[gambit-list] planning Gambit 5

Marc Feeley feeley at iro.umontreal.ca
Tue Dec 3 08:47:53 EST 2019


> On Nov 29, 2019, at 7:46 AM, Sven Hartrumpf <hartrumpf at gmx.net> wrote:
> 
> Hi Marc.
> 
> Marc wrote, 2019-11-20 08:55:
>> With all the changes to Gambit recently it is time to think about Gambit v5.0.0
>> which will probably be released in the next 6-12 months
> 
> Very good news!
> 
> What are your plans for the native backend?

Gambit’s development is advancing on several fronts including the module system, SMP scheduler, universal backend, native backends, and various side projects.

Thanks to Laurent Huberdeau’s work in 2018 and Abdelhakim Qbaich’s work in 2019, the native backend for x86 has 
advanced nicely and is currently in a usable state (it can compile most Scheme programs you throw at it).  However it is not fully integrated with the rest of the system and some primitives still rely on the C backend.  I would like the native backend to be able to compile all of the Gambit runtime system without needing the C backend as a fallback.  This goal is not high-priority currently, but if you or others are interested in helping out I can guide you on how you can contribute.

If you are interested in trying it out you need to use the --enable-lowlevel-exec configure flag, and compile/execute programs like this:

% gsc -target x86-64 prog.scm 
% gsi proc

A higher priority goal is implementing a better linker that takes advantage of the new module system to simplify the linking of programs (for the programmer) and generate much smaller compiled programs.  This will have a positive impact on the C, native and universal backends.

Marc





More information about the Gambit-list mailing list