Re: Status update: CMUCL port to Darwin/OS X

Raymond Toy said:
>>>>>> "Pierre" == Pierre R Mai <pmai@xxxxxxx> writes:
>
>     Pierre> Hi!
>     Pierre> This is just a short notice that, after only a short year of
> delay ;),
>     Pierre> I've resurrected the PPC backend of CMUCL and  more or less
> completed a
>     Pierre> port of CMUCL to Darwin/OS X.  Current status, as well as
> binaries and
>     Pierre> sources can be had at
>
> Great news!  I assume it was a fairly difficult port.  Did you

Yes and no, actually.  It was a bit of work because the PPC backend had
bit-rotted a bit, and I needed to wrap my head around a couple of
peculiarities, but OTOH I had the SBCL port as a baseline, for most parts.

What mattered more was the horrible state of affairs with gdb and
exceptions/signals on OS X, which cost me lots of hours in debugging.  All
in all the whole porting effort probably took me around 2 weeks of work,
but spread out over the last 9 months, due to a distinct lack of spare
time.

It was more work than merely porting to a new OS, but definitely much less
work than going for a new architecture, due to the great head-start Gary
Byers gave us with the PPC backend.

> cross-compile from x86?

Yes, x86/Linux specifically.  I've got a working cross-compile script for
this combination, if people are interested (though now the port
self-hosts, obviously).

Regs, Pierre.

-- 
Pierre R. Mai <pmai@xxxxxxx>                    http://www.pmsf.de/pmai/
 The most likely way for the world to be destroyed, most experts agree,
 is by accident. That's where we come in; we're computer professionals.
 We cause accidents.                           -- Nathaniel Borenstein