recursive makefiles 
>> risk correctness of the build.
>>
>>     
> Hm... With a set of correct rules, this should not be a problem.
> What do you mean with incorrect build?
>
> I think, the best solution for SIMH would be a single improved makefile.
>   
In rewriting the Makefile, be careful about handling
the command line options that vary between simulators
for the same source files, i.e. -DECLIPSE, -DUSE_INT64,
and others.

FYI: The current makefile looks a lot like a batch file,
because it started out as a batch file. I created it
during the vax/pdp10 development era when the
compilation method was to type the commands at the
command line (no build scripts existed), and the
commands were getting too complicated. This was
just a quick and dirty to make the existing build
process possible without having to type in a huge
command (multiple lines on the terminal). This
hack was turned that into a makefile so that you could
specify what you wanted to build, and made it much
more portable. Others have hacked on it, and you
have what exists today.

It might make sense to use autoconf/automake to
handle all the system dependent stuff and the
multiple objects due to -D stuff. Things like 64 bit
file systems, networking, threads, etc.. I don't
think autoconf works under windows, unless you
run it under cygwin, so a regular makefile will
also be necessary.




More information about the Simh mailing list