[gambit-list] Porting Gambit

Paul Wolneykien manowar at altlinux.org
Thu Nov 22 00:57:47 EST 2018


21.11.2018 19:50, Marc Feeley пишет:
> When starting a new port you should try to build using the latest release (not a clone of the git repo) so that you avoid the bootstrap process.  Also it is useful to configure with --enable-ansi-c to only use the ANSI-C libraries.  Once that works you can remove that configure flag to debug the more advanced features.
> 
> Are you building Gambit on the E2K itself?  Cross-compiling may cause some issues.  Make sure the configuration process yielded reasonable values.  One aspect that is not handled automatically is the endianness of the processor.  The endianness can be forced explicitly with:
> 
> ./configure --enable-debug --enable-ansi-c CC="gcc -D___LITTLE_ENDIAN"
> make

  Thanks, Marc! It builds (and seems to work!) with these options. Now
I'll try to build it with production options, such as
enable-single-host, etc. There are some minor problems with outdated
config.guess. I'll report back all changes if they arise.


> Try using the debugger to run the generated gsi or gsc, to see where the segfault occurs. For example:
> 
> gdb gsi/gsi
> run
> 
> If you get it working please report back the changes needed.
> 
> Marc
> 
> 
> 
>> On Nov 21, 2018, at 9:42 AM, Paul Wolneykien <manowar at altlinux.org> wrote:
>>
>>
>>  Hi! Tell me please, is it really hard to port Gambit to another
>> platform (CPU)? Is there any platform-specific code?
>>
>>  Currently, when I'm trying to build the vanilla sources from
>> github.com/gambit/gambit.git , it segfaults at the gsc/ part. The gsi
>> and gsc-boot are built, but they seems to be unusable: both segfaults
>> instantly.
>>
>>  The arch I'm trying to build for is E2K
>> (https://en.wikipedia.org/wiki/Elbrus_2000).
>> _______________________________________________
>> Gambit-list mailing list
>> Gambit-list at iro.umontreal.ca
>> https://mailman.iro.umontreal.ca/cgi-bin/mailman/listinfo/gambit-list




More information about the Gambit-list mailing list