[Openmcl-devel] CCL on Sonoma

Tim McNerney mc at media.mit.edu
Mon Nov 13 11:11:32 PST 2023

Thanks for the update on Apple possibly fixing Rosetta 2.  I worked with 
an engineer friend at Apple to look into this.

FWIW, Ayla Kurdak is available and interested, and I am willing to 
supervise her again, but she'd need to get paid.
I had been planning to tap her to port my company's Forth implementation 
to a Lisp-based VM, until we had a sort-of hiring freeze.


On 11/13/23 1:03 PM, Ron Garret wrote:
> I recently upgraded one of my M1 machines to Sonoma and tried building CCL and putting it through some basic stress tests.  It (mostly -- see below) worked.  In particular, I did not get any of the weird Rosetta crashes that have been observed on earlier versions of MacOS.  I think it's possible that Apple may have fixed this bug, and CCL under Rosetta 2 may be reliable now.  I have not done anywhere near enough testing to be confident that all of the problems have been fixed, but it's looking promising enough that it seemed worth mentioning here so that other people might be motivated to try it.
> Note: I did have one weird error while loading code via Quicklisp: at one point I got a "bad file descriptor" error.  I was able to restart the process and it completed successfully, but this is still super-weird.  Given that there have been weird intermittent failures of CCL running under Rosetta 2 in the past I thought I would mention it in case anyone else encounters this.
> As long as I'm writing, is there any news on the porting effort?
> rg

More information about the Openmcl-devel mailing list